article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
we consider a scalar field satisfying the helmholtz equation with frequency @xmath1 in @xmath2 . given a prescribed incident field @xmath3 , a non - singular solution of @xmath4 we are interested in the solution @xmath5 of @xmath6 where , for @xmath7 , @xmath8 , and @xmath9 equals @xmath10 inside the inhomogeneity and @xmath11 outside .
we take the inhomogeneity to be a ball of radius @xmath0 .
the coordinate system is chosen so that the inhomogeneity is centered at the origin . in other words
@xmath12 we assume that both @xmath11 and @xmath10 are real and positive .
we assume that the scattered field satisfies the classical silver - mller @xcite outgoing radiation condition , given by @xmath13 where , as usual @xmath14 .
the purpose of this paper is to provide sharp estimates for the scattered field @xmath15 , for any contrast @xmath16 and any frequency @xmath17 .
the norms we use to describe the scattered field are the following . given any @xmath18 , its restriction to the circle @xmath19
can be decomposed in terms of the spherical harmonics given by , in the following way @xmath20 and @xmath21 can be measured in terms of the following sobolev norm @xmath22 for any real parameter @xmath23 . by density , this norm can be defined for less regular functions . for radius
independent estimates , we shall use the norm @xmath24 it is easy to see that this norm is finite for a smooth @xmath25 with bounded radial variations . for a radial function , this is simply the sup norm for @xmath25 .
finally , to document the sharpness of our estimates , we will provide lower bounds in terms of the semi - norms @xmath26 where @xmath27 are integers and @xmath28 is a real parameter .
these norms are satisfy the following inequality @xmath29 and if for all @xmath30 , @xmath31 only has one non zero spherical harmonic coefficient , @xmath32 they are the natural extension of the norms introduced in @xcite for the two dimensional companion problem . when the incident field is a plane wave , @xmath33 where @xmath34 is a unit vector in @xmath2 , for all @xmath35 , @xmath36 whereas for any @xmath37 and @xmath28 , @xmath38 where @xmath39 for all @xmath40 and @xmath10 ( see ) .
the motivation from this work comes from imaging . in electrostatics ,
the small volume asymptotic expansion for a diametrically bounded conductivity inclusion is now well established , and the first order expansion has been shown to be valid for any contrast @xcite . it is natural to ask whether such expansions could also hold for non - zero frequencies , even in a simple case .
another inspiration for this work is recent results concerning the so - called cloaking - by - mapping method for the helmholtz equation . in @xcite ,
the authors show that cloaks can be constructed using lossy layers , and that non - lossy media could not be made invisible to some particular frequencies ( the quasi - resonant frequencies ) . within the range of non - lossy media , one can ask whether such ` cloak busting ' frequencies are a significant phenomenon , that is , would appear with non - zero probability in any large frequency set , or on the contrary if they are contained in a set whose measure tends to zero with @xmath0 .
these questions were considered in two dimensions in @xcite . in this paper
we show that these results extend , after some adjustments , to the three dimensional case .
the proofs presented in this paper are very similar to the ones of the two - dimensional paper , but we believe the results , more than their derivation , could be of interest to researchers in various areas of mathematics . in numerical analysis
they could be used as a validation test for broadband helmholtz solvers , since we provide both upper and lower bounds for the scattering data . in the area of small volume expansion for arbitrary geometries , or in the mathematical developments related to cloaking , they provide a best case scenario which can be used to document the sharpness of more general estimates . to make the results of this paper accessible to readers who are not familiar with bessel functions ,
the main estimates are written in terms of the norms @xmath41 and @xmath42 introduced above , and powers of @xmath43 . because no unknown constant @xmath44 appears in the results , this paper can be used as a black - box if the reader wishes to do so .
bessel functions do appear in one place , to describe quasi - resonances , but they turn out to be of the same nature as the two - dimensional ones , and so we refer to @xcite in that case . the main results of the paper are presented in section [ sec : main ] . the proofs are given in section [ sec : proofs ] .
to state our results , we introduce the rescaled non - dimensional frequency @xmath45 , and the contrast factor @xmath46 given by @xmath47 the following theorem provides our estimates for either small frequencies or for any frequency .
[ thm : estim - all ] for any @xmath48 , when @xmath49 there holds @xmath50 if @xmath51 is the first integer such that the @xmath52 spherical harmonic decomposition coefficient of @xmath53 is non zero for some @xmath54 , then holds for all @xmath55 .
furthermore , for any @xmath48 , when @xmath49 , the scattered field @xmath15 also satisfies @xmath56 when @xmath57 there holds @xmath58 naturally , the variant of incorporating the more precise estimate given by for the fourier coefficient corresponding to @xmath59 also holds by linearity .
it is easy to verify that the dependence on @xmath45 in is optimal by a taylor expansion around @xmath60 ( or @xmath61 ) for a incident wave @xmath3 with only one ( or two ) non - zero spherical harmonic coefficients for @xmath59 ( and @xmath62 ) .
theorem [ thm : estim - all ] shows that this estimate is valid up to rescaled frequencies of order @xmath63 when @xmath64 , and of order @xmath65 when @xmath66 . to prove the optimality of these ranges , we define , for @xmath67 , @xmath68 and for @xmath69 , @xmath70 [ pro : lowerbounds ] when @xmath71 and @xmath72 , @xmath73 when @xmath66 , @xmath74 for any integer @xmath75 , @xmath76 in fact estimate holds if the supremum is taken in the set @xmath77 note that @xmath78 when @xmath74 . to illustrate the sharp contrast between what these estimates show when @xmath64 and when @xmath79 , let us consider the case of a plane wave .
when @xmath64 , theorem [ thm : estim - all ] and proposition [ pro : lowerbounds ] show that for any @xmath48 there holds @xmath80 when @xmath66 , theorem [ thm : estim - all ] shows that for any @xmath81 , @xmath82 and for any @xmath83 , @xmath84 the combination of these two estimates do not provide a bound when @xmath85 for the near field @xmath86 .
proposition [ pro : lowerbounds ] provides a lower bound in that case . for any @xmath87 and any @xmath23 , @xmath88 this lower bound grows geometrically with the upper bound of the interval of frequencies considered .
in particular , for any @xmath66 , any @xmath89 , and @xmath90 , @xmath91 this unbounded behavior of the scattered field is due to the existence of quasi - resonant frequencies , just as in the two - dimensional case . to characterize these quasi - resonances ,
bessel functions are required . for @xmath92 , we denote by @xmath93 the hankel function of the first kind of order @xmath94 .
the bessel functions of the first and second kind of order @xmath94 are given by @xmath95 , and @xmath96 .
we denote by @xmath97 the m - th positive solution of @xmath98 .
we denote by @xmath99 the m - th positive solution of @xmath100 .
finally , we write @xmath101 the first positive solution of @xmath102 .
[ def : qr ] for any @xmath92 , the triplet @xmath103 is called quasi - resonant if @xmath104 and @xmath105 the following proposition is proved in @xcite in the case when @xmath94 is an integer , but the proof is unchanged for any @xmath106 .
[ pro : dixon ] for any @xmath106 and @xmath107 , in every interval @xmath108 there exists a unique frequency @xmath109 such that the triplet @xmath110 is quasi - resonant .
there are no other quasi - resonances .
in particular , no quasi - resonance exists in the interval @xmath111 , or when @xmath112 .
since for any @xmath46 there is only a countable number of quasi - resonant triplets , one could hope that outside security sets around the quasi - resonant frequencies , the scattered field could be bounded from above , even in the near field .
this means excluding a countable union of intervals : a trade - off occurs between how much in the near - field one wishes to go , and how large the set of authorized frequencies is .
the theorem below is the result of such a trade - off .
[ thm : broadband ] for all @xmath113 , all @xmath114 , @xmath115 , and @xmath116 $ ] , there exists a set @xmath117 depending on @xmath0 , @xmath46 and @xmath118 such that @xmath119 and @xmath120 if the contrast @xmath46 is less than @xmath121 , this holds with @xmath122 .
altogether , the conditions ( [ eq : eq - intro-1],[eq : eq - intro-2],[eq : eq - intro-3 ] ) imply that the incident field @xmath3 , the scattered field @xmath123 and the transmitted field @xmath124 for @xmath125 , admit series expansions in terms of special functions , namely @xmath126 in the above formulae , @xmath127 , and @xmath128 is the spherical hankel function of the first kind of order @xmath129 , and @xmath130 are the spherical harmonics , given in terms of the polar coordinate @xmath131 $ ] and @xmath132 by @xmath133 where @xmath134 is the associated legendre polynomial .
the reflection and transmission coefficients @xmath135 and @xmath136 are given by the transmission problem on the boundary of the inhomogeneity , that is , at @xmath137 .
they are the unique solutions of @xmath138 which are @xmath139 and , after a simplification using the wronskian identity satisfied by @xmath140 and @xmath141 , @xmath142 in , and , the @xmath143 symbol is an equality if the right - hand - side is replaced by its real part , the fields being real . by a common abuse of notations ,
in what follows we will identify @xmath3 and @xmath123 with the full complex right - hand - side . to verify that this is the correct solution , we need to check that @xmath135 and @xmath136 are well defined .
the fact that there is a unique solution to problem satisfying the radiation condition is well known ( see e.g. @xcite ) .
@xmath144 is non zero for all @xmath145 , @xmath146 and @xmath113 .
assume , for contradiction , @xmath147 .
+ then , as @xmath148 and @xmath149 do not have common zeroes ( see e.g. @xcite ) , either @xmath150 is non zero , in which case - has the following solution @xmath151 or @xmath152 is non zero , and - has the following solution @xmath153 both and would be solution of problem - without an incident wave .
there is of course another solution to that problem , @xmath154 .
since problem - is well posed , see @xcite , we have a contradiction .
we chose the three ( semi-)norms @xmath155 , @xmath156 and @xmath157 because they are compatible with expansions , and .
in particular , for any @xmath35 , we have @xmath158 and @xmath159 whereas the other norms are @xmath160 and for any @xmath27 and @xmath161 , @xmath162 the two dimensional results found in @xcite are easily translated into three dimensional ones for the following reason .
[ pro:2d3d ] let @xmath163 be the reflection coefficient associated to problem posed in a disk of radius @xmath0 in dimension 2 , with the appropriate out - going radiation condition .
this reflection coefficient is defined ( by the same formula ) when @xmath129 is an arbitrary positive real number , and @xmath45 is replaced by any real @xmath164 .
then , for any @xmath165 and any @xmath164 there holds @xmath166 from this point
onwards , we use the short - hand @xmath167 to represent the number @xmath168 . for @xmath169 we introduce @xmath170 then , the reflection function @xmath171 introduced in two dimensional problem considered in @xcite is @xmath172 where @xmath173 . for any @xmath174 and any @xmath175 such that @xmath176
, we write @xmath177 the properties of the function @xmath178 were studied in @xcite for any @xmath179 .
when @xmath94 is an integer , @xmath178 is the reflection coefficient associated to problem in dimension @xmath43 .
note that the identities @xmath180 valid for any @xmath181 yield that @xmath182 we introduce the notations @xmath183 then , the reflection coefficient corresponding to is given by @xmath184 if we introduce for any @xmath181 , outside the zeroes of @xmath185 and @xmath186 , @xmath187 we can rewrite , when @xmath188 is not a zero of @xmath189 , @xmath190 or @xmath191 , @xmath192 from it follows that for any @xmath193 , @xmath194 where these functions are defined .
this in turns implies that @xmath195 and this identity extends to the zeroes of @xmath189 , @xmath190 or @xmath191 by continuity .
the following lemma then follows mostly from @xcite .
[ lem:7.1 ] for any @xmath113 and @xmath165 , * for @xmath196 there holds @xmath197 * for @xmath198 we have @xmath199 * furthermore , when @xmath200 @xmath201 and when @xmath202 and @xmath203 , @xmath204 * when @xmath71 and @xmath205 , we have @xmath206 * when @xmath207 there exists @xmath208 such that @xmath209 the existence of @xmath210 satisfying follows from proposition [ pro : dixon ] .
thanks to proposition [ pro:2d3d ] , and because for any @xmath165 , we have @xmath211 the inequalities , and are proved when @xmath203 in @xcite , lemma 7.1 .
we will now check that holds when @xmath59 .
from there holds @xmath212 since @xmath213 we have for all @xmath214 @xmath215 the zero @xmath216 is given the first positive solution of @xmath217 .
it satisfies @xmath218 .
we can thus consider only @xmath219 .
we have from proposition [ pro:2d3d ] , for @xmath220 @xmath221 and using the wronskian identity satisfied by @xmath222 , @xmath223 and the recurrence relations satisfied by bessel functions , we obtain ( see @xcite for details ) that @xmath224 with @xmath225 note that @xmath226 , @xmath222 and @xmath227 have simple analytic formulae .
for example @xmath228 in particular , it is easy to show that @xmath229 is decreasing on @xmath230 .
when @xmath59 , we have the following bound for all @xmath231 .
@xmath232 as a consequence , @xmath233 which concludes the proof of when @xmath59 . for any @xmath220 , and @xmath234
, we have @xmath235 where @xmath236 for @xmath220 .
therefore @xmath237 to proceed , note that @xmath238 is negative and decreasing , and for @xmath239 , @xmath240 thus a taylor expansion shows that for @xmath241 , @xmath242 and @xmath243 , we obtain @xmath244 for all @xmath245 , which implies , and since @xmath246 . inserting in together with the values of @xmath247 and @xmath248 for @xmath249 and @xmath250
, we obtain for @xmath251 @xmath252 for all @xmath253 , it is known that @xcite for @xmath254 , @xmath255 in particular , for all @xmath256 , we have @xmath257 combining , and we obtain .
we may now conclude the proof of theorem [ thm : estim - all ] . for convenience
we write @xmath258 .
formula shows that @xmath259 note that @xmath260 is decreasing ( see e.g. @xcite @xmath261 13.74 ) , therefore for all @xmath262 , @xmath263 estimate in lemma [ lem:7.1 ] shows that when @xmath264 .
@xmath265 inserting this bound in we obtain @xmath266 which is estimate . using instead , we obtain for @xmath203 and @xmath267 , @xmath268 inserting this bound in together with we obtain @xmath269 which proves since @xmath270 .
let us now turn to .
formula shows that @xmath271 when @xmath57 , since @xmath260 is decreasing @xmath272 we define two sets of indices , @xmath273 and @xmath274 . since @xmath275 is decreasing , estimate shows that for all @xmath276 , we have @xmath277 on the other hand , @xmath278 , therefore when @xmath279 , @xmath280 , and it is known ( see @xcite @xmath261 13.74 ) that when @xmath281 , @xmath282 is an increasing function of @xmath175 with limit @xmath283 .
furthermore , it is also known ( see @xcite @xmath261 15.3 ) that for all @xmath165 @xmath284 therefore , since @xmath285 , we have @xmath286 on the other hand , @xmath287 as it is known ( see @xcite ) that for all @xmath165 @xmath288 this shows that @xmath289 we have obtained that , for @xmath290 , @xmath291 inserting this bound in , we obtain .
to prove proposition [ pro : lowerbounds ] , we shall use the following intermediate result , related to bessel functions .
[ lem : boundhn ] for any @xmath165 , there holds @xmath292 for any @xmath293 there holds @xmath294 for any @xmath66 , @xmath203 , and @xmath295 , there holds @xmath296 we prove this lemma below .
we can now conclude the proof of proposition [ pro : lowerbounds ] .
let us start with the case @xmath64 .
starting as before from formula , we have @xmath297 using now the bounds in lemma [ lem : boundhn ] and in lemma [ lem:7.1 ] , we have @xmath298 therefore @xmath299 as claimed .
we now turn to the case @xmath66 .
note that for all @xmath300 , using the upper bound given in we see that @xmath301 .
therefore , choosing for each @xmath10 the frequency @xmath302 given by lemma [ lem:7.1 ] , we have @xmath303 the conclusion follows from estimate .
to estimate the maximum of @xmath304 , we proceed as follows .
note that the maximal value occurs at the first positive solution of @xmath305 , which we will denote @xmath306 .
we compute that @xmath307 it is known ( see e.g. @xcite proposition a.1 ) that for @xmath308 and @xmath309 , @xmath310 therefore @xmath311 .
this implies that @xmath312 next , note that from , and the bound @xmath313 we have @xmath314 on the other hand , @xmath315 is an increasing function ( see @xcite @xmath261 15.8 ) , therefore @xmath316 it is well known ( see e.g. @xcite ) that for all @xmath164 , @xmath317 .
it is also known ( see @xcite @xmath261 15.8 ) that @xmath318 is a decreasing function .
note if @xmath293 then @xmath319
. therefore @xmath320 combining these two bounds we obtain that @xmath321 together with this shows that @xmath322 let us now turn to .
it is known ( see @xcite ) that @xmath323 where @xmath324 is a universal constant , @xmath325 .
let us first assume that @xmath326 , where @xmath327 is given by @xmath328 then , using we find @xmath329 it is known ( see @xcite ) that for all @xmath308 , @xmath330 therefore @xmath331 where @xmath332 since @xmath333 is increasing for @xmath334 , we have @xmath335 using the definition of @xmath327 , we see that the right hand side of this last inequality is an explicit function of @xmath336 , which is negative when @xmath337 .
using , we see that for @xmath338 for all @xmath339 .
thus for all @xmath339 , there holds @xmath340 next , we note that @xmath341 is minimal on @xmath342 at @xmath343 for all @xmath344 , where it equals @xmath345 .
therefore @xmath346 and , using that @xmath347 we obtain @xmath348 let us now suppose that @xmath349 . since @xmath350 is decreasing when @xmath351 , we obtain an upper bound on @xmath352 by replacing @xmath353 by its lower bound and @xmath354 by its upper bound , given in .
the resulting expression is an explicit increasing function of @xmath129 , with limit @xmath355 .
it is then possible to verify by inspection on a finite range for @xmath167 that for all @xmath356 , @xmath357 this inequality being automatically satisfied when @xmath358 for example .
the maximum occurs near @xmath359 . on
the other hand , @xmath360 and @xmath361 is decreasing , and see e.g. @xcite @xmath362 therefore when @xmath363 . @xmath364 the bounds and show that when @xmath349 , we have @xmath365 combining and we have obtained that holds for all @xmath366 , all @xmath203 and all @xmath367 . to conclude , note that using we have @xmath368 the proof of theorem [ thm : broadband ] follows the line of the proof of the corresponding result in the two - dimensional case proved in @xcite .
the first step is the following proposition .
[ pro : ito ] for any @xmath369 and @xmath370 , we define @xmath371 and @xmath372 then @xmath373 where @xmath374 is defined by @xmath375 and where @xmath376 is the set of all positive @xmath129 such that @xmath377 . furthermore , @xmath378 when @xmath59 , the same result holds for @xmath379 .
first , note that lemma [ lem:7.1 ] shows that @xmath380 .
furthermore , we have shown in the proof of theorem [ thm : estim - all ] that when @xmath381 , @xmath382 , thus @xmath383 .
thirdly , using the bound @xmath285 , we see that when @xmath384 , we have @xmath385 on @xmath384 , @xmath386 is decreasing , and @xmath387 is a decreasing function of @xmath167 .
therefore @xmath388 and we have obtained that @xmath389 next , it is known ( see @xcite , proposition 8.3 ) that when @xmath390 and @xmath391 for some @xmath392 , then @xmath393 when @xmath203 .
the argument is simple .
it turn out that by a simple calculus argument using the formula for @xmath135 , when @xmath394 , @xmath395 , and @xmath396 , @xmath397 , then @xmath393 . when @xmath398 , @xmath394 , @xmath395 , and @xmath396 therefore the inclusion holds .
since @xmath399 , @xmath400 , and @xmath401 on @xmath402 , the same is true when @xmath59 .
the proof of proposition [ pro : ito ] will be complete once estimate is established , for @xmath403 .
since it is proved in @xcite proposition 8.2 for @xmath203 , we only need to consider the case @xmath59 , and @xmath404 .
we have @xmath405 and @xmath406 . introducing @xmath407{rcl } ( 0,\frac{\pi}{2})\setminus \cup_k \{k\pi/\lambda\ } & \to & \mathbb{r } \\ x & \to & \displaystyle \frac{g_{\frac{1}{2 } } \left(\lambda x\right)}{k_{\frac{1}{2 } } ( x ) } , \end{array}\ ] ] we have @xmath408.\ ] ] we first verify that @xmath409 is one - to - one on @xmath410 , for @xmath411 small enough and @xmath46 large enough .
differentiating @xmath412 we find @xmath413 when @xmath414 , we have @xmath415 therefore @xmath416 when @xmath417 and @xmath418 , we have @xmath419 , and @xmath420 since @xmath421 . finally , @xmath422 when @xmath403 , thus for any @xmath423 , we have obtained that @xmath424 for all @xmath418 . in particular , if @xmath425 $ ] , we have @xmath426 therefore @xmath427 since @xmath428 is increasing on @xmath429 , and @xmath430 , @xmath431 where @xmath432 is a convenient choice ( but any number smaller than @xmath216 and greater than zero would do to write the riemann sum ) .
the distance between two distinct positive solutions of @xmath217 is at least @xmath433 , therefore @xmath434 since @xmath403 , and @xmath435 we have obtained that @xmath436 and this concludes our proof .
the second step is to use proposition [ pro : ito ] to derive an estimate for the scattered field , for large enough contrast .
[ lem : highcontrast - lemma ] suppose @xmath370 .
let @xmath437 be the following decreasing function of the contrast @xmath438 given @xmath439 , for any @xmath440 such that @xmath441 there exists a set @xmath117 depending on @xmath442 and @xmath46 such that @xmath443 and , for any @xmath262 @xmath444 recall that we established in , that for all @xmath48 we have @xmath445 for @xmath446 , and @xmath447 , let @xmath448 be given by @xmath449 if @xmath450 , we have thanks to proposition [ pro : ito ] @xmath451 from proposition [ pro : ito ] , we also know that @xmath452 to conclude , note that the set of excluded frequencies for @xmath453 is @xmath454 . we can now conclude the proof of theorem [ thm : broadband ] .
when @xmath455 , theorem [ thm : estim - all ] implies theorem [ thm : broadband ] , with @xmath122 .
when @xmath456 .
theorem [ thm : estim - all ] shows that for all @xmath457 , we have @xmath458 so we can again select @xmath122 .
suppose now @xmath459 , with @xmath460 .
then @xmath370 , and we can apply lemma [ lem : highcontrast - lemma ] .
choosing @xmath461 we have for all @xmath446 , @xmath462 and there exists a set @xmath117 depending on @xmath463 and @xmath0 such that for any @xmath262 , @xmath464 the size of the set @xmath117 is bounded by @xmath465 since @xmath466 is decreasing when @xmath467 .
this work was completed in part while george leadbetter and andrew parker were visiting oxpde during a summer undergraduate research internship awarded by oxpde in 2011 , and they would like to thank the centre for the wonderful time they had there . c. mller , _ foundations of the mathematical theory of electromagnetic waves _ , revised and enlarged translation from the german .
die grundlehren der mathematischen wissenschaften , band 155 , springer - verlag , new york , 1969 .
mr 0253638 ( 40 # 6852 ) h .-
nguyen and m. s. vogelius , _ a representation formula for the voltage perturbations caused by diametrically small conductivity inhomogeneities .
proof of uniform validity _ , ann .
h. poincar anal .
non linaire * 26 * ( 2009 ) , no . 6 , 22832315 .
mr 2569895 ( 2011f:78003 ) f. w. j. olver , d. w. lozier , r. f. boisvert , and c. w. clark ( eds . ) , _ nist handbook of mathematical functions _
department of commerce national institute of standards and technology , washington , dc , 2010 , with 1 cd - rom ( windows , macintosh and unix ) . mr 2723248 | we consider the solution of a scalar helmholtz equation where the potential ( or index ) takes two positive values , one inside a ball of radius @xmath0 and another one outside . in this paper , we report that the results recently obtained in the two dimensional case in @xcite can be easily extended to three dimensions . in particular , we provide sharp estimates of the size of the scattered field caused by this ball inhomogeneity , for any frequencies and any contrast .
we also provide a broadband estimate , that is , a uniform bound for the scattered field for any contrast , and any frequencies outside of a set which tends to zero with @xmath0 . |
nitrogen - vacancy ( nv ) centers found in diamond are a promising system for realizing optically addressable solid - state spin qubits . in single - crystal diamond ,
negatively charged nv@xmath4 centers have been used in a number of experimental demonstrations relevant to quantum information processing , including single photon generation @xcite , coherent population trapping @xcite , and optical readout and manipulation of single nuclear spins @xcite . in order to utilize these properties in quantum information processing applications @xcite , efficient and scalable optical coupling between nvs and photonic devices such as waveguides and microcavities
@xcite is necessary .
photonic crystal nanocavities @xcite confine photons to sub - wavelength mode volumes , @xmath5 , enabling purcell enhanced coupling between an optical dipole , such as an nv - center or quantum dot , and the cavity mode @xcite .
the coherent dipole cavity coupling rate scales as @xmath6 , and small @xmath5 is particularly beneficial in cavity - qed ( quantum electrodynamic ) systems where intrinsic optical loss or dipole dephasing rates are large .
recently , optical coupling between high-@xmath0 microcavities and nvs hosted in diamond nanocrystals has been reported @xcite , and devices for studying nvs hosted in nanocrystalline diamond films have been fabricated @xcite .
however , to - date many desirable properties of nvs in single - crystal diamond have not yet been observed in nanocrystalline diamond .
although spontaneous lifetime - limited optical transition linewidths have been measured in single nvs hosted in diamond nanocrystals , they exhibit @xmath7 thz inhomogeneous broadening and have not been measured in high yield from an ensemble of candidate nanocrystals @xcite . in bulk single crystal diamond ,
nv@xmath4 optical transitions with lifetime limited linewidths @xcite , inhomogeneous broadening of less than 10 ghz , and time - averaged spectral diffusion below @xmath8 mhz @xcite have been observed . because of the superior properties of nvs in bulk single crystal diamond ,
some cavity geometries using only single - crystal diamond and air have been proposed @xcite . however , there are difficulties in fabricating photonic devices from single - crystal diamond , since vertical optical confinement within the diamond requires either a three dimensional etching process , or a method for fabricating thin single - crystal diamond films @xcite .
an alternative to fabricating devices directly from single - crystal diamond is to integrate a patterned high - index optical waveguiding layer on top of a diamond substrate , from which photons can couple evanescently to nvs close to the diamond surface .
recently , a hybrid gap - diamond material system , consisting of an optically thin gap film attached to a diamond substrate , was used to demonstrate optical coupling between nvs close to the diamond surface and ridge waveguides patterned in the gap film @xcite .
owing to the enhanced local density of states in the near field of the gap waveguide modes , relatively efficient evanescent coupling between nvs and the gap waveguide can be achieved . in order to efficiently couple to individual nvs ,
three dimensional confinement provided by an optical cavity is necessary .
here we present designs for a gap - diamond photonic crystal nanowire waveguide and nanocavity which can be fabricated using available materials and processing techniques . using numerical simulations , we study the sensitivity of the optical properties of these devices on structural parameters .
we also simulate the coupling of spontaneous emission from a broadband source , such as phonon sideband emission from an nv - center , into the proposed waveguide design , and show that efficient broadband collection of nv emission is possible with these structures .
photonic crystal nanocavities , formed by introducing localized perturbations to planar periodic structures , can support ultra - high @xmath9 resonances .
ultra - high @xmath9 devices have been demonstrated at near - ir wavelengths in free standing membranes such as si with @xmath10 @xcite . in si films supported by low - index sio@xmath11 substrates ( soi ) , devices with @xmath12
have been realized @xcite .
although simulations of photonic crystal cavities fabricated from diamond membranes ( @xmath13 ) predict radiation loss limited resonances with @xmath14 @xcite , material loss has limited @xmath15 @xcite in nanocavities fabricated from nanocrystalline membranes .
photonic crystal nanocavities fabricated from single crystal diamond membranes @xcite have not yet been demonstrated . here
we study photonic crystal cavities formed in a thin waveguiding layer supported by a diamond substrate hosting high quality nvs near the diamond surface , as indicated in fig .
[ fig : pc_schematic ] . in choosing a film from which to form the waveguide layer ,
we are limited to materials which are transparent at the nv@xmath4 zero phonon transition wavelength , @xmath16 nm , and whose refractive index exceeds @xmath17 .
epitaxially grown gap films have a nominally high refractive index ( @xmath18 and low optical absorption at @xmath19 .
photonic crystal cavities formed in gap membranes have recently been demonstrated with @xmath20 @xcite , and in ref .
@xcite , optical loss of @xmath21 db / cm was measured in a gap waveguide supported by a single crystal diamond substrate .
the waveguide loss measurements presented in ref .
@xcite indicate that devices with absorption and surface scattering limited @xmath22 can be realized in this system . with improved processing to reduce scattering loss ,
it is expected that material loss limited @xmath0 exceeding this value should be possible . realizing a high-@xmath0 photonic crystal nanocavity in this gap - diamond system is complicated by the diamond substrate s moderately high refractive index ( @xmath23 ) .
compared to photonic crystal membrane or soi devices , a diamond substrate expands the light cone into which photons can radiate out of the nanocavity .
in addition , the broken vertical mirror symmetry from the diamond substrate precludes the existence of a bound mode in the gap waveguide layer for arbitrarily low frequencies . as discussed below , the underlying gap - diamond nanowire waveguide structure from which the nanocavity studied here is realized does not support a non - radiating mode in the wavelength range of interest .
this is a result of the subwavelength dimensions of the waveguide cross - section , which provides strong modal confinement at the expense of reducing the effective index of the waveguide modes below @xmath24 .
however , by extending the waveguide sidewalls into the diamond substrate , radiation into the diamond substrate can be made arbitrarily small .
as we will show below , for a realistic diamond sidewall height , the proposed structure supports waveguide modes whose radiation loss is smaller than the expected intrinsic material or scattering loss , and forms a low - loss structure for realizing a high-@xmath0 nanocavity . the underlying structure of the photonic crystal nanocavity illustrated in fig .
[ fig : pc_schematic ] consists of a gap photonic crystal nanowire waveguide whose sidewalls have been extended into a diamond substrate .
this waveguide supports bloch modes which form the basis for localized resonances formed when the periodic symmetry of the waveguide is broken . in general , the bloch modes are either leaky or guided , depending on their frequency , @xmath25 , and wavenumber , @xmath26 , relative to the air and diamond lightlines @xcite .
waveguide modes with an effective index smaller than that of diamond , @xmath27 ( setting @xmath28 ) , will leak into the diamond substrate .
high-@xmath0 cavity resonances are realized by engineering a perturbation which couples guided or low - loss bloch modes @xcite .
using 3-dimensional finite difference time domain ( fdtd ) simulations @xcite to study the photonic crystal waveguide loss , we can predict an upper limit on the @xmath0 of a nanocavity formed from this structure .
figure [ fig : pc_wg_band](a ) shows the first brillouin zone of the band - structure for the lowest frequency gap - diamond photonic crystal nanowire waveguide modes .
also shown are the air , diamond , and gap lightlines .
the calculation used waveguide cross - section dimension @xmath29 = \left[192,128\right]$ ] nm , hole spacing @xmath30 nm , and hole radius @xmath31 nm . the hole spacing was chosen to approximately position the lowest energy waveguide mode band edge in the wavelength range of interest .
the waveguide cross - section was not systematically optimized , and when normalized by the hole spacing @xmath32 , is similar to that used in previous nanowire based cavity designs @xcite .
@xmath33 was chosen based upon initial simulations of the @xmath0 of the nanocavity design studied below in sec .
[ sec : defect ] , and is also similar to values used in other work .
a mesh with resolution @xmath34 , perfectly matched layer ( pml ) boundary conditions in the @xmath35 and @xmath36 directions ( @xmath37 nm pml layer thickness ) , and bloch boundary conditions in the @xmath38 direction were used in these simulations . for clarity , only modes with even parity in the @xmath35 direction are calculated .
odd modes lie higher in frequency , and as will become clear below , we are primarily interested in the lowest frequency waveguide mode .
note that parity in the @xmath36 dimension is not conserved for these gap - diamond structures , since they are not vertically symmetric .
both the lowest frequency te - like ( dominantly @xmath39 polarized ) and tm - like ( dominantly @xmath40 polarized ) modes are shown in fig .
[ fig : pc_wg_band ] .
we label these modes te-1 and tm-1 , respectively . also evident in fig .
[ fig : pc_wg_band](b ) is a mode ( black open points ) whose dispersion lies between the te-1 and tm-1 bands .
this mode corresponds to a leaky mode whose field is predominantly confined within the diamond substrate .
leaky modes with waveguide quality factor , @xmath41 are not shown .
@xmath42 is related to the waveguide loss per unit length , @xmath43 , by @xmath44 , where @xmath45 is the group velocity of the waveguide mode . from fig .
[ fig : pc_wg_band](a ) , it is immediately clear that for the specified waveguide dimensions , no waveguide modes lie below the diamond lightline . in particular , at the band edge ( @xmath46 ) the effective index of the lowest frequency waveguide mode is smaller than @xmath17 ; this mode is expected to leak into the diamond substrate .
this leakage could be reduced by increasing the gap waveguide dimensions ( @xmath47 ) , increasing the effective index of the mode at the expense of decreasing the peak single photon field strength , and decreasing the frequency and momentum mismatch between the te-1 and higher order waveguide modes .
alternatively , we can extend the gap sidewalls ( including the holes ) by depth @xmath48 into the diamond substrate . in the limit that @xmath49 , the bulk diamond lightline is no longer relevant in determining whether a gap - diamond waveguide mode is bound or leaky .
instead , the relevant substrate is the approximately two dimensional vertical diamond photonic crystal slab underneath the gap photonic crystal waveguide . the band - structure of the lowest frequency @xmath39 polarized mode supported by this vertical slab
is shown in fig .
[ fig : pc_wg_band](b ) . above this band exists a continuum of radiating slab modes with non - zero vertical momentum ; this band forms a renormalized `` structured lightline '' . at @xmath50 , the te-1 mode lies below this structured lightline , indicating that this mode is not leaky in the @xmath49 limit .
the existence of this mode below the structured lightline is not sensitive to the choice of @xmath51 , since decreasing @xmath51 decreases the effective index of both the te-1 mode and the lowest energy radiating slab mode . decreasing @xmath52 decreases the te-1 effective index in a manner typical of slab and ridge waveguides @xcite , and can push the te-1 mode above the structured lightline at a given frequency . in fig .
[ fig : pc_wg_q ] the waveguide loss per optical cycle , @xmath53 , of the te-1 mode is shown as a function of the etch depth @xmath48 , at @xmath54 , for @xmath55 nm , where @xmath56 is the `` cavity '' hole spacing of the nanocavity design presented in the following section and indicated in fig .
[ fig : pc_schematic ] .
this was calculated using fdtd simulations , by monitoring the power radiated into the absorbing @xmath35 and @xmath36 boundaries of a waveguide bloch unit cell .
we expect @xmath42 at @xmath57 to place an upper limit on the @xmath0 of a cavity mode formed predominantly from the te-1 mode @xcite .
figure [ fig : pc_wg_q ] indicates that for @xmath58 nm , the te-1 waveguide mode is very leaky with @xmath59 , but that for increased etch depth @xmath60 nm , the loss can be reduced and @xmath61 . for @xmath62 nm , @xmath42 increases exponentially with @xmath48 .
the nanocavity studied here is formed in the photonic crystal nanowire waveguide described in sec .
[ sec : waveguide ] by locally perturbing the hole spacing , @xmath63 , as indicated in figs .
[ fig : pc_heterostructure](a ) and [ fig : pc_cavity_field](a ) , forming a heterostructure @xcite cavity @xcite .
the hole spacing is varied slowly over 6 periods from @xmath64 nm in the center of the cavity , to a constant `` bulk '' value of @xmath65 nm . the cavity is symmetric about the @xmath66 and @xmath67 axes . for simplicity , the hole radius remains constant .
this type of photonic crystal nanowire waveguide cavity @xcite supports localized cavity modes with simulated radiation loss limited @xmath68 when applied to structures with air undercladding ( waveguide - substrate index contrast @xmath69 ) @xcite . in soi material systems ( @xmath70 ) , nanowire based cavities with @xmath71 have been theoretically studied @xcite .
ultrahigh @xmath9 nanocavities formed in photonic crystal nanowire waveguides on a high - index substrate such as diamond have not been previously investigated .
adjusting the local hole spacing within the photonic crystal waveguide shifts the frequency range of the photonic crystal waveguide stop - band . in a quasi-1d picture
, the photonic crystal waveguide forms an `` optical potential '' @xcite whose band edge is locally modulated by the variation in the hole spacing or hole size @xcite . in the cavity design considered here , the local reduction in hole spacing shifts the valence and conduction band edges up in frequency so that the cavity supports an `` acceptor '' defect mode @xcite formed by superpositions of the valence band states of the unperturbed photonic crystal nanowire waveguide @xcite . following an analysis similar to that in ref .
@xcite , the photonic crystal cavity optical potential is illustrated by fig.[fig : pc_heterostructure](b ) , which shows the frequency @xmath72 of the te-1 valence band edge as a function of the `` local '' hole spacing , @xmath63 , at positions @xmath73 midway between two holes in the cavity .
the band edge frequency @xmath72 was calculated at each @xmath73 by using fdtd simulations to determine the band structure of the nanowire photonic crystal waveguide in sec.[sec : waveguide ] , with the hole spacing set to @xmath74 .
also shown are the frequencies of the te-1 valence and conduction band edges in the `` bulk '' waveguide region , where @xmath75 , and of a high-@xmath0 localized cavity mode discussed in detail below .
note that these photonic crystal nanowire cavities have also been analyzed in the context of the unperturbed photonic crystal waveguide sections forming mirrors for photons trapped in the central cavity region @xcite .
fdtd simulations of this structure predict that it supports a mode at a resonance wavelength @xmath76 close to the nv@xmath4 zero phonon line at 637 nm ( @xmath77 ) , with @xmath78 and mode volume @xmath79 , where @xmath5 is defined by the peak electric field energy density : @xmath80 .
the maximum field amplitude inside the diamond is @xmath81 , where @xmath82 is the field maximum at the location of maximum energy density .
@xmath82 is located inside the high index gap layer for the device and mode studied here .
the field maximum decays to @xmath83 at a depth @xmath84 nm below the diamond surface .
simulations indicate that similar designs with the number of grading periods reduced to 4 support more tightly confined modes ( @xmath85 ) with a maximum @xmath86 .
the mesh used in these simulations had uniform resolution @xmath87 , and extended @xmath88 periods ( @xmath89 ) in @xmath90 beyond the end of the graded region , @xmath91 in @xmath92 above and below the diamond surface , and @xmath93 along the @xmath94 directions .
a pml layer thickness of @xmath95 was used in the highest-@xmath0 simulations . after exciting the structure with a narrow band source , the steady state ratio of stored and radiated power
was used to calculate the modal @xmath0 . the radiated power along each axis
was monitored using reference frames positioned @xmath96 away from each pml boundary .
mirror symmetry was enforced about the @xmath38 ( odd parity ) and @xmath35 ( even parity ) axes .
higher resolution ( @xmath97 ) simulations of the @xmath98 nm structures were also conducted , and it was verified that the structure supports a mode with @xmath99 .
the role of the diamond substrate in limiting @xmath0 is illustrated by simulating its dependence on the depth of the etched diamond ridge .
figure [ fig : pc_cav_q_vs_p](a ) shows @xmath0 as a function of @xmath48 , and indicates the relative contributions to @xmath0 due to radiation into the bottom diamond substrate ( @xmath100 ) , top air cladding ( @xmath101 ) , end of the cavity ( @xmath102 ) and side of the cavity ( @xmath103 ) . the total cavity @xmath0 is given by @xmath104 .
when @xmath48 is reduced to @xmath105 nm , increased radiation loss into the substrate degrades @xmath106 . for @xmath107 nm , radiation loss through the end of the cavity into
the photonic crystal waveguide begins to play a non - negligible role in limiting @xmath0 .
figure [ fig : pc_cav_q_vs_p](b ) shows the dependence of @xmath0 on the number @xmath108 of waveguide periods between the cavity grading edge and the pml absorbing layer used in the simulation . for @xmath109
, @xmath102 is approximately constant , indicating that it is limited by coupling between the cavity mode and either photonic crystal waveguide modes which do not exhibit a bandgap at @xmath76 ( e.g. , the tm-1 mode ) or radiation modes @xcite . for @xmath110 ,
the cavity loss is dominated by `` mirror '' leakage . in a practical device
, @xmath108 represents the number of `` mirror '' periods between the edge of the cavity grading and an integrated waveguide , and would be set to a value that ensures that radiation loss from of the cavity is predominantly into the waveguide .
the sensitivity of @xmath0 to the central cavity hole spacing , @xmath56 , is shown in fig .
[ fig : pc_cav_q_vs_p](c ) .
when @xmath56 is varied between @xmath111 nm , @xmath0 remains @xmath112 . in these simulations ,
the bulk hole spacing is maintained fixed at @xmath113 nm , and the grading of @xmath56 remains parabolic .
@xmath0 and @xmath102 are maximized when @xmath64 nm , while @xmath100 trends upward as @xmath114 . in this limit , the depth of the parabolic cavity defect becomes increasingly small , minimizing coupling to lossy waveguide modes close to the diamond lightline . as a result , @xmath100 increases .
@xmath102 is maximized when a sufficiently small @xmath56 is chosen to position @xmath76 well within the te-1 bandgap , as illustrated in fig .
[ fig : pc_heterostructure](b ) .
however , pushing @xmath76 too far above the te-1 valence band edge can result in efficient phase matched coupling between the cavity mode and the lossy tm-1 valence band edge modes ( see fig . [
fig : pc_wg_band ] ) , reducing @xmath102 .
these competing dependencies determine the optimum @xmath56 .
note that @xmath5 varies between @xmath115 @xmath116 over the simulated range of @xmath56 . from the fdtd calculations of @xmath5 and @xmath0 presented above
, we can predict the coherent coupling rate , @xmath117 , for a single nv@xmath4 center located near the cavity and coupled to a single photon stored in the nanocavity mode . for experiments in cavity qed we must be careful to distinguish between zero - phonon optical transitions ( these become spectrally narrow at low temperature ) and phonon - assisted transitions .
the total spontaneous emission rate of an nv@xmath4 center is measured to be @xmath118 mhz ; however the rate into the zpl alone , @xmath119 , is only @xmath120 of @xmath121 @xcite .
the coherent coupling rate between a single photon and the nv@xmath122 zpl is given by @xmath123 where @xmath124 , @xmath125 is the magnitude of the electric field at the nv location , and @xmath82 is the field strength at the electric field energy density maximum , which is located within the gap for the device studied here .
this expression assumes that the spatial orientation of the nv transition dipole moment and electric field polarization are parallel , which is possible for one of four allowed nv orientations in @xmath126 diamond . for the cavity mode volume given in sec .
[ sec : defect ] , eq .
( [ eqn : g ] ) gives @xmath127 ghz for an nv optimally located at the diamond surface .
assuming that the cavity @xmath0 is limited by radiation loss as calculated in sec .
[ sec : defect ] , we see that @xmath128/2\pi = \left[2.25,0.16,0.013\right]$ ] ghz , where @xmath129 , indicating that it is possible to reach the strong coupling regime @xcite using this nanocavity design . also of interest is the purcell enhanced spontaneous emission rate of the nv@xmath4 zpl into the cavity mode . assuming the cavity and nv@xmath4 zpl
are aligned spectrally , the purcell enhancement factor to the zpl spontaneous emission rate is given by , @xmath130 for an nv optimally positioned at the diamond surface , eq .
( [ eqn : fp ] ) gives @xmath131 . in a realistic experiment , it may not be possible to couple to nvs with desirable optical and spin coherence properties arbitrarily close to the diamond surface @xcite .
in addition , material loss and fabrication imperfections will likely limit the nanocavity @xmath0 below the highest values presented in sec .
[ sec : defect ] .
if we assume that the nv is optimally positioned @xmath132 nm below the diamond surface , and that the cavity @xmath0 is limited to @xmath133 ( e.g. , due to material absorption and surface scattering @xcite ) , we find @xmath134 , indicating that @xmath135 of the _ total _ nv@xmath4 spontaneous emission will radiate into the nanocavity mode .
the @xmath136 of uncoupled nv emission will radiate into phonon sidebands not resonant with the cavity mode .
in addition to forming the basic structure for the the nanocavity presented in sec . [ sec : defect ] , the waveguide studied in sec.[sec : waveguide ] can function as an efficient broadband collector of radiation from nv phonon sidebands .
measurements of broadband phonon - sideband emission are used in experiments to gather information regarding the relative populations of the nv spin triplet ground states @xcite , as well as the nuclear spin state of neighboring impurities @xcite . from coupled - mode theory
@xcite , the normalized power spectral density , @xmath137 , coupled into a waveguide mode from a nv dipole located at position @xmath138 in the photonic crystal waveguide near field is approximately given by @xmath139 where @xmath140 is the frequency dependent group index of the waveguide mode , @xmath141 is the waveguide mode field amplitude , and @xmath142 is the waveguide mode area defined by @xmath143 where volume @xmath144 is a unit cell of the waveguide .
we have assumed that the dipole and waveguide field at @xmath138 are parallel .
@xmath145 is normalized by the `` bulk '' power spectral density of the nv dipole embedded far below an unpatterned diamond surface .
equation ( [ eqn : a_wg ] ) indicates that coupling into a given waveguide mode can be increased by maximizing @xmath146 .
photonic crystal waveguides are ideally suited in this regard @xcite : they can support modes with sub - wavelength modal area and large @xmath147 . to gain a quantitative estimate of the nv - waveguide coupling , and to study enhancement or suppression of nv coupling into other radiation modes of the waveguide structure , it is useful to conduct fdtd simulations of the emission spectrum of a broadband dipole source positioned at @xmath138 .
here we consider a broadband dipole , polarized along @xmath35 and located @xmath148 nm below the diamond surface and midway between two holes of the photonic crystal waveguide studied in sec .
[ sec : waveguide ] . in order to channel the dipole emission into a forward propagating waveguide mode , the hole spacing of the waveguide structure used here is asymmetric about the dipole position , as shown in fig .
[ fig : spectrum](a ) . for @xmath149 ,
the valence band edge of this structure is lower on the left side of the dipole than on the right side ; the band edge of the left ( `` mirror '' ) region has a lower frequency than the band edge of the right ( `` waveguide '' ) region . for efficient collection of nv emission ,
the structure should be designed such that the majority of the nv phonon sideband emission is lower in frequency than the `` waveguide '' te-1 valence band edge , and higher in frequency than the `` mirror '' te-1 valence band edge .
note that reflections from the waveguide termination can dramatically alter the coupling between a dipole and the waveguide , as studied in @xcite .
here we have chosen a termination designed to suppress these reflections in the simulations . in fdtd simulations of periodic structures supporting propagating modes , pml terminations of the propagating axis
can create significant reflections .
a waveguide which is invariant along the propagating dimension , such as the termination region of the structure considered here , is better suited for a pml boundary .
figure [ fig : spectrum](b ) shows the fdtd calculated spectrum for the system described above .
power spectral density of radiation into the waveguide , and into all other channels ( e.g. , the substrate ) are shown . in the simulations studied here ,
@xmath64 nm and @xmath113 nm . the hole radius and waveguide cross - section are the same as in sec . [
sec : waveguide ] , and @xmath150 .
the spectra are normalized by the spectrum of the identical dipole source when it is positioned at depth @xmath151 below an unpatterned diamond surface . from fig.[fig : spectrum](b ) , we see that for frequencies below the valence band edge of the waveguide region ( @xmath152 ) , the emission into the waveguide exceeds the total `` bulk '' emission in absence of the waveguide .
the radiated waveguide power was calculated by monitoring the power flux through an area overlapping the waveguide cross - section , with dimensions @xmath153 , located near the simulation pml boundary .
note that emission into non - waveguide modes is also enhanced . as shown in fig .
[ fig : spectrum](c ) , radiation into the substrate ( @xmath154 direction ) and into the waveguide ( @xmath155 direction ) are the dominant radiation channels .
also , note that relatively little power is coupled into the backward propagating waveguide mode ( @xmath156 direction ) in the displayed frequency range . for @xmath113 nm ,
the bandwidth of this efficient waveguide coupling is approximately @xmath157 nm , allowing efficient collection of a large portion of the nv@xmath4 zero phonon line and phonon sideband emission
. this may be particularly useful in room - temperature nv experiments which do not rely upon readout of a sharp optical zpl , but require efficient collection of sideband emission , for electron spin readout , for example .
fabrication of the devices studied in this paper requires three key processing abilities : ( i ) patterning gap or another high index film with the design presented above , ( ii ) attaching high index films to a diamond substrate , and ( iii ) transferring the thin film pattern into the diamond substrate . proofs of principle of these individual steps have been demonstrated .
results presented in refs .
@xcite show that it is possible to lithographically define and etch sufficiently small holes using current electron beamwriting and plasma etching tools .
large area patterned gap films were transferred and adhered @xcite onto a diamond sample in ref.@xcite , as have smaller micron - scale structures @xcite . transferring the pattern into the diamond
can be accomplished using oxygen based plasma etching , e.g. , in ref .
@xcite a highly selective plasma etch was used to extend a sin pattern into a diamond substrate . the degree
to which the precise design studied in sec . [
sec : defect ] can be replicated will affect the maximum achievable @xmath0 , and achieving the high aspect ratio diamond sidewalls of the @xmath98 nm structure may be challenging .
theoretical studies analyzing the importance of the sidewall slope may be necessary in future work .
finally , note that although precise ( 2 nm ) tuning of @xmath56 was considered in sec .
[ sec : defect ] in order to maximize @xmath0 , the cavity supports modes with @xmath158 over the range 132 nm @xmath159 150 nm .
the gap - based cavity proposed here is designed specifically for coupling to negatively charged nv centers positioned within 50 nm of a diamond surface .
it is well established that a dense layer of nv centers close to a diamond surface can be created in high - nitrogen diamond by ion implantation and annealing @xcite .
such a dense layer is well - suited for initial testing of the optical characteristics of gap - based microcavities .
however , for the proposed devices to be useful for quantum information applications , single nv centers with good spectral characteristics must be fabricated close to a surface in high - purity diamond .
initial tests suggest that in high - purity diamond , a nearby surface can have a large effect on nv properties @xcite that include charge stability and optical linewidth .
a thorough study of how to optimize the properties of single nv centers close to a surface will be required for any device based on nv centers coupled to an optical microcavity , including the gap - based design presented here .
in this paper we have analyzed a photonic crystal geometry which supports low - loss , subwavelength nanocavities and waveguides suitable for coupling to diamond nv - centers in single crystal diamond .
the structures presented here can be fabricated using existing materials and processing techniques , and promise to allow efficient collection of photons emitted from nv - centers , a crucial requirement for proposed applications @xcite .
efficient collection of nv zpl emission is necessary for single photon and quantum repeater applications of nvs , and efficient collection of nv sidebands is required for high bandwidth readout of the nv electron spin .
the photonic crystal nanocavity design presented above supports modes with the necessary @xmath0 and @xmath5 to reach the strong coupling or large purcell factor regimes with the zpl of an nv - center .
the photonic crystal waveguide analyzed here allows broadband collection of nv emission , and is promising for efficient collection of nv sidebands . in future work
, it is expected that integration of multiple cavities , and realization of nanophotonic circuits for quantum information processing using nv - centers , will be possible using the device designs presented here . | a design for an ultra - high @xmath0 photonic crystal nanocavity engineered to interact with nitrogen - vacancy ( nv ) centers located near the surface of a single crystal diamond sample is presented .
the structure is based upon a nanowire photonic crystal geometry , and consists of a patterned high refractive index thin film , such as gallium phosphide ( gap ) , supported by a diamond substrate .
the nanocavity supports a mode with quality factor @xmath1 and mode volume @xmath2 , and promises to allow purcell enhanced collection of spontaneous emission from an nv located more than 50 nm below the diamond surface .
the nanowire photonic crystal waveguide can be used to efficiently couple light into and out of the cavity , or as an efficient broadband collector of nv phonon sideband emission .
the proposed structures can be fabricated using existing materials and processing techniques .
10 url # 1#1urlprefix a. gruber , a. drbenstedt , c. tietz , l. fleury , j. wratchtrup , and c. v. borczyskowski , `` scanning confocal optical microscopy and magnetic resonance on single defect centers , '' science * 276 * , 20122014 ( 1997 ) . c. kurtsiefer , s. mayer , p. zarda , and h. weinfurter , `` stable solid - state source of single photons , '' phys
. rev .
lett . * 85 * , 290 ( 2000 ) .
a. beveratos , r. brouri , t. gacoin , j .-
p .
poizat , and p. grangier , `` nonclassical radiation from diamond nanocrystals , '' phys .
rev .
a * 64 * , 061802 ( 2001 ) . c. santori , p. tamarat , p. neumann , j. wrachtrup , d. fattal , r. g. beausoleil , j. rabeau , p. olivero , a. d. greentree , s. prawer , f. jelezko , and p. hemmer , `` coherent population trapping of single spins in diamond under optical excitation , '' phys .
rev .
lett . * 97 * , 247401 ( 2006 ) .
f. jelezko , t. gaebel , i. popa , a. gruber , and j. wrachtrup , `` observation of coherent oscillations in a single electron spin , '' phys .
rev .
lett .
* 92 * , 076401 ( 2004 )
. m. v. gurudev dutt , l. childress , l. jiang , e. togan , j. maze , f. jelezko , a. s. zibrov , p. r. hemmer , and m. d. lukin , `` quantum register based on individual electronic and nuclear spin qubits in diamond , '' science * 316 * , 13121316 ( 2007 ) .
l. childress , m. v. gurudev dutt , j. m. taylor , a. s. zibrov , f. jelezko , j. wrachtrup , p. r. hemmer , and m. d. lukin , `` coherent dynamics of coupled electron and nuclear spin qubits in diamond , '' science * 314 * , 281285 ( 2006 ) .
l. childress , j. m. taylor , a. s. srensen , , and m. d. lukin , `` fault - tolerant quantum communication based on solid - state photon emitters , '' phys .
rev .
lett . * 96 * , 070504 ( 2006 ) .
s. c. benjamin , b. w. lovett , and j. m. smith , `` prospects for measurement - based quantum computing with solid state spins , '' arxiv:0901.3092v2 [ quant - ph ] ( 2009 ) .
c .- h .
su , a. d. greentree , and l. c. l. hollenberg , `` towards a picosecond transform - limited nitrogen - vacancy based single photon source , '' opt .
express * 16 * , 62406250 ( 2008 ) .
a. young , c. y. hu , l. marseglia , j. p. harrison , j. l. obrien , and j. g. rarity , `` cavity enhanced spin measurement of the ground state spin of an nv center in diamond , '' new j. phys . * 11 * , 013007 ( 2009 ) .
o. painter , r. k. lee , a. yariv , a. scherer , j. d. obrien , p. d. dapkus , and i. kim , `` two - dimensional photonic band - gap defect mode laser , '' science * 284 * , 18191824 ( 1999 ) .
y. akahane , t. asano , b .- s .
song , and s. noda , `` high - q photonic nanocavity in a two - dimensional photonic crystal , '' nature * 425 * , 944947 ( 2003 ) .
k. srinivasan , p. e. barclay , m. borselli , and o. painter , `` optical - fiber based measurement of an ultra - small volume high - q photonic crystal microcavity , '' phys .
rev .
b * 70 * , 081306(r ) ( 2004 ) . b .- s .
song , s. noda , t. asano , and y. akahane , `` ultra - high - q photonic double - heterostructure nanocavity , '' nature materials * 4 * , 207210 ( 2005 ) .
y. takahashi , h. hagino , y. tanaka , b .- s .
song , t. asanoa , , and s. noda , `` high - q nanocavity with a 2-ns photon lifetime , '' opt .
express * 15 * , 1720617213 ( 2007 ) .
t. yoshie , a. scherer , j. hendrickson , g. khitrova , h. gibbs , g. rupper , c. ell , o. shchekin , and d. deppe , `` vacuum rabi splitting with a single quantum dot in a photonic crystal nanocavity , '' nature * 432 * , 200203 ( 2004 ) . k. hennessy , a. badolato , m. winger , d. gerace , m. atature , s. gulde , s. faelt , e. l. hu , and a. imamoglu , `` quantum nature of a strongly coupled single quantum dot - cavity system , '' nature * 455 * , 896899 ( 2007 ) . d. englund , a. faraon , i. fushman , n. stoltz , p. petroff , and j. vukovi , `` controlling cavity reflectivity with a single quantum dot , '' nature * 450 * , 857861 ( 2007 ) .
y .- s .
park , a. k. cook , and h. wang , `` cavity qed with diamond nanocrystals and silica microspheres , '' nano lett .
* 6 * , 20752079 ( 2006 ) .
p. e. barclay , o. painter , c. santori , k .-
m .
fu , and r. beausoleil , `` coherent interference effects in a nano - assembled optical cavity - qed system , '' opt .
express * 19 * , 8081 ( 2009 ) .
s. schietinger , t. schroder , and o. benson , `` one - by - one coupling of single defect centers in nanodiamonds to high - q modes of an optical microresonator , '' nano lett .
* 8 * , 39113915 ( 2008 ) .
c. f. wang , y .- s .
choi , j. c. lee , e. l. hu , j. yang , and j. e. butler , `` observation of whispering gallery modes in nanocrystalline diamond microdisks , '' appl .
. lett . *
90 * , 081110 ( 2007 ) . c. f. wang , r. hanson , d. d. awschalom , e. l. hu , t. feygelson , j. yang , and j. e. butler , `` fabrication and characterization of two - dimenstional photonic crystal microcavities in nanocrystalline diamond , '' appl .
. lett .
* 91 * , 201112 ( 2007 ) .
y. shen , t. m. sweeney , and h. wang , `` zero - phonon linewidth of single nitrogen vacancy centers in diamond nanocrystals , '' phys .
rev .
b * 77 * , 033201 ( 2008 ) .
p. tamarat , t. gaebel , j. r. rabeau , m. khan , a. d. greentree , h. wilson , l. c. l. hollenberg , s. prawer , p. hemmer , f. jelezko , and j. wrachtrup , `` stark shift control of single optical centers in diamond , '' phys .
rev .
lett . * 97 * , 083002 ( 2006 ) .
s. tomljenovic - hanic , m. steel , c. de sterke , and j. salzman , `` diamond based photonic crystal microcavities , '' opt . express * 14 * , 35563562 ( 2006 ) . c. kreuzer , j. riedrich - mller , e. neu , and c. becher , `` design of photonic crystal microcavities in diamond films , '' opt .
express * 16 * , 16321644 ( 2008 ) .
b. a. fairchild , p. olivero , s. rubanov , a. d. greentree , f. waldermann , i. w. robert a. taylor and , j. m. smith , s. huntington , b. c. gibson , d. n. jamieson , and s. prawer , `` fabrication of ultrathin single - crystal diamond membranes , '' adv .
mater . * 20 * , 47934798 ( 2008 ) .
k .-
m . c. fu , c. santori , p. e. barclay , i. aharonovich , s. prawer , n. meyer , a. m. holm , and r. g. beausoleil , `` coupling of nitrogen - vacancy centers in diamond to a gap waveguide , '' appl .
. lett . * 93 * , 234107 ( 2008 ) .
a. r. m. zain , n. p. johnson , m. sorel , and r. m. d. l. rue , `` ultra high quality factor one dimensional photonic crystal / photonic wire micro - cavities in silicon - on - insulator ( soi ) , '' opt . express * 16 * , 1208412089 ( 2008 ) .
i. bayn and j. salzman , `` ultra high - q photonic crystal nanocavity design : the effect of a low-@xmath3 slab material , '' opt .
express * 16 * , 49724980 ( 2008 ) .
k. rivoire , a. faraon , and j. vukovi , `` gallium phosphide photonic crystal nanocavities in the visible , '' appl .
. lett . * 93 * , 063103 ( 2008 ) .
s. g. johnson , p. r. villeneuve , s. fan , and j. d. joannopoulos , `` linear waveguides in photonic - crystal slabs , '' phys .
rev .
b * 62 * , 82128222 ( 2000 ) .
p. lalanne , `` electromagnetic analysis of photonic crystal waveguides operating above the light cone , '' ieee j. quantum electron . *
38 * , 800804 ( 2002 ) .
k. srinivasan and o. painter , `` momentum space design of high - q photonic crystal nanocavities in two - dimensional slab waveguides , '' opt .
express * 10 * , 670684 ( 2002 ) .
http://www.opticsexpress.org/abstract.cfm?uri=opex-10-15-670 o. painter and k. srinivasan , `` localized defect states in two - dimensional photonic crystal slab waveguides : a simple model based upon symmetry analysis , '' phys . rev .
b * 68 * , 035110 ( 2003 ) .
all fdtd simulations used the meep software package .
http://ab-initio.mit.edu/wiki/index.php/meep
p. velha , j. c. rodier , p. lalanne , j. p. hugonin , d. peyrade , e. picard , t. charvolin , and e. hadji , `` ultra - high - reflectivity photonic - bandgap mirrors in a ridge soi waveguide , '' new j. phys .
* 8 * , 204 ( 2006 ) .
a. w. snyder and j. d. love , _ optical waveguide theory _
( chapman and hall , new york , ny , 1983 ) .
e. istrate and e. h. sargent , `` photonic crystal heterostructures and interfaces , '' rev .
mod .
phys .
* 78 * , 455 ( 2006 ) .
m. notomi , e. kuramochi , and h. taniyama , `` ultrahigh - q nanocavity with 1d photonic gap , '' opt .
express * 16 * , 11095 ( 2008 ) .
j. s. foresi , p. r. villeneuve , j. ferrera , e. r. thoen , g. steinmeyer , s. fan , j. d. joannopoulos , l. c. kimerling , h. i. smith , and e. p. ippen , `` photonic - bandgap microcavities in optical waveguides , '' nature * 390 * , 143145 ( 1997 ) .
a. jugessur , p. pottier , and r. de la rue , `` one - dimensional periodic photonic crystal microcavity filters with transition mode - matching features , embedded in ridge waveguides , '' electron . lett . * 39 * , 367 369 ( 2003 ) . m. w. mccutcheon and m. lonar , `` design of an ultrahigh quality factor silicon nitride photonic crystal nanocavity for coupling to diamond nanocrystals , '' opt .
express * 16 * , 19136 ( 2008 ) .
j. chan , m. eichenfield , r. camacho , and o. painter , `` optical and mechanical design of a `` zipper '' photonic crystal optomechanical cavity , '' opt .
express * 17 * , 38023817 ( 2009 ) .
o. painter , k. srinivasan , and p. e. barclay , `` wannier - like equation for the resonant cavity modes of locally perturbed photonic crystals , '' phys .
rev .
b * 68 * , 035214 ( 2003 ) . p. e. barclay , k. srinivasan , and o. painter , `` design of photonic crystal waveguides for evanescent coupling to optical fiber tapers and integration with high - q cavities , '' j. opt
. soc .
amer .
b * 20 * , 22742284 ( 2003 ) .
d. morie and t. baba , `` dispersion - controlled optical group delay device by chirped photonic crystal waveguides , '' appl .
. lett . * 85 * , 11011103 ( 2004 ) .
p. lalanne , s. mias , and j. p. hugonin , `` two physical mechanisms for boosting the quality factor to cavity volume ratio of photonic crystal microcavities , '' opt .
express * 12 * , 458467 ( 2004 ) . c. sauvan , g. lecamp , p. lalanne , and j. hugonin , `` modal - reflectivity enhancement by geometry tuning in photonic crystal microcavities , '' opt .
express * 13 * , 245255 ( 2005 ) .
n. b. manson , j. p. harrison , and m. j. sellars , `` nitrogen - vacancy center in diamond : model of the electronic structure and associated dynamics , '' phys .
rev .
b * 74 * , 104303 ( 2006 ) .
h. j. kimble , `` strong interactions of single atoms and photons in cavity qed , '' phys
. scr . * t76 * , 127137 ( 1998 ) . c. santori , p. e. barclay , k .-
m . c. fu , and r. g. beausoleil , `` vertical distribution of nitrogen - vacancy centers in diamond formed by ion implantation and annealing , '' phys .
rev .
b * 79 * , 125313 ( 2009 ) . v. s. c. manga rao and s. hughes , `` single quantum dot spontaneous emission in a finite - size photonic crystal waveguide : proposal for an efficient `` on chip '' single photon gun , '' phys .
rev .
lett .
* 99 * , 193901 ( 2007 ) .
e. yablonovitch , d. hwang , t. j. gmitter , l. t. florez , and j. p. harbison , `` van der waals bonding of gaas epitaxial liftoff films onto arbitrary substrates , '' appl .
. lett . * 56 * , 24192421 ( 1990 ) .
p. e. barclay , c. santori , k .-
m .
fu , and r. g. beausoleil , `` microcavities for cavity - qed in single crystal diamond , '' in _ conference on lasers and electro - optics / international quantum electronics conference proceedings _ p. imf3 ( 2009 ) .
g. davies , s. c. lawson , a. t. collins , a. mainwood , and s. j. sharp , `` vacancy - related centers in diamond , '' phys .
rev .
b * 46 * , 1315713170 ( 1992 ) .
j. meijer , b. burchard , m. domhan , c. wittmann , t. gaebel , i. popa , f. jelezko , and j. wrachtrup , `` generation of single color centers by focused nitrogen implantation , '' applied physics letters * 87 * , 261909 ( 2005 ) .
f. waldermann , p. olivero , j. nunn , k. surmacz , z. wang , d. jaksch , r. taylor , i. walmsley , m. draganski , p. reichart , a. greentree , d. jamieson , and s. prawer , `` creating diamond color centers for quantum optical applications , '' diamond and related materials * 16 * , 1887 1895 ( 2007 ) . |
the frequency conversion of light fields has been an important theme in optical physics for around half a century . in quantum information physics the conversion of single photons to and from the telecom wavelength band is a topic of more recent vintage , and is motivated by the desire to minimize optical fiber transmission losses when distributing entangled states over distant quantum memory elements in a quantum repeater @xcite .
an associated technical problem is that telecom light is not readily stored in ground level atomic memory coherences .
retrieval processes in atomic ensembles , for example using electromagnetically induced transparency @xcite , or more specifically the dark - polariton mechanism @xcite , generate shorter wavelength radiation correlated to the stored atomic excitation by raman scattering .
such radiation , optically resonant to the ground level of typical atoms and ions , has been retrieved in numerous experiments @xcite .
an important advance would involve generation of atomic memory coherences quantum - correlated with telecom wavelength radiation , thereby minimizing transmission losses over long distances .
recently there has been a breakthrough in this direction using a pair of cold , nondegenerate rubidium gas samples @xcite .
the stored excitation is correlated with an infrared field ( idler ) in one gas sample , and the idler is then frequency converted to a telecom wavelength signal field in the other ensemble .
the frequency conversion mechanism involves the diamond configuration of atomic levels shown in fig . 1 . in a probabilistic protocol
it is important to maximize all efficiencies , e.g. , fiber transmission , single - photon detection and quantum memory lifetime @xcite . in the present work
we investigate the efficiency of frequency up- and down- conversion in the diamond atomic configuration @xcite , as a function of the ensemble s optical thickness , and the intensity and detuning of the pump fields involved in the near - resonant , four - wave mixing process .
the remainder of this paper is organized as follows . in sec .
ii we discuss the four wave mixing process and present solutions for the up- and down- converted fields ; the dressed state picture is used as a guide to understand the characteristic features of the absorption and signal - idler field coupling , as a function of the microscopic interaction parameters . in sec .
iii we present results of an optimization of the conversion efficiency as a function of the optical depth of the atomic ensemble . in sec .
iv we investigate the effects of finite pulse duration by integrating the maxwell - bloch equations for the system .
section v presents our conclusions .
the derivations of the maxwell - bloch and parametric equations are relegated to the appendix . and propagated probe fields ( single line ) @xmath0 interact with the atomic medium .
various detunings are defined in the appendix and the atomic levels used in the experiment @xcite are @xmath1,width=340,height=245 ]
in this paper , we consider a cold and cigar - shaped @xmath2rb atomic ensemble with co - propagating light fields similar to the experimental setup in ref .
@xcite . the conversion scheme shown in fig .
[ four ] involves two pump lasers with frequencies @xmath3 and @xmath4 , respectively ; their rabi frequencies are given by @xmath5 and @xmath6 .
two weak probe fields , signal and idler , with frequency @xmath7 and @xmath8 , respectively , propagate through the optically thick atomic medium . unlike the cascade driving scheme , where two - photon excitation generates a photon pair spontaneously @xcite , pump laser b experiences a transparent medium if both the signal and idler fields are in the vacuum state . with an incident signal field , four wave mixing with the pumps generates an up - converted idler field , while an incident idler field generates a down - converted signal .
the maxwell - bloch equations for the interacting system of light and four light fields is derived in the appendix . by linearizing the equations with respect to the signal and idler field amplitudes , and adiabatically eliminating the atoms ,
one arrives at coupled parametric equations for the signal and idler fields .
we discuss their solution in this section , and leave numerical solutions of the maxwell bloch equations to sec .
the calculation of conversion efficiencies can also be carried out with the quantized heisenberg - langevin version of the coupled parametric equations .
the resulting conversion efficiencies are identical to the semi - classical treatment ; the additional quantum noise contributions vanish as the @xmath9 transition driven by pump laser b has vanishing populations and atomic coherence .
a similar simplification occurs in the calculation of the storage efficiency of spin waves in a system of atoms in the @xmath10 configuration @xcite . the co - moving propagation equation for c - number signal and idler fields ( respectively , @xmath11 and @xmath12 ) under energy conservation ( @xmath13 ) and four - wave mixing conditions ( @xmath14 ) are@xmath15 the coupled equations are similar to those found for the double @xmath10 system @xcite .
the self - coupling coefficients @xmath16 @xmath17 and parametric coefficients @xmath18 , @xmath19 are defined in the appendix .
the solution , under conditions of down conversion with boundary condition @xmath20 , is @xmath21 \end{bmatrix } , \nonumber\\ & \end{aligned}\ ] ] where @xmath22 , and @xmath23 .
we define the down conversion efficiency @xmath24 and transmission of input idler field @xmath25 by @xmath26\right\vert ^{2}.\end{aligned}\ ] ] for up - conversion , ( @xmath27 ) symmetry gives the corresponding coefficients @xmath28\right\vert ^{2}.\end{aligned}\ ] ] the up and down conversion efficiencies differ only by the interchange @xmath29 in the strong parametric coupling regime where @xmath30 , the coefficients can be simplified to @xmath31 and @xmath32 and @xmath33 two strong fields @xmath34 shift the levels with energy @xmath35 and wavy lines represent the idler field resonances.,width=377,height=283 ] under the further assumptions @xmath36 and @xmath37 are pure imaginary , we find @xmath38 @xmath39 $ ] and @xmath40 $ ] .
this result was recently derived by gogyan using a dressed - state approach @xcite , in the case of resonant pump fields @xmath41 @xcite . in this ideal limit
there is a conservation condition @xmath42 the parametric coupling coefficients are not identical , but in the regime of strong coupling they approach each other . as noted by gogyan ,
when the pump - a intensity is large ( @xmath43 ) the @xmath44 is saturated , the atomic coherence is negligible and @xmath45 .
alternatively , in the limit @xmath46 and @xmath47 the atomic coherence of @xmath44 dominates and once again @xmath48 .
the ac - stark splitting induced by the pump lasers shifts the resonant absorption condition for the idler and signal fields . the idler and signal experience resonant absorption at the transition frequency of the dressed atom .
the corresponding transitions for the idler are shown in fig.[dress ] .
the bare states are shifted by @xmath49 and @xmath50 respectively .
note that our rabi frequencies are smaller by a factor 2 than the standard definitions to avoid a plethora of prefactors in the equations of the appendix . for resonant pump fields , @xmath51 .
the idler transition resonances are at @xmath52 @xmath53 @xmath54 @xmath55 and these delineate three windows separated by these four absorption peaks . for @xmath56
the centers of these windows are at @xmath57 @xmath58 and @xmath5 , respectively . choosing the idler detuning @xmath59 as in @xcite , the idler interacts with the atomic medium at the center of the left or right window . as an example of the strong coupling windows created by intense pump lasers , we show in fig .
[ coefficient ] the self- and cross - coupling coefficients for the signal and idler fields as a function of the idler frequency .
note that the corresponding frequency of signal field is determined by @xmath60 .
the dimensionless quantities @xmath61 , @xmath62 , and @xmath63 are shown under the conditions of maximum conversion efficiency to be discussed in the next section .
we choose the optical depth ( opd ) @xmath64 where @xmath65 is the number density , @xmath66 the resonant absorption cross section , and @xmath67 the atomic ensemble length in the propagation direction .
three parametric coupling windows are separated by two strong absorption peaks on the left and two relatively weak ones on the right .
the imaginary part of the self - coupling coefficients are seen to vanish in each window at a certain point , while the real parts are small away from resonances . at the same time
the cross - coupling coefficients have a large imaginary part .
the positive gradient of im@xmath68 and im@xmath69 , inside the windows is indicative of normal dispersion .
it is important to ascertain the parameters which allow maximum efficiency of conversion due its potential in practical quantum information processing . in principle
we need to search the three parametric coupling windows to find the optimum conditions for an atomic ensemble of a given optical thickness . in the previous section
we have discussed how three parametric coupling windows appear for some particular values of pump laser parameters . in the search for the maximal conversion efficiency , five parameters @xmath5 , @xmath6 , @xmath70 , @xmath71 , and @xmath72
are varied to maximize the conversion efficiency for a fixed optical depth of atomic ensemble , using functional optimization .
the optical depth @xmath73 appears through the dependence on atomic number @xmath74 in the arecchi - courtens cooperation time @xmath75
@xcite @xmath76 in fig .
[ opd ] , we show the maximum of down conversion efficiency using eq.([down ] ) for different optical depths from @xmath77 to @xmath78 . the maximum is found by varying five parameters mentioned above and the conversion efficiency reaches @xmath79 asymptotically when the optical depth becomes larger . in the strong parametric coupling regime as we discussed in the previous section , @xmath80 $ ] and
it has a maximum when @xmath81 see fig.[coefficient ] .
since @xmath82 is proportional to optical depth and inversely proportional to the rabi frequencies of the driving lasers , an order of magnitude estimate of the optical depth necessary for near unit conversion efficiency is opd@xmath83 .
the behavior of the cross - coupling coefficient @xmath84 as a function of idler detuning indicates where large conversion is to be found , as a comparison with fig.[trans ] shows
. the maximum efficiency of about @xmath85 is located in the left parametric coupling window at the intersection of @xmath84 and @xmath86 . inside the windows
the trade - off between conversion and transmission is clear . in the region where absorption is large , on the sides of the window ( especially for the left window ) , the efficiency and the transmission are both low although the valley in conversion efficiency corresponds to a peak in transmission as expected in parametric coupling .
the transmission approaches unity when the incident idler field is far off - resonance .
we note that the symmetry @xmath87 gives degenerate optimal conversion conditions .
the effect of finite - duration input probe pulses , which are often employed in practice , can be assessed by numerically solving the maxwell - bloch equations for the coupled atoms - fields system .
the characteristic scales of time and length are given by the arecchi - courtens time @xmath75 and @xmath88 , respectively , which are inversely proportional to the square root of the opd .
the cooperative electric field is the product of the atomic number and the idler electric field per photon , i.e. , @xmath89 scaling the space , time , and electric field amplitude accordingly , indicated by tildes , the light propagation equation becomes @xmath90 a consistent scaling can also be applied to the atomic dynamical equations presented in the appendix .
the maxwell - bloch equations were integrated with a semi - implicit finite difference method @xcite .
the mid - point integration method is stable and has high accuracy without sacrificing memory for finer grids @xcite .
the algorithm has been tested by comparing with the parametric equations solutions in appropriate limits , and these solutions are recovered when fine enough grids are employed . to illustrate the influence of finite pump pulse duration , we compute the down conversion efficiency @xmath91 in fig . [ pulse ] , we show the computed values of @xmath24 for two different input idler pulse durations .
we fix the opd@xmath92 and use the near optimum parameters ( @xmath5 , @xmath6 , @xmath70 , @xmath71 , @xmath72 ) @xmath93 , @xmath94 , @xmath95 , @xmath96 , @xmath97 determined from the coupled parametric equations .
the temporal shape of the pump laser intensities are also shown .
pump laser b is taken to be continuous wave , while pump a is a square pulse with duration large enough to completely overlap the input idler pulse . to compare with the steady state solutions , we choose the rabi frequency of idler as @xmath98 which is small compare to those of the pumps .
we find that the conversion efficiency is reduced for shorter idler pulse inputs .
a @xmath99 ns idler pulse is long enough that it has _ * a * _ almost the same maximum conversion efficiency of @xmath85 as in fig.[opd ] for opd=150 . while for the shorter idler pulse of @xmath100 ns , the signal develops significant temporal modulation and this reduces the conversion efficiency , although it is still quite appreciable .
the modulation frequency is at the generalized rabi frequency of pump - a @xmath101 .
we note the characteristic time and space scales of the calculations are @xmath102 ns and @xmath103 mm for a moderate atomic density @xmath104@xmath105 and @xmath106 mm .
the grid size for dimensionless time @xmath107 and space @xmath108 were chosen for both @xmath99 and @xmath100 ns idler pulse durations and the convergence is reached with an estimated relative error less than @xmath109 .
we have studied light frequency conversion in an atomic ensemble with a diamond configuration of atomic levels such as @xmath2rb .
the motivation stems from the need to efficiently convert light resonant with ground state transitions ( storable in the sense of quantum memories ) to and from the telecom wavelength band for low - loss quantum network communication .
the optically thick atomic sample is driven by two strong co - propagating pump fields , and a probe idler or signal field depending on whether we consider down- or up - conversion .
parametric equations for the probe fields are derived and used to compute conversion efficiencies . by performing a global search
we find conditions of pump rabi frequencies , detunings and signal / idler input frequency to maximize the conversion efficiency as a function of optical depth of the ensemble . only in the limit of very large optical depth
does the maximum efficiency approach the ideal strong coupling result @xcite . under conditions routinely obtained in cold , non - degenerate rubidium gas , with opd @xmath110 ,
optimal conversion efficiencies of the order @xmath111 to @xmath112 are predicted . numerical solution of the maxwell - bloch equations confirms the solution of the parametric equations in the limit of long pulse duration , and indicates that for shorter pulses , pump pulse induced modulation may reduce the conversion efficiency .
* we acknowledge support from nsf and we thank alex kuzmich , alex radnaev and stewart jenkins for helpful discussions .
to derive the coupled maxwell - bloch equations it is convenient to employ a quantized description of the electromagnetic field @xcite and use heisenberg - langevin equation methods , and then invoke a standard semiclassical factorization assumption .
the propagation length @xmath67 is discretized into @xmath113 elements .
the positive frequency component of the electric field operator is given by @xmath114 where @xmath115=\delta_{nn^{\prime}}$ ] , @xmath116 @xmath117 @xmath118 and @xmath119is the central frequency .
define the local boson operators @xmath120 where @xmath121=\delta_{ll^{\prime}}$ ] .
similar definitions hold for the signal , s , and idler field , @xmath122 , which carry an additional index in the following .
the hamiltonian for the interacting system depicted in fig.[four ] is given by , ( we ignore the interactions responsible for atomic spontaneous emission for the moment ) @xmath123 where @xmath124 and @xmath125 where @xmath126 , the rabi frequency @xmath127 is half the standard definition and @xmath128 is a slowly varying temporal profile without spatial dependence ( ensemble size much less than pulse length ) . the dipole matrix element @xmath129 @xmath130 , coupling strength @xmath131 and @xmath132 @xmath133 . the matrix @xmath134 accounts for field propagation by coupling the local mode operators . the dynamical equations including dissipation due to spontaneous emission may be treated by standard langevin - heisenberg equation methods @xcite , and we define @xmath135 as the natural transition rate from @xmath136 since we are interested in a semiclassical description , we replace the field operators by c - numbers in the langevin equations , and drop the zero - mean langevin noise sources . all atomic spin operators are also replaced by their expectation values . finally , in the co - moving frame coordinates @xmath137 and @xmath138 the atomic equations are where
@xmath140 @xmath141 the four - wave mixing mismatch wavevector @xmath142 the frequency mismatch @xmath143 and various detunings are defined as @xmath144 , @xmath145 @xmath146 @xmath147 , @xmath148 .
the slow - varying atomic operators are defined@xmath149@xmath150 where @xmath151 . for energy and momentum conservation ( @xmath154 ) , and in the weak field limit
, we solve the atomic equations in steady state after linearizing with respect to the probe fields @xmath155 where @xmath156 @xmath157 @xmath158 @xmath159 @xmath160 @xmath161 and note that @xmath162 @xmath163 @xmath164 @xmath165 are expressed in first order of fields and @xmath166 @xmath167 in zeroth order . for population operators , we solve them in the zeroth order of fields and the nonzero steady states of population and coherence operator are ( s denotes steady state solution ) @xmath168 substitute the latter results into eq.([steady ] ) , and solve for @xmath169 and @xmath170 the parametric coupling equations for the signal and idler fields become@xmath171 where @xmath172\end{aligned}\ ] ] @xmath173\end{aligned}\ ] ] @xmath174\end{aligned}\ ] ] @xmath175\end{aligned}\ ] ] and @xmath176 | the efficiency of frequency up and down conversion of light in an atomic ensemble , with a diamond level configuration , is analyzed theoretically .
the conditions of pump field intensities and detunings required to maximize the conversion as a function of optical thickness of the ensemble are determined .
the influence of the probe pulse duration on the conversion efficiency is investigated by numeric solution of the maxwell - bloch equations . |
the interaction between sufficiently intense laser pulses and spherical nanoclusters can lead to the complete expulsion of all the electrons from the cluster , as discussed in @xcite , and to a consequently rapid explosion of the left - behind ion cloud .
the process is a well known mechanism of ion acceleration @xcite with many potential applications , ranging from fusion research @xcite to biomolecular imaging @xcite .
while the explosion of a single species ion plasma leads to a broad energy spectrum @xcite , it has been verified that in the presence of ion mixtures conditions exist producing a quasi - monoenergetic distribution of the species having the larger charge - to - mass ratio @xcite .
numerical simulations also pointed out a connection between the monochromaticity of the spectrum and the occurrence of overtakings among faster ions , leading to the formation of shock shells @xcite .
the present work investigates in detail and from a theoretical point of view the energy spectrum of charged particles accelerated during the coulomb explosion of ion mixtures .
in particular , a mixture composed by two types of ions uniformly distributed is considered .
the ions of the two species have masses @xmath0 , @xmath1 and electric charges @xmath2 , @xmath3 , with @xmath4 the ions , which are initially at rest , start moving under the influence of the repulsive electrostatic forces , and the species with a larger charge - to - mass ratio moves faster . in this way , two concentric spherical regions , @xmath5 and @xmath6 ,
are created having radius @xmath7 respectively , with @xmath8 .
the sphere @xmath6 contains fast and slow particles and @xmath9 represents the frontline of the slow ions ; instead , the spherical shell in @xmath5 outside @xmath6 contains only fast particles .
the ion dynamics is different in these two regions , and it is studied in detail in sect .
2 ( inner region ) and in sect .
3 ( outer region ) . the dynamics of the explosion depends strongly on the presence of shock shells .
it is proved rigorously that the appearance of shocks depends only on the ratio between the total charge of the fast ions and the one of the slow ions , and not on their masses ( sect .
4 ) . in sect . 5 , by using a hamiltonian approach the properties of the energy spectra are studied analytically . in particular , a rigorous proof is provided for the existence of a threshold for the generation of shock shells ( previously this proof was known only for the limit situation of @xmath10 @xcite ) , so confirming the results of numerical simulations @xcite .
finally , some conclusions are reported in sect .
whenever possible , the properties of the explosions have been deduced analytically , otherwise the study has been performed numerically with the so called `` shell method '' @xcite , which provides extremely precise results by exploiting the spherical symmetry of the problem .
initially , all the ions are uniformly distributed inside a sphere of radius @xmath11 . as
the charge density is constant , the radial electric field is linear .
therefore , there will be a uniform expansion for both slow and fast ions , and consequently the charge density will remain constant inside the sphere @xmath6 , producing again a linear behaviour of the electric field . from this consideration
, one can infer that the electric field inside @xmath6 remains a linear function of the radius @xmath12 at any time : @xmath13 under this assumption , the equations of motion for the fast and the slow ions inside @xmath6 can be written in a simple way .
the uniform expansion can be described by introducing two functions , @xmath14 and @xmath15 , such that slow ions and fast ions initially at @xmath16 are moved at time @xmath17 to @xmath18 and @xmath19 , respectively .
moreover , the initial densities of the two species , @xmath20 and @xmath21 , evolve in time as @xmath22 using gauss s law , the electric field inside the sphere @xmath6 can be written as : @xmath23 which has the correct dependence on @xmath12 , as in eq . .
after introducing expression into the equations of the motion for fast and slow ions , by writing the acceleration of the ions of the two species as @xmath24 and @xmath25 , one finally obtains @xmath26 in which @xmath27 does not appear .
equations can be rewritten in a more compact way , as @xmath28 being @xmath29^{1/2 } , \quad \alpha=\dfrac{n_{f,0}}{n_{f,0}+n_{s,0 } } , \quad \beta=\dfrac{q_s}{q_f}. \ ] ] the quantity @xmath30 represents a characteristic frequency for the fast ion expansion , while @xmath31 is the fraction of fast ions in the cluster .
the system of equations describes the dynamics of slow ions during the whole explosion and the motion of fast ions as long as they are inside @xmath6 .
a fast ion , initially at @xmath32 , reaches the border of the inner region when @xmath33 .
the time the ion crosses the sphere @xmath6 , indicated in the following as @xmath34 , is obtained by solving the equation @xmath35 considering a fast ion initially at @xmath27 , the electric field for @xmath36 can be written as @xmath37 , where @xmath38 is the charge in the sphere of radius @xmath12 at time @xmath17 ; as long as overtakings between fast ions do not occur , @xmath38 is the sum of the total charge of slow ions and the charge of the fast ions initially inside a sphere of radius @xmath27 : @xmath39 consequently , for @xmath40 , the equation of motion for a fast ion outside @xmath6 is @xmath41 equation can be integrated in order to obtain the asymptotic kinetic energy of the fast ions , @xmath42 , as @xmath43 the energy spectrum of the fast ions , @xmath44 , can be written as @xmath45 where @xmath46 is the total number of fast ions , while @xmath47 represents the number of ions having asymptotic energy in the interval @xmath48 . making use of eqs . and , the spectrum can be expressed parametrically as a function of @xmath49 , as : @xmath50 as pointed out in a previous work @xcite , shock shells arise when the fraction of fast ions , @xmath51 , is smaller than a critical value , @xmath52 .
typical phase - space evolutions of the fast ions are showed in fig .
[ noshock ] for @xmath53 and in fig .
[ shock ] for @xmath54 .
in particular , the curves in fig .
[ shock ] show the typical wave - breaking behaviour that indicates the presence of a shock @xcite . with @xmath55 , for @xmath17 in the range 12 - 57 fs . a cluster with ion density @xmath56=10@xmath57 @xmath58 and @xmath11 = 6.5 nm has been considered . ] with @xmath55 , for @xmath17 in the range 12 - 57 fs . a cluster with ion density @xmath56=10@xmath57 @xmath58 and @xmath11 = 6.5 nm has been considered . ] in fact , there is a simple way to discriminate coulomb explosions without shocks from the ones in which this phenomenon appears without requiring a full numerical simulation of the explosion . starting from eq . , which , in principle , is valid only when shocks are absent , the derivative of @xmath59 with respect to @xmath27 is calculated .
if the condition @xmath60 is verified for every @xmath61 $ ] , inner ions take a slower velocity respect to the ones in the outer zones and can not reach them . in this case , no overtaking is present . instead
, if @xmath62 is negative for some @xmath27 , in principle there will be overtakings .
of course , eq . is no longer valid in this case and it can not be used to calculate the energy spectrum of fast ions , whose dynamics can be solved only numerically . using this criterion ,
the existence of a limit value for @xmath51 can be readily verified numerically . in general ,
if @xmath63 the derivative @xmath62 remains positive and no shock is formed ; for @xmath64 , @xmath62 changes its sign , meaning that the fast particles overtake each other , giving rise to shock shells .
the calculations show that @xmath52 depends only on the ratio between the ions charges and it is independent of their masses @xcite .
in particular , @xmath65 for mixtures @xmath66 , @xmath67 , @xmath68 and @xmath69 .
starting with the case of no shock , by indicating with @xmath70 and @xmath71 the total charge of slow and fast ions , the dynamics of a fast ion initially at @xmath16 is governed by the hamiltonian @xmath72 where @xmath73 is the electrostatic potential in @xmath12 due to fixed ions uniformly distributed in a sphere of radius @xmath74 : @xmath75 in the limit situation @xmath76 , @xmath77 is constant and equal to @xmath11 .
consequently , as @xmath78 does not depend explicitly on time , it is a constant of motion , and the asymptotic energy @xmath42 expressed as a function of @xmath27 can be written immediately as @xmath79 , \label{energy_infinity}\ ] ] from which the energy spectrum can be readily calculated @xcite . ) and hydrogen ions , for different value of @xmath51 in the range @xmath80 $ ] . the same initial density and radius as in fig .
[ noshock ] have been considered .
curves a , b and c represent mixtures with @xmath53 ; in absence of shocks the numerical spectra correspond to the analytic curve .
the limit case of @xmath81 is shown in curve d , where the energy spectra is singular .
curves e , f and g display situations with @xmath53 , for which the spectra have been computed by means of the shell method . ]
two situations can occur , depending on the sign of @xmath82 . if @xmath83 the ion velocity is an increasing function of @xmath27 ,
so overtakings do not occur , and the analytic spectrum is correct , as shown in fig .
[ spectra ] , curves a , b , c. when @xmath84 , according to eq .
, @xmath42 is a decreasing function of @xmath27 , but this implies that the ions closer to the center of the cluster are bound to overtake the ions close to the border of the sphere . in other words ,
the hypothesis of no overtakings , and consequently eq . ( [ energy_infinity ] ) , are no longer valid . in fig .
[ spectra ] , curves e , f , g , the energy spectra have been calculated correctly by using the shell method .
the case @xmath85 ( fig .
[ spectra ] , curve d ) can be considered as a limit for @xmath86 and , consequently , eq .
can still be used . in the general case ,
where the mass of slow ions is finite , in the hamiltonian the potential due to slow ions , @xmath87 , is a function of time , as it depends on the frontline @xmath77 , and @xmath78 is no longer a constant of motion : @xmath88 in principle , the value of the asymptotic energy can be obtained by integrating eq .
( [ hamiltonian ] ) in time : @xmath89 where the upper integration limit can be set equal to @xmath90 , as for @xmath91 the ion is outside @xmath92 and the term @xmath93 vanishes .
noticing that @xmath94 is equal to the asymptotic energy for fixed slow ions , eq .
, and expressing @xmath95 and @xmath77 as functions of @xmath96 and @xmath97 , one finally obtains @xmath98-\dfrac{3 q_s q_f}{2 r } i(r_0 ) \label{energy_finite}\ ] ] where @xmath99 is defined as @xmath100 \dfrac{\xi_s'(t)}{\xi_s^2(t ) } \textrm{d}t .
\label{i0}\ ] ] the quantity @xmath101 is always non negative , and , in fact , the motion of slow ions takes some energy away from the fast ones .
+ the condition @xmath102 for @xmath61 $ ] is now investigated .
starting from eq .
, one obtains : @xmath103 from this formula , considering that @xmath104 , one has @xmath105,\quad \dfrac{\textrm{d}i}{\textrm{d}r_0 } ( r ) = 0 .
\end{matrix}\ ] ] being @xmath106 if @xmath83 the derivative @xmath107 is always non negative , as it is the sum of two non - negative quantities . instead ,
if @xmath84 a sum of a negative term and a non negative term is present in , and in principle nothing can be said about the sign of @xmath107 for a generic @xmath27 .
however , @xmath108 vanishes for @xmath109 , and therefore @xmath110 must be negative ; consequently , in this case there are overtakings .
+ in summary , the condition @xmath85 is valid in any case and can be used in order to discriminate between explosions with and without shocks .
this is the rigorous proof of a property of the coulomb explosions of mixtures that was found numerically in a previous work @xcite . the condition @xmath85 provides the critical value of @xmath51 : @xmath111 which depends only on the charge ratio
the analysis performed for the simplified situation of @xmath112 suggest that fast ion spectra could be particularly narrow when @xmath51 approaches @xmath52 . to confirm that , a full set of calculations has been carried out for the mixtures @xmath113 , @xmath114 , @xmath115 and @xmath116 .
clusters with ion density @xmath56=10@xmath57 @xmath58 and @xmath11 = 6.5 nm ( as in figs .
[ noshock]-[spectra ] ) have been considered . in figs .
[ fig : emedia_carbonio ] and [ fig : std_dev_carbonio ] the mean value of the energy and the energy spread have been calculated for each type of mixture for @xmath31 . moreover , in figs .
[ fig : c1]-[fig : c4 ] the energy spectra are reported for each mixture for @xmath51 close to @xmath52 .
all the calculations have been performed with the shell method .
+ the plots of the mean ion energy , @xmath117 , as function of @xmath51 , show a linear behaviour .
in fact , the total kinetic energy of the fast ions , @xmath118 , can be written as the difference between the potential energy at @xmath119 ad the total kinetic energy of the slow ions @xmath120 ^ 2\ ] ] and , therefore , it does not depend on the possible presence of shocks . with different values of @xmath51 in the range
@xmath121 $ ] .
the thick curve indicates the distribution with the minimum value of @xmath122 . ] with different values of @xmath51 in the range @xmath121 $ ] .
the thick curve indicates the distribution with the minimum value of @xmath122 . ] from the numerical solution of system , it has been found out that @xmath123 ^ 2 $ ] depends linearly on @xmath51 , and using this result the average kinetic energy @xmath124 becomes a linear function of @xmath51 . with different values of @xmath51 in the range
@xmath125 $ ] .
the thick curve indicates the distribution with the minimum value of @xmath122 . ] with different values of @xmath51 in the range @xmath125 $ ] .
the thick curve indicates the distribution with the minimum value of @xmath122 . ] the standard deviation of the energy of fast ions , @xmath126^{1/2 } $ ] , presents a minimum for a value of @xmath51 which is close to @xmath52 when @xmath127 .
in fact , for @xmath128 the spectrum has a large plateau and a sharp peak .
probably , the plateau gives a non negligible contribution to @xmath122 , while for practical applications the presence of the peak and its energy spread are more important .
in the paper , the dynamics of particles during the coulomb explosion of ion mixtures has been investigated , with particular interest to the energy spectrum of the component with the larger charge - to - mass ratio . exploiting the hypothesis of no overtakings among ions of the same species and the new approach based on the linear behavior for the electric field inside the inner sphere , @xmath6 , containing both slow and fast ions , analytical formulas describing the acceleration both inside and outside @xmath6 ( where only fast ions are present )
, have been derived together with the energy spectrum .
for the first time ( at least to the authors knowledge ) , a simple way to deduce if overtakings between fast ions will take place during the expansion is presented and discussed , confirming the existence , empirically deduced in a previous work by using numerical simulations @xcite , of a limit value for the mixture composition , @xmath52 , above which no shocks are forming .
the analysis here proposed is based on the calculation of the derivative of @xmath42 with respect to the initial ion position @xmath27 : when this quantity is negative for some @xmath27 , the hypothesis of no overtaking is no longer valid .
moreover , making use of the hamiltonian of fast particles and starting with the simple case where the mass of the slow species is considered infinite , a rigorous proof of the existence of @xmath52 , which is valid for every value of @xmath1 and @xmath0 , has been obtained . in the limit situation of @xmath76 ,
when @xmath53 approaches the critical value , the energy spectrum becomes narrower and narrower , meaning that the ions tend to acquire very similar velocities , and for @xmath81 the spectrum becomes monoenergetic . on the other hand ,
the condition @xmath81 is the limit situation for the rising of shock shells .
as most of these considerations are valid also for the general case in which @xmath0 is finite , one can conclude that the rising of shock shells is induced by a very narrow energy spectrum , and not vice versa . in conclusion ,
conditions to obtain nearly monoenergetic ions from coulomb explosions of heteronuclear clusters have been individuated and supported by a rigorous theoretical analysis .
the analysis has been carried out supposing that the electrons are completely expelled from the cluster ; also in order to explain recent experimental results @xcite , future investigation should take into proper account also the electron dynamics and the laser - cluster interaction in order to have a clearer picture of the phenomenon . + this work was partially supported by the european research council ( erc-2010-adg grant no . | the paper presents a theoretical work on the dynamics of coulomb explosion for spherical nanoplasmas composed by two different ion species .
particular attention has been dedicated to study the energy spectra of the ions with the larger charge - to - mass ratio .
the connection between the formation of shock shells and the energy spread of the ions has been the object of a detailed analysis , showing that under particular conditions the width of the asymptotic energy spectrum tends to become very narrow , which leads to a multi - valued ion phase - space . the conditions to generate a quasi mono - energetic ion spectrum
have been rigorously demonstrated and verified by numerical simulations , using a technique that , exploiting the spherical symmetry of the problem , allows one to obtain very accurate and precise results . |
the high statistics data by the superkamiokande collaboration @xcite has confirmed the deficit of atmospheric muon neutrinos , especially at small zenith angles , opening a new era in neutrino physics . on the other hand
the persistent disagreement between solar neutrino data and theoretical expectations has been a long - standing problem in physics @xcite .
altogether these constitute the only solid evidence we now have in favour of physics beyond the present standard model , providing a strong hint for neutrino conversion .
although massless neutrino conversions @xcite can be sizeable in matter , and may even provide alternative solutions of the neutrino anomalies @xcite , it is fair to say that the simplest interpretation of the present data is in terms of massive neutrino oscillations . taking for granted such an interpretation , the present data do provide an important clue on the pattern of neutrino masses and mixing .
the atmospheric data indicate to flavour oscillations with maximal mixing @xcite , while the solar data can be accounted for in terms of either small ( sma ) and large ( lma ) mixing msw solutions @xcite , as well as through vacuum or _ just - so _
solutions @xcite . a large mixing among @xmath8 and @xmath9
is excluded both by the atmospheric data and by reactor data on neutrino oscillations @xcite .
there has indeed been an avalanche @xcite of papers trying to address this issue in the framework of unified models adopting _ ad hoc _ texture structures for the yukawa couplings .
here we propose an alternative approach to describe the structure of lepton mixing which accounts for the atmospheric and solar neutrino anomalies @xcite based on the simplest extension of minimal supergravity with bi - linear r parity violation @xcite .
the particles underlying the mechanism of neutrino mass generation are the neutral supersymmetric partners of the standard model gauge and higgs bosons which have mass at the weak - scale and are thus accessible to accelerators .
our model breaks lepton number and therefore necessarily generates non - zero majorana neutrino masses @xcite . at tree - level only one of the neutrinos picks up a mass by mixing with neutralinos @xcite , leaving the other two neutrinos massless @xcite . while this can explain the atmospheric neutrino problem , to reconcile it with the solar neutrino data requires going beyond the tree - level approximation
this is the purpose of the present paper .
here we improve the work of ref . @xcite by performing a full one - loop calculation of the neutrino mass matrix and also update the discussion in the light of the recent global fits of solar and atmospheric neutrino data .
this can also be used to improve the discussion given in @xcite where the tree approximation was assumed . for
simplified analyses including only the atmospheric neutrino problem in the tree - level approximation see ref .
@xcite and a number of papers in ref .
@xcite .
we have performed a full one - loop calculation of the neutralino - neutrino mass matrix in the bi - linear mssm , showing that , in order to explain the solar and atmospheric neutrino data , it is necessary and sufficient to work at the one - loop level , provided one performs the renormalization of the heaviest neutrino .
in contrast to all existing papers @xcite , we have taken special care to verify the gauge invariance of the calculation , thus refining the approximate approaches so far used in the literature .
we find that if the soft - supersymmetry breaking terms are universal at the unification scale then only the small mixing angle ( sma ) msw solution to the solar neutrino problem exists . on the other hand
if these assumptions are relaxed then one can implement large mixing angle solutions , either msw or just - so .
bilinear r - parity breaking supersymmetry has been extensively discussed in the literature @xcite .
it is motivated on the one hand by the fact that it provides an effective truncation of models where r parity breaks _ spontaneously _ by singlet sneutrino vevs around the weak scale @xcite .
moreover , they allow for the radiative breaking of r - parity , opening also new ways to unify gauge and yukawa couplings @xcite and with a potentially slightly lower prediction for @xmath10 @xcite . for recent papers on phenomenological implications of these models see ref . @xcite .
if present at the fundamental level tri linear breaking of r parity will always imply bi - linear breaking at some level , as a result of the renormalization group evolution .
in contrast , bi - linear breaking may exist in the absence of tri linear , as would be the case if it arises spontaneously .
this paper is organized as follows . in sections
[ themodel ] , [ thepot ] and [ ewbreaking ] we describe the model , the minimization of the scalar potential and the radiative breaking of the electroweak symmetry . in section [ treelevelmasses ] the tree level masses and mixings are described , while the contributions to the one loop mass matrix and the gauge invariance issue are studied in section [ oneloopmassmatrix ] .
finally the neutrino masses and mixings are discussed in section [ oneloopmasses ] where we show our results for solar and atmospheric oscillation parameters .
the more technical questions regarding the mass matrices , couplings and one loop results as well as further details of gauge invariance are given in the appendices .
we also briefly discuss how , despite the smallness of neutrino masses indicated by experiment , the theory can lead to observable phenomena at high - energy accelerators .
using the conventions of refs .
@xcite we introduce the model by specifying the superpotential , which includes brpv @xcite in three generations .
it is given by @xmath11 \label{eq : wsuppot}\ ] ] where the couplings @xmath12 , @xmath13 and @xmath14 are @xmath15 yukawa matrices and @xmath16 and @xmath17 are parameters with units of mass .
the bilinear term in eq .
( [ eq : wsuppot ] ) violates lepton number in addition to r parity .
supersymmetry breaking is parameterized with a set of soft supersymmetry breaking terms . in the mssm
these are given by @xmath18\label{eq : vsoft } \\ & & \!\!\!\!+\varepsilon_{ab}\left [ a_u^{ij}\widetilde q_i^a\widetilde u_j h_u^b + a_d^{ij}\widetilde q_i^b\widetilde d_j h_d^a + a_e^{ij}\widetilde l_i^b\widetilde r_j h_d^a
-b\mu h_d^a h_u^b\right ] \,.\nonumber \end{aligned}\ ] ] in addition to the mssm soft susy breaking terms in @xmath19 the brpv model contains the following extra term @xmath20 where the @xmath21 have units of mass . in what follows
, we neglect intergenerational mixing in the soft terms in eq .
( [ eq : vsoft ] ) .
the electroweak symmetry is broken when the two higgs doublets @xmath22 and @xmath23 , and the neutral component of the slepton doublets @xmath24 acquire vacuum expectation values .
we introduce the notation : @xmath25 where we shift the neutral fields with non zero vevs as @xmath26\,,\quad h_u^0\equiv{1\over{\sqrt{2}}}[\sigma^0_u+v_u+i\varphi^0_u]\,,\quad \tilde{l}_i^0\equiv{1\over{\sqrt{2}}}[\tilde\nu^r_i+v_i+i\tilde\nu^i_i]\ , .
\label{shiftedfields}\ ] ] note that the @xmath27 boson acquires a mass @xmath28 , where @xmath29 .
we introduce the following notation in spherical coordinates for the vacuum expectation values : @xmath30 which preserves the mssm definition @xmath31 . in the mssm limit , where @xmath32 , the angles @xmath33 are equal to @xmath34 .
in addition to the above mssm parameters , our model contains nine new parameters , @xmath17 , @xmath35 and @xmath21 .
the three vevs are determined by the one loop tadpole equations , and we will assume universality of the @xmath36terms , @xmath37 at the unification scale .
therefore , the only new and free parameters can be chosen as the @xmath17 .
the electroweak symmetry is broken when the higgs and lepton fields acquire non zero vevs .
these are calculated via the minimization of the effective potential or , in the diagramatic method , via the tadpole equations .
the full scalar potential at tree level is @xmath38 where @xmath39 is any one of the scalar fields in the superpotential in eq .
( [ eq : wsuppot ] ) , @xmath40 are the @xmath41-terms , and @xmath42 is given in eq .
( [ softbrpv ] ) .
the tree level scalar potential contains the following linear terms @xmath43 where the different @xmath44 are the tadpoles at tree level .
they are given by @xmath45 where we have defined @xmath46 and @xmath47 . a repeated index @xmath48 in eq .
( [ eq : tadpoles ] ) implies summation over @xmath49 .
the five tree level tadpoles @xmath50 are equal to zero at the minimum of the tree level potential , and from there one can determine the five tree level vacuum expectation values .
it is well known that in order to find reliable results for the electroweak symmetry breaking it is necessary to include the one loop radiative corrections .
the full scalar potential at one loop level , called effective potential , is v_total = v_total^0 + v_rc where @xmath51 include the quantum corrections . in this paper
we use the diagramatic method , which incorporates the radiative corrections through the one loop corrected tadpole equations .
the one loop tadpoles are @xmath52 where @xmath53 and @xmath54 are the finite one loop tadpoles . at the minimum of the potential we have @xmath55 , and the vevs calculated from these equations are the renormalized vevs . neglecting intergenerational mixing in the soft masses ,
the five tadpole equations can be conveniently written in matrix form as @xmath56^t= { \bf{m}^2_{tad}}\left[v_u , v_d , v_1,v_2,v_3\right]^t \label{tadpolemat}\ ] ] where the matrix @xmath57 is given by @xmath58 \label{mtad}\ ] ] and depends on the vevs only through the @xmath41 term defined above . in order to have approximate solutions for the tree level vevs ,
consider the following rotation among the @xmath22 and lepton superfields : @xmath59 where the rotation @xmath60 can be split as @xmath61 \times\left[\matrix{c_2 & 0 & 0 & -s_2 & 0 \cr 0 & 1 & 0 & 0 & 0 \cr 0 & 0 & 1 & 0 & 0 \cr s_2 & 0 & 0 & c_2 & 0 \cr 0 & 0 & 0 & 0 & 1 } \right ] \times\left[\matrix{c_1 & 0 & -s_1 & 0 & 0 \cr 0 & 1 & 0 & 0 & 0 \cr s_1 & 0 & c_1 & 0 & 0 \cr 0 & 0 & 0 & 1 & 0 \cr 0 & 0 & 0 & 0 & 1 } \right]\ , .
\label{rotationr}\ ] ] where the three angles are defined as @xmath62 it is clear that this rotation @xmath60 leaves the @xmath41 term invariant .
the rotated vevs are given by @xmath63^t= { \bf{r}}\left[v_u , v_d , v_1,v_2,v_3\right]^t\ , , \label{rotatedvevs}\ ] ] and under the assumption that @xmath64 , these three small vevs have the approximate solution @xmath65\,,\nonumber\\\nonumber\\ & & v'_2\approx-{{\mu'\epsilon_2}\over{m'^2_{l_2}+d}}\left [ { { m'^2_{h_d}-m_{l_2}^2}\over{\mu''\mu'''}}v'_d+{{b_2-b'}\over{\mu''}}v'_u \right]\,,\label{vevspapprox}\\\nonumber\\ & & v'_3\approx-{{\mu''\epsilon_3}\over{m'^2_{l_3}+d}}\left [ { { m''^2_{h_d}-m_{l_3}^2}\over{\mu'''^2}}v'_d+{{b_3-b''}\over{\mu'''}}v'_u \right]\,,\nonumber\end{aligned}\ ] ] where we have defined the following rotated soft terms : @xmath66 the approximation @xmath64 is justified in sugra models with universality of soft masses at the weak scale , as shown in the next section .
it was demonstrated in ref .
@xcite that brpv can be succesfully embedded into sugra with universal boundary conditions at the unification scale , and with a radiatively broken electroweak symmetry . at @xmath67 we assume the standard minimal supergravity unification assumptions , @xmath68 we run the rge s from the unification scale @xmath69 gev down to the weak scale , giving random values to the fundamental parameters at the unification scale : ccccc 10 ^ -2 & & h^2_t_u / 4 & & 1 10 ^ -5 & & h^2_b_u / 4 & & 1 -3 & & a_0 a / m_0&&3 0&&^2_u / m_0 ^ 2&&10 0&&m_1/2/m_0&&5 the yukawa couplings are determined by requiring that three eigenvalues of the chargino / charged - lepton mass matrix corrrespond to the experimentally measured tau , muon , and electron masses as in the mssm , the electroweak symmetry is broken because the large value of the top quark mass drives the higgs mass parameter @xmath70 to negative values at the weak scale via its rge @xcite . in the rotated basis , the parameter @xmath71 is determined at one loop by @xmath72 + { { \left(m'''^2_{h_d}+\widetilde
t^{\overline{dr}}_{v'_d}\,\right)- \left(m_{h_u}^2+\widetilde t^{\overline{dr}}_{v'_u}\,\right)t'^2_{\beta } } \over{t'^2_{\beta}-1 } } \label{muppp2}\ ] ] where @xmath73 is defined in the rotated basis and is analogous to @xmath74 in eq .
( [ eq : vevs ] ) defined in the original basis .
the finite @xmath75 z - boson self energy is @xmath76 , and the one loop tadpoles @xmath77 and @xmath78 are obtained by applying to the original tadpoles in eq .
( [ tadpoles ] ) the rotation @xmath79 defined in eq .
( [ rotationr ] ) .
the radiative breaking of the electroweak symmetry is valid in the brpv model in the usual way : the large value of the top quark yukawa coupling drives the parameter @xmath70 to negative values , breaking the symmetry of the scalar potential .
as we will see a radiative mechanism is also responsible for the smallness of the neutrino masses in models with universality of soft mass parameters at the unification scale .
the relevant parameters are the bilinear mass parameters @xmath36 and @xmath21 , the higgs mass parameter @xmath80 , and the slepton mass parameters @xmath81 .
the rge s for the @xmath36 parameters are @xmath82 where we do not write the effect of yukawa couplings of the first two generations .
similarly , the rge for the down - type higgs mass is @xmath83 and the rge s for the slepton mass parameters are @xmath84 where @xmath85 and @xmath86 . with the aid of these rge s we can find an approximate expression for the slepton vev s in the rotated basis @xmath87 , given in eq .
( [ vevspapprox ] ) .
the relevant soft term differences , defined as @xmath88 and @xmath89 , are approximated by @xmath90 for the @xmath36 terms , and by @xmath91 for the mass squared terms .
this way , if we assume that @xmath92 we can neglect the rotations in eq .
( [ vevspapprox ] ) and we find @xmath93 which give us an approximate expression for the sneutrino vev s @xmath87 in the basis where the @xmath17 terms are absent from the superpotential . in a model with unified universal boundary conditions on the soft susy breaking terms ( sugra case , for short ) the @xmath87 are calculable in terms of the renormalization group evolution due to the non - zero bottom quark yukawa coupling .
we should stress here that for our subsequent numerical calculation we solve the tadpole equations _
exactly_. the symmetry of the neutralino / neutrino mass matrix implies that only one neutrino acquires a tree level mass , and the other two remain massless @xcite ( see next section ) .
the massive neutrino will have the largest component along @xmath94 , @xmath16 or @xmath95 if the largest vev is @xmath96 , @xmath97 , or @xmath98 respectively . on the other hand ,
the most obvious difference between the third generation sneutrino vev and the first two generations is in the extra contribution from @xmath99 to @xmath100 in eq .
( [ dbiapprox ] ) and to @xmath101 in eq .
( [ dm2approx ] ) for the first two generations . due to the tau lepton contribution ,
@xmath102 and @xmath103 are larger than @xmath104 , and similarly for the @xmath105 , specially if @xmath106 .
however , we have checked that it is possible without fine - tuning the parameters in an unnatural way to arrange for the heaviest of the neutrinos to be an equal mixture of and @xmath8 as needed in order to obtain an explanation of the atmospheric neutrino anomaly . that this is possible can be understood by noticing that there can be a cancellation between the @xmath107 and @xmath108 terms in eq .
( [ vpiapprox ] ) for @xmath98 and @xmath97 .
here we discuss the tree level structure of neutrino masses and mixings . for a complete discussion of the fermion mass matrices in this model see appendix [ massmatrices ] . ] ] in the basis @xmath109 the neutral fermion mass matrix @xmath110 is given by _
n = where _ ^0 - 2pt=-4pt is the standard mssm neutralino mass matrix and m= characterizes the breaking of r - parity .
the mass matrix @xmath110 is diagonalized by ( see appendix [ massmatrices ] ) ^*m_nn^-1=diag(m_^0_i , m__j ) [ chi0massmat ] where @xmath111 for the neutralinos , and @xmath112 for the neutrinos .
we are interested in the case where the neutrino mass which is determined at the tree level is small , since it will be determined in order to account for the atmospheric neutrino anomaly .
the above form for @xmath110 is especially convenient in this case in order to provide an approximate analytical discussion valid in the limit of small violation parameters . indeed in this case
we perform a perturbative diagonalization of the neutral mass matrix , using the method of @xcite , by defining @xcite = m _
^0 ^ -1 [ defxi ] if the elements of this matrix satisfy _ ij 1 then one can use it as expansion parameter in order to find an approximate solution for the mixing matrix @xmath113 .
explicitly we have @xmath114 where _
i = v_i + v_d _ i v_i [ lambdai ] are the alignment parameters . from and
one can see that @xmath115 in the mssm limit where @xmath116 , @xmath117 . in leading order in @xmath118 the mixing matrix @xmath113
is given by , ^ * -2pt = -2pt ( -1pt cc n^ * & 0 + 0 & v_^t -1pt ) ( -1pt cc 1 -1 2 ^ & ^ + - & 1 -1 2 ^ -1pt ) the second matrix above block - diagonalizes the mass matrix @xmath119 approximately to the form diag(@xmath120 ) , where m_eff & -3mm=-3 mm & - m _ ^0 ^ -1 m^t & -3mm=-3 mm & ( -2 mm ccc _
e^2 -1pt&-1pt _ e _ -1pt&-1pt _ e _ + _ e _ -1pt&-1pt _ ^2 -1pt&-1pt _ _ + _ e _ -1pt&-1pt _ _ -1pt&-1pt _ ^2 -3 mm ) the sub - matrices @xmath121 and @xmath122 diagonalize @xmath123 and @xmath124
n^*m_^0 n^ = diag(m_^0_i ) , v_^t m_eff v _ = diag(0,0,m
_ ) , where [ mnutree ] m _
= tr(m_eff ) = ||^2 .
clearly , one neutrino acquires mass due to the projective nature of the effective neutrino mass matrix @xmath125 , a feature often encountered in models @xcite . as a result
one can rotate away one of the three angles @xcite in the matrix @xmath122 , leading to @xcite v_= ( ccc 1 & 0 & 0 + 0 & _ 23 & -_23 + 0 & _ 23 & _ 23 ) ( ccc _ 13 & 0 & -_13 + 0 & 1 & 0 + _ 13 & 0 & _ 13 ) , where the mixing angles can be expressed in terms of the _ alignment vector _ @xmath126 as follows : [ tetachooz ] _ 13 = - , [ tetatm ] _ 23 = .
one loop radiative corrections to the neutralino / neutrino mass matrix in the brpv model were calculated first in @xcite , working in the thooft feynman gauge ( @xmath127 ) .
our analysis improves the previous work in that we check explicitly the gauge invariance using the @xmath128 gauge .
we use dimensional reduction to regularize the divergences @xcite and include all possible mssm particles with consistently determined mass spectra and couplings in the relevant loops .
we denote the sum of all one loop graphs contributing to the 2-point function as ( 120,40)(0,22 ) ( 5,25)(40,25 ) ( 60,25)200.5 ( 80,25)(115,25 ) ( 25,34)[]@xmath129 ( 100,34)[]@xmath130 @xmath131 + the most general expression for the one loop contribution to the unrenormalized neutralino / neutrino two point function is @xmath132-\left[\vb{12}p_l\pi^l_{ij}(p^2)+ p_r\pi^r_{ij}(p^2)\right]\right\}\ ] ] where the indices @xmath48 and @xmath133 run from 1 to 7 , @xmath134 and @xmath135 are the right and left projection operators , and @xmath136 is the external four momenta .
the functions @xmath137 and @xmath138 are unrenormalized self energies and depend on the external momenta squared , @xmath139 .
the neutral fermions @xmath129 are a mixture of weak eigenstate neutralinos and neutrinos and given by @xmath140 where @xmath113 is the @xmath141 matrix that diagonalizes the neutralino / neutrino mass matrix according to .
the inverse propagator at one loop is obtained by adding to the tree level propagator , this self energy previously renormalized with the dimensional reduction @xmath75 scheme and denoted as @xmath142 and @xmath143 . in the @xmath75 scheme ,
the counterterms cancel only the divergent pieces of the self energies . in this way
, they become finite and dependent on the arbitrary scale @xmath144 .
the tree level masses are promoted to running masses in order to cancel the explicit scale dependence of the self energies .
thus , the inverse propagator of the neutral fermion @xmath129 is @xmath145 the physical pole mass is given by the zero of the inverse propagator , in the limit where @xmath146 , and may be found using @xmath147u(p)=\overline{u}(p)\big[p_{\mu}\gamma^{\mu}-m_{f_i}(q)+ \widetilde{\bf\sigma}_{ff}^{ii}(p , q)\big]u(p ) \label{zequation}\ ] ] where @xmath148 and @xmath149 are two on - shell spinors , @xmath150 and @xmath151 are the neutral fermion pole and running masses respectively , and @xmath152 is the renormalized two point function in the @xmath75 scheme .
the quantity @xmath153 corresponds to the finite ratio of the infinite wave function renormalization constants in the @xmath75 scheme and the on shell scheme , and it accounts for the fact that the residue of the @xmath75 propagator at the pole is not one @xcite .
the renormalized function @xmath152 is calculated by subtracting the pole terms proportional to the regulator of dimensional reduction = -_e + 4where @xmath154 is the euler s constant and @xmath155 is the number of space time dimensions . in practice
we have @xmath156_{\delta=0}\ , .
\label{delta0eq}\ ] ] since @xmath157 , the terms proportional to @xmath158 in @xmath159 do not contribute . from eq .
( [ zequation ] ) we find @xmath160 where ^v=(^l+^r ) , ^v=(^l+^r ) , and the tilde implies renormalized self energies . a given set of input parameters in the neutralino / neutrino mass matrix defines the set of tree level running masses @xmath151 , among them two massless and degenerate neutrinos .
the one loop renormalized masses @xmath150 are then found through eq .
( [ massdiff ] ) , and the masslessness and degeneracy of the two lightest neutrinos is lifted .
the tree level masslessness of the lightest neutrinos implies an indetermination of the corresponding eigenvectors . in order to find the correct neutrino mixing angles
we diagonalize the one loop corrected neutralino / neutrino mass matrix .
we define m^pole_ij= m^_ij(q ) + m_ij with m_ij= - , [ deltam ] where the symmetrization is necessary to achieve gauge invariance . of course
, the diagonal elements of @xmath161 correspond to the difference between the pole and running masses defined in eq .
( [ massdiff ] ) . as explained in section 2.2 , the one loop corrected vacuum expectation values are found by solving the one loop corrected tadpole equations in eq .
( [ tadpoles ] ) . of course
, it is desirable to work with gauge invariant vevs . in order to achieve the gauge invariance of the @xmath162 s ,
the one loop tadpole @xmath163 must be independent of the gauge parameter @xmath118 . as it is shown in the appendix [ ap : tadpoles ] the following set of tadpoles is gauge invariant : ( 140,60)(0,26 ) ( 20,50)(20,0,360)321 ( 20,30)(20,0)4 ( 60,30)[]@xmath164 ( 120,50)2 ( 118.5,57.7)2 ( 114.1,64.1)2 ( 107.7,68.5)2 ( 100,70)2 ( 92.3,68.5)2 ( 85.9,64.1)2 ( 81.5,57.7)2 ( 80,50)2 ( 81.5,42.3)2 ( 85.9,35.9)2 ( 92.3,31.5)2 ( 100,30)2 ( 107.7,31.5)2 ( 114.1,35.9)2 ( 118.5,42.3)2 ( 100,30)(100,0)4 ( 45,70)[]@xmath27 ( 125,70)[]@xmath165 ( 30,10)[]@xmath166 ( 110,10)[]@xmath166 @xmath167^{w,\eta^{\pm } } $ ] + where @xmath166 denote neutral scalar bosons in the weak basis ( see appendix a ) @xmath168 s are the fadeev - popov ghosts .
a similar set for the @xmath169 gauge boson exists .
nevertheless , the tadpole with a charged goldstone boson in the loop introduces a gauge dependence that can not be canceled .
for this reason , the goldstone boson loops are removed from the tadpoles @xmath170 and introduced into the self energies .
this in turn allows us to achieve the gauge invariance for the two point functions , as well as for the vev s , as explained below . among the loops contributing to the self energies ,
consider for example the @xmath27boson loop , which in the general @xmath128 gauge is ( 120,40)(0,22 ) ( 5,25)(40,25 ) ( 60,25)(20,180,0 ) ( 60,25)(20,0,180)310.5 ( 80,25)(115,25 ) ( 25,34)[]@xmath129 ( 100,34)[]@xmath130 ( 75,50)[]@xmath27 ( 75,0)[]@xmath171 @xmath172 + where the dots indicate terms proportional to @xmath158 which are irrelevant for us , @xmath171 are charged fermions resulting from mixing between charginos and charged leptons , and ( ^v_ij)^w & -5pt=-5pt & - _ k=1 ^ 5 ( o^ncw_l jk o^cnw_l ki + o^ncw_r jk o^cnw_r ki ) \ { 2 b_1(p^2,m^2_k , m^2_w ) + b_0(p^2,m^2_k , m^2_w ) + & & -b_0(p^2,m^2_k , m^2_w ) - } [ wgraph ] + ( ^v_ij)^w & -5pt=-5pt & _ k=1 ^ 5 ( o^ncw_l jk o^cnw_r ki + o^ ncw_r jk o^cnw_l
ki)m_k this graph introduces an explicit dependence on the gauge parameter @xmath118 .
the other self energy graph with @xmath118 dependence is the one that includes the charged goldstone boson .
the charged goldstone boson is one of the eight charged scalars @xmath173 resulting from mixing between the two charged higgs fields and the six charged sleptons .
this contribution is ( 120,40)(0,22 ) ( 5,25)(40,25 ) ( 60,25)(20,180,0 ) ( 60,25)(20,0,180)4 ( 80,25)(115,25 ) ( 25,34)[]@xmath129 ( 100,34)[]@xmath130 ( 75,50)[]@xmath174 ( 75,0)[]@xmath171 @xmath175 + where again , the dots indicate terms proportional to @xmath158 , and ( ^v_ij)^s^+&= & - _ r=1 ^ 8 _ k=1 ^ 5 ( o^ncs_r jkr o^cns_l kir + o^ncs_l jkr o^cns_r kir ) b_1(p^2,m^2_k , m^2_r ) ( ^v_ij)^s^+&= & -
_ r=1 ^ 8 _ k=1 ^ 5 ( o^ncs_l jkr o^cns_l kir + o^ncs_r jkr o^cns_r kir ) m_k b_0(p^2,m^2_k , m^2_r ) [ chascaloop ] with the couplings given in appendix b. nevertheless , gauge dependence is not canceled after combining eqs .
( [ wgraph ] ) and ( [ chascaloop ] ) . in order to achieve it the inclusion of the goldstone boson tadpole graphs , left over from the tadpole equations ,
is necessary : ( 80,60)(0,26 ) ( 20,50)(20,0,360)4 ( 20,30)(20,0)4 ( -20,0)(20,0 ) ( 20,0)(60,0 ) ( 45,70)[]@xmath176 ( 30,15)[]@xmath177 ( -20,8)[]@xmath129 ( 65,8)[]@xmath130 @xmath178 + where @xmath179 and @xmath180 the @xmath181 , @xmath182 and @xmath183 appearing above are passarino - veltman functions @xcite , @xmath184 being the neutral scalar coupling to a pair of charged goldstone bosons , and @xmath185 the neutral scalar couplings to a pair of neutral fermions ( neutralino / neutrinos ) .
numerically , we have checked that , by adding the goldstone tadpoles to the self energies our results do not change by varying the gauge parameter from @xmath127 to @xmath186 , thus establishing the gauge invariance of the calculation . similarly we have also checked that the corresponding set of diagrams involving the neutral gauge boson tadpole ( @xmath169 ) plus neutral ghost tadpoles is gauge invariant and , similarly , the contribution to the self energies due to @xmath169 exchange plus neutral pseudoscalars and neutral goldstones is also gauge invariant .
before we close this section we would like to add a short discussion on the basic structure of the loops which will be useful in the following .
it is useful to do this in the approximation where the parameters are small , as discussed above .
as seen from the expression for @xmath125 , at tree - level the effective neutrino mass matrix in this limit has the structure @xmath187 , and at this level of sophistication neutrino angles are simple functions of ratios of @xmath188 .
the one - loop corrections , however , in general destroy this simple picture . this can be seen as follows .
the one - loop corrections have the general form , ( _ ij,_ij ) ~(o_i , a o_j , b + o_i , co_j , d ) ( b_1,m b_0 ) [ gen1lp ] where the @xmath189 stand symbolically for the various couplings .
now , since the expansion matrix @xmath118 , defined in can be written as @xmath190 ( see ) a product of two couplings involving neutrino - neutralino mixing has the general structure , _ i , a o_j , b ~(f _ _ i + g _ _
i ) ( f _ _ j + g _ _ j ) f ( ... ) [ genstruc ] where all the other dependence on the susy parameters has been hidden symbolically in @xmath191 .
the one - loop corrections therefore also carry a certain index structure , which can be written as m^1-loop_ij ~a _ i _ j + b ( _ i _ j + _ i _ j ) + c _ i _ j [ symb1lp ] where @xmath192 are again complicated functions of susy parameters involving couplings , the passarino - veltman functions etc .
clearly , the terms proportional to @xmath193 in above will lead only to a renormalization of the heaviest neutrino mass eigenstate . on the other hand
the terms proportional to @xmath194 in are genuine loop corrections .
consider the simple case where all @xmath195 .
clearly in this case the tree - level neutrino mass is absent , but the one - loop effective neutrino mass has the same index structure as before , but now in terms of @xmath196 s instead of @xmath197 s . in this idealized case angles
are given as simple funtions of @xmath17 ratios .
for non - zero @xmath198 the terms proportional to @xmath199 in eq .
, however , destroy this simple picture .
any mismatch between @xmath200 and @xmath188 will lead , in general , to a very complex parameter dependence of the neutrino angles .
here we collect our numerical results on neutrino masses and mixings . as we have seen , a characteristic of the brpv model is the appearence of vacuum expectation values for the sneutrino fields , @xmath201 which imply a tree
level mass for one of the neutrinos given by .
the one loop corrected neutrino mass matrix gives important contributions to the heaviest neutrino mass which we have determined through the renormalization procedure sketched above .
first we note that in order to solve the atmospheric @xmath202 solar neutrino problem one requires @xmath203 . if this is fullfilled it is essentially trivial within our model to solve the atmospheric neutrino problem .
it is simply equivalent to choosing an adequate size of the alignment vector @xmath204 , as can be seen from and is also demonstrated in .
however there are regions of parameters where the one - loop contributions are comparable to the tree - level neutrino mass .
this is discussed in quite some detail below , where we give an illustrative parameter study in order to isolate the main features of the dependence on the underlying parameters .
first we get a rough idea of the magnitude of the neutrino masses including the one - loop corrections by displaying in the three lightest eigenvalues of the neutrino / neutralino mass matrix as a function of the parameter @xmath205 .
other parameters are fixed as follows : a ) mssm parameters : @xmath206 gev , @xmath207 gev , @xmath208 , @xmath209 .
b ) rpv parameters : @xmath210 @xmath211 , @xmath212 and @xmath213 . in the left panel
we give the predicted masses in the general case , while in the one on the right we apply the sign condition , ( _ / _ ) ( _ / _ ) 0 [ signcondition ] to be discussed in more detail below .
( 50,70 ) ( -37,-150 ) ( 50,70 ) ( -37,-150 ) one notices that the parameter @xmath205 determines the importance of the loop contribution relative to the tree - level - induced masses .
for example , from the right panel one sees that , below @xmath214 the heaviest neutrino mass @xmath215 is mainly a tree - level mass , while for @xmath216 the loop
induced masses are important relative to the tree
level one .
similar results are obtained for other choices of mssm parameters .
it is also interesting to analyse the dependence of the neutrino mass spectrum obtained in this model as a function of other supersymmetric parameters . in
we show the three lightest eigenvalues of the neutrino neutralino mass matrix as a function of @xmath74 , keeping the other parameters fixed as in , fixing @xmath217 . again in the right figure we have applied the sign condition discussed in more details below .
loop contributions are very strongly correlated with @xmath74 .
( 50,70 ) ( -37,-150 ) similarly one can compute the three lightest eigenvalues of the neutrino / neutralino mass matrix as a function of @xmath218 , as shown in .
larger @xmath218 leads to smaller loop masses , as expected . from - one
sees that , as expected , the pattern of neutrino masses obtained in the bilinear scenario , for almost all choices of parameters , is a _ hierarchical _ one .
( 50,70 ) ( -37,-150 ) in the above we have not paid attention to whether or not the parameter values used in the evaluation of the neutrino mass spectrum are indeed solutions of the minimization tadpole conditions of the higgs potential .
we now move to a more careful study of the magnitude of the neutrino mass spectrum derived in the scenario . in order to proceed further with the discussion of the solutions to the solar neutrino anomalies in this model we must distinguish two cases : 1 .
unified universal boundary conditions on the soft susy breaking terms ( sugra case , for short ) 2 .
non universal boundary conditions on the soft susy breaking terms ( mssm case , for short ) in what follows we refer to these two possibilities as sugra and mssm cases , accordingly . for the analysis of the neutrino _ masses _ these two scenarios are very similar
so we focus on the case where the low scale paramaters are derivable from a universal supergravity scheme . in
we show the mass squared difference @xmath219 which is relevant for the analysis of neutrino oscillations and therefore relevant to the interpretation of solar data , as a function of the parameter @xmath220 . in the left panel
we display @xmath221 while on the right panel @xmath222 .
small values prefer @xmath223 in the range of the vacuum solution to the solar neutrino problem , while large values give masses in the range of the msw solutions .
points shown in the following figures were obtained scanning the relevant parameters randomly over the region : @xmath224 and @xmath225 from 0 to 500 gev , @xmath218 [ 0.2 tev , 1.0 tev ] , @xmath226 and @xmath227 [ -3,3 ] and @xmath74 [ 2.5,10 ] , and for the parameters , @xmath228 , @xmath229 , @xmath230 , @xmath231 and @xmath232 @xmath211 .
they were subsequently tested for consistency with the minimization ( tadpole ) conditions of the higgs potential and for phenomenological constraints from supersymmetric particle searches .
( 50,70 ) ( -37,-150 ) one can also explicitly determine the attainable range of @xmath223 for which the corresponding @xmath233 ( see below ) lies in the range required for the correct interpretation of the atmospheric neutrino data .
the result obtained is displayed in in which we show @xmath223 as function of @xmath234 for those points which solve the atmospheric neutrino problem .
( 50,70 ) ( -30,-100 ) we now turn to the discussion of the three neutrino mixing angles and of how they must be identified in terms our our underlying parameters . following the usual convention the relation ^1l
= u_k _ k connecting mass - eigenstate and weak - eigenstate neutrinos are recovered in our notation as u_k= n^1l_4+k , 4 + where the mixing coefficients are determined numerically by diagonalizing the neutral fermion mass matrix .
note that , without loss of generality , in the bilinear model one can always choose as basis the one in which the charged lepton mass matrix is already diagonal .
the neutrino mixing angles relevant in the interpretation of solar and atmospheric data are identified as ( if @xmath235 , as indicated by the atmospheric data and the reactor neutrino constraints ) .
^2 ( 2 _ 13 ) = 4 u_3 ^ 2 ( 1 -u_3 ^ 2 ) ^2 ( 2 _ 12 ) = 4 u_e1 ^ 2 u_e2 ^ 2 ( 50,70 ) ( -30,-100 ) note that the maximality of the atmospheric angle is achieved for @xmath236 ( see ) and @xmath237 is smaller than the other two , as required by the chooz data ( see below ) .
in fact we have found @xcite , that if @xmath238 then the approximate formula holds u_3 _ /|| in we show the expected magnutide of @xmath239 versus the relevant ratio of parameters . in order to comply with the reactor data from the chooz experiment one should have @xmath239 below 0.05 .
this implies a bound on @xmath237 which can be read off from the figure .
( 50,70 ) ( -30,-100 ) the discussion on the solar mixing angle is more involved .
first note that it has no meaning before adding the one loop corrections to the neutrino mass , since in that limit the two low - lying neutrinos would be degenerate in mass . in order to proceed further with the discussion of the solutions to the solar neutrino problem in this model we must analyse carefully the implications of .
here it is important to distinguish between case 1 ( sugra ) and case 2 ( mssm ) discussed above . in the sugra case by taking the ratio of the first two equations in @xmath240 we conclude that , since @xmath241 and , since the relevant ratio of susy soft breaking terms is close to one , it follows that @xmath242 is small .
the predictions for the solar angle as a function of the breaking parameters is indicated in .
( 50,70 ) ( -30,-100 ) more precisely , the interpretation of the solar data @xcite in terms of the small angle msw solution indicates that ^2 ( 2 _ ) 10 ^ -3 - 10 ^ -2 and this in turn selects the required ratio of @xmath237 to @xmath243 and @xmath244 .
therefore in this case the large angle solutions , including the vacuum or just - so solutions do not fit in the scheme .
we now move to the general mssm case . in this case
the ratio of susy soft - breaking terms appearing in is in general arbitrary and thus the ratios of @xmath188 is no longer tied up to the ratios of @xmath200 s .
this opens up the possibility for large angle solutions to the solar neutrino problem . at first sight
it would seem that all predictivity of the solar angle is lost in this case , as seen in left panel of .
( 50,110 ) ( 0,-70 ) -5 cm the ability of our model to determine the solar neutrino angle may be understood in terms of .
for example in the sugra case we see from that the @xmath245 and @xmath246 ratios are fixed within a narrow range , leading to the small mixing angle prediction for the solution to the solar neutrino problems .
there is however another way to obtain predictivity for the general mssm case , namely by applying .
the possibility of our model predicting the solar angle even in the general mssm case by assuming can be understood as follows .
consider first the simplified limit @xmath247 . in this case
@xmath248 at tree - level and there is no mixing at all between the electron neutrino and the other two states , but a finite mixing exists at one - loop , due to the terms proportional to @xmath249 . in this case the sign condition , defined in introduces two more zeros into the matrix proportional to @xmath199 in above , if @xmath250 .
this fact simplifies the calculation of the solar angle very much , since one of the neutrino eigenvectors ( the one for @xmath9 ) has no dependence on the @xmath198 ratios but only on the @xmath17 ratios . for a non - zero @xmath237 ( and small departures from equality of @xmath251 or latexmath:[$|\lambda_{\mu}| , @xmath237 dependence reintroduced in the solar angle .
however , as long as the one - loop contributions are smaller than the tree - level one and as long as @xmath253 , the `` cross - talk '' between the @xmath237 and @xmath249 pieces is sufficiently small , such that some predictivity of the solar angle is retained , as illustrated in figure ( right panel ) .
the discussion on mixing angles may be summarized as follows . in the case that one - loop corrections are not larger than the tree - level contributions , the approximate formula u_3 _
/ | | holds .
this allows one to fix the atmospheric angle and at the same time obey the chooz constraint .
for the solar angle , however , the results depend on whether one wants to work in a sugra motivated scenario or not . for the sugra scenario
we have found that our model allows only the small mixing angle msw solution ( sma ) , while for the general case also lma and vacuum oscilation solutions are possible .
we have shown that the simplest unified extension of the minimal supersymmetric standard model with bi - linear r parity violation typically predicts a hierarchical neutrino mass spectrum , offering a natural theory for the solar and atmospheric neutrino anomalies . in this model
only one neutrino acquires mass due to mixing with neutralinos , while the other two get mass only as a result of radiative corrections .
we have performed a full one - loop calculation of the effective neutrino mass matrix in the bi - linear mssm , taking special care to achieve a manifestly gauge invariant calculation and performing the renormalization of the heaviest neutrino , needed in order to get reliable results .
the atmospheric mass scale and maximal mixing angle arise from tree - level physics , while the solar neutrino scale and oscillations follow from calculable one - loop corrections . under the assumption of universal boundary conditions for the soft - supersymmetry breaking terms at the unification scale
we find that the atmospheric scale is calculable by the renormalization group evolution due to the non - zero bottom quark yukawa coupling . in this case one
predicts the small mixing angle ( sma ) msw solution to be the only viable solution to the solar neutrino problem .
in contrast , for the general mssm model , where the above assumptions are relaxed , one can implement a _ bi - maximal _ @xcite neutrino mixing scheme , in which the solar neutrino problem is accounted for through large mixing angle solutions , either msw or just - so .
a great advantage of our approach is that the parameters required in order to solve the neutrino anomalies can be independently tested at high energy accelerators , as originally proposed in @xcite .
in fact , as shown in ref .
@xcite the bilinear model predicts the lightest supersymmetic particle ( lsp ) decay to be observable at high - energy colliders , since the expected decay path can easily be shorter the typical detector sizes .
this happens despite the smallness of neutrino masses indicated by the superkamiokande data .
this provides a way to test this solution of the atmospheric and solar neutrino anomalies and potentially discriminate between the large and small mixing solutions to the solar neutrino problem .
this work was supported by dgicyt grant pb98 - 0693 and by the eec under the tmr contract erbfmrx - ct96 - 0090 .
m.h . has been supported by the marie - curie program under grant no erbfmbict983000 and w.p . by a fellowship from the spanish ministry of culture under the contract sb97-bu0475382 .
m.a.d . was supported by conicyt grant no .
the mass matrix of the charged scalar sector follows from the quadratic terms in the scalar potential @xmath254 where the unrotated charged scalars are @xmath255 . for convenience
we will divide this @xmath256 matrix into blocks in the following way : @xmath257 + \xi m_w^2 \left[\matrix { { \bold m_{a}^2 } & { \bold m_{b}^{2 t } } \cr { \bold m_{b}^2 } & { \bold m_{c}^2 } } \right ] \label{eq : subdivm}\ ] ] where the charged higgs block is @xmath258 \label{eq : submhh } \end{aligned}\ ] ] this matrix reduces to the usual charged higgs mass matrix in the mssm when we set @xmath259 and we call @xmath260 .
the slepton block is given by @xmath261\end{aligned}\ ] ] where ( m^2_ll)_ij & = & v_d^2 ( h_e^ * h_e^t)_ij + g^2(-_k=1 ^ 3 v_k^2-v_d^2+v_u^2)_ij + g^2 v_i v_j - b_i _ i _ ij + _ ij & & + _ i _ ij -_i ( _ k=1 ^ 3 _ k ) _
ij + _ i _ j + m^2_l ji & & -_k=1 ^ 3 ( m^2_l ik+m^2_l ki)_ij = ( v_d a_e^ * -v_u h_e^ * ) = ( m^2_lr)^ @xmath262 we recover the usual stau mass matrix again by replacing @xmath259 ( note that we need to replace the expression of the tadpole @xmath263 in eq .
( [ eq : tadpoles ] ) before taking the limit ) .
the mixing between the charged higgs sector and the slepton sector is given by the following @xmath264 block ( repeated indices are not summed unless an explicit sum appears ) : @xmath265 \label{eq : subhtau}\ ] ] and as expected , this mixing vanishes in the limit @xmath259 .
the charged scalar mass matrix in eq .
( [ eq : subdivm ] ) , after setting @xmath266 , has determinant equal to zero for @xmath115 , since one of the eigenvectors corresponds to the charged goldstone boson with zero eigenvalue . for our one loop calculations one has to had the gauge fixing . the part of the mass matrix in eq .
( [ eq : subdivm ] ) that comes from the gauge fixing reads for the @xmath267 @xmath268 block @xmath269\ ] ] for the @xmath270 @xmath36 and the @xmath271 @xmath272 blocks @xmath273 \hskip 10 mm ; \hskip 10 mm { \bold m_{c}^2}= \left[\matrix { { \bold m_{d}^2}&0\cr \vb{18 } 0 & 0 } \right]\ ] ] where the @xmath274 @xmath41 block is @xmath275\ ] ] the charged scalar mass matrices are diagonalized by the following rotation matrices , @xmath276 with the eigenvalues @xmath277 .
the quadratic scalar potential includes @xmath278 { \bold m^2_{s^0}}\left[\matrix { \sigma^0_d \cr \sigma^0_u \cr \tilde\nu_{i}^r } \right]+ \cdots \label{eq : neutscallag}\ ] ] where the neutral cp - even scalar sector mass matrix in eq .
( [ eq : neutscallag ] ) is given by = where = [ eq : mss ] = and ( m__r _ r^2)_ij -10pt&=-10pt & ( _ iv_d - b_i_iv_u -_i _ k=1 ^ 3 _ k - _
k=1 ^ 3 ( m^2_l ik+m^2_l ki ) + t_i)_ij + g^2_zv_iv_j & & + _ i _ j + ( m^2_l ij+m^2_l ji ) where we have defined @xmath279 . in the upper left @xmath280 block , in the limit @xmath259 , the reader can recognize the mssm mass matrix corresponding to the cp even neutral higgs sector . to define the rotation matrices
let us define the unrotated fields by s^0=(_d^0,_u^0,_1^r , _ 2^r , _
2^r ) then the mass eigenstates are @xmath281 given by @xmath282 with the eigenvalues @xmath283 .
the quadratic scalar potential includes @xmath284 { \bold m^2_{p^0}}\left[\matrix { \varphi^0_1 \cr \varphi^0_2 \cr \tilde\nu_{i}^i } \right ] + \cdots \label{eq : pseudoscallag}\ ] ] where the cp - odd neutral scalar mass matrix is = + m_z^2 [ eq : mp0 ] where = [ eq : mpp ] = and ( m__i _ i^2)_ij&= & ( _ iv_d - b_i_iv_u -_i _ k=1 ^ 3 _ k - _
k=1 ^ 3 ( m^2_l ik+m^2_l ki ) + t_i)_ij & & + _ i _ j + ( m^2_l ij+m^2_l ji ) finally the part of the mass matrix in eq . ( [ eq : mp0 ] ) that comes from the gauge fixing reads for the @xmath267 @xmath285 block @xmath286\ ] ] for the @xmath287 @xmath288 block @xmath289\ ] ] and for the @xmath290 @xmath291 block @xmath292\ ] ] the charged pseudo scalar mass matrices are diagonalized by the following rotation matrices , @xmath293 with the eigenvalues @xmath294 . where the unrotated fields are p^0=(_d^0,_u^0,_1^i , _ 2^i , _
2^i ) in the unrotated basis @xmath295 and @xmath296 we get v_quadratic=^ ^2
+ ^ ^2 where ^2= ( ) [ eq : masssquarks ] with @xmath297 .
the blocks are different for up and down type squarks .
we have & = & v_u^2 h_u^ * h_u^t + m^2_q + ( 4 m_w^2 -m_z^2 ) 2 & = & v_u^2 h_u^t h_u^ * + m^2_u + ( m_z^2 -m_w^2 ) 2 & = & a_u^ * - h_u^ * + _
i=1 ^ 3 _ i h_u^ * & = & ^ and & = & v_d^2 h_d^ * h_d^t + m^2_q - ( 2 m_w^2 + m_z^2 ) 2 & = & v_d^2 h_d^t h_d^ * + m^2_d - ( m_z^2 -m_w^2 ) 2 & = & a_d^ * - h_d^ * & = & ^ we define the mass eigenstates = which implies _i=_ji^ _ j the rotation matrices are obtained from ^ ( ) ^2 = ^2 in our case the matrices in are real and therefore the rotation matrices @xmath298 are orthogonal matrices . the charginos mix with the charged leptons forming a set of five charged fermions @xmath299 , @xmath300 in two component spinor notation . in a basis where @xmath301 and @xmath302 , the charged fermion mass terms in the lagrangian are @xmath303 where the chargino / lepton mass matrix is given by @xmath304 \label{eq : cham6x6}\ ] ] and @xmath305 is the @xmath306 gaugino soft mass .
we note that chargino sector decouples from the lepton sector in the limit @xmath32 . as in the mssm
, the chargino mass matrix is diagonalized by two rotation matrices @xmath307 and @xmath308 defined by f_i^-=u_ij _ j^- ; f_i^+=v_ij _ j^+ then ^ * ^-1 = where @xmath309 is the diagonal charged fermion mass matrix . to determine @xmath310 and @xmath311 we note that = ^-1 = ( )
^-1 implying that @xmath311 diagonalizes @xmath312 and @xmath313 diagonalizes @xmath314 . for future reference
we note that ^-_j=_kj f_k^- ; ^+_j=_kj f_k^+ in the previous expressions the @xmath299 are two component spinors .
we construct the four component dirac spinors out of the two component spinors with the conventions , @xmath315 and @xmath316 to be the particles and not the anti particles .
] , _ i^-= ( )
using four component spinor notation the relevant part of the lagrangian can be written as -15pt l&= & ^ ( o^cnw_l ij p_l + o^cnw_rij p_r ) _
j^0 w_^- + ^ ( o^ncw_l ij p_l + o^ncw_lij p_r ) _ j^- w_^+ & & + ( o^cns_l ijk p_l + o^cns_rijk p_r ) _ j^0 s_k^- + ( o^ncs_l jik p_l + o^ncs_ljik p_r ) _ j^- s_k^+ & & + ^ ( o^nnz_lij p_l + o^nnz_rij p_r ) _ j^0 z_^0 + ( o^nnh_l ijk p_l + o^nnh_r ijk p_r ) _ j^0 h_k^0 & & + i ( o^nna_l ijk p_l + o^nna_r ijk p_r ) _ j^0 a_k^0 & & + ( o^qns_lijk p_l + o^qns_rijk p_r ) _
j^0 q_k + ( o^nqs_lijk p_l + o^nqs_rijk p_r ) q_j q_k^ * where @xmath317 can be either @xmath155 or @xmath148 . the various couplings are : o^cnw_l ij= g _ i _
j o^cnw_r ij = g ( - _ j2 _ i1 + _ j4 _ i2 ) o^ncw_l ij= ( o^cnw_l ji ) ^ * ; o^ncw_r ij= ( o^cnw_r ji ) ^ * o^nnz_l ij= _ i _ j ( _ i4 _ j4 -_i3 _
j3 -_k=1 ^ 3 _ i,4+k _ j,4+k ) o^nnz_rij=- ( _ i4 _ j4 -_i3 _
j3 -_k=1 ^ 3 _ i,4+k _ j,4+k ) o^cns_l ijk= _
j o^cns_r ijk=_i o^ncs_l ijk= ( o^ncs_r jik ) ^ * ; o^cns_r ijk= ( o^ncs_l jik ) ^ * o^nnh_l ijk=_j o^nnh_r ijk= ( o^nnh_l jik ) ^ * o^nna_l ijk= - _ j o^nna_r ijk=- ( o^nna_l jik ) ^ * the factors @xmath318 are the signs one has to include if we consider @xmath319 , @xmath320 and @xmath321 as real matrices and the mass of the fermion @xmath48 is negative .
o^uns_l ijk = ( ) _ w n^*_j1 r^u^*_k , m+3 r^u_ri , m - ( h_u)_mlr^u^*_k , m r^u_ri , ln^*_j4 o^uns_r ijk = - ( ) ( n_j2 + _ w n_j1 ) r^u^*_k , m r^*^u_lm , i - ( h_u^*)_mlr^u^*_k+3,l r^*^u_lm , in_j4 and , o^nus_l ijk= ( o^uns_r jik ) ^ * ; o^nus_r ijk= ( o^uns_l jik ) ^ * o^dns_l ijk = - ( ) _ w n^*_j1 r^d^*_k , m+3 r^d_ri , m - ( h_d)_mlr^d^*_k , m r^d_ri , ln^*_j3 o^dns_r ijk = ( ) ( n_j2 - _ w n_j1 ) r^d^*_k , m r^*^d_lm , i - ( h_d^*)_mlr^d^*_k , l+3 r^*^d_lm , in_j3 and , o^nds_l ijk= ( o^dns_r jik ) ^ * ; o^nds_r ijk= ( o^dns_l jik ) ^ * to evaluate the tadpoles we need the couplings of the neutral scalars with all the fields in the model .
these couplings are easier to write in the unrotated basis .
the couplings for the mass eigenstates can always be obtained by appropriate multiplication by the rotation matrices . as an example , and to fix the notation ( repeated indices are understood to be summed unless otherwise stated ) , the couplings of three neutral scalars in the two basis will be related by g_ijk^s^0s^0s^0=r^s^0_ip r^s^0_jq r^s^0_kr g_pqr^s^0s^0s^0 sometimes we will also use partially rotated couplings , for instance g_ijk^s^0s^0s^0= r^s^0_jq r^s^0_kr g_iqr^s^0s^0s^0 in an obvious notation .
these couplings are defined as follows g_ijk^s^0s^0s^0&= & g_ijk^s^0s^0s^0&= & g_ijk^s^0s^0s^0=- ( g^2+g^2 ) u_m ( _ mi _ jk + _ mj _ ik + _ mk _ ij ) where we have defined u_m ( v_d , v_u , v_1,v_2,v_3 ) ; _ ij(+,-,+,+,+ ) for future reference we also define v_m(v_1,v_2,v_3 ) while @xmath322 without the _ hat _ is the usual kronecker delta .
g_ijk^s^0p^0p^0=- ( g^2+g^2 ) u_m _ mi _ jk we define g_ijk^s^0s^+s^-= ( ) where ( g_ihh^s^0s^+s^-)_jk&= & g^2 & & - g^2 u_m _ i m _ jk & & - v_m _ j1 _ k1 ( h_e h_e^ + h_e^ * h_e^t ) _ m , i-2 ( g_ihl^s^0s^+s^-)_jk&= & - g^2 + v_m ( h_e^ * h_e^t ) _ mk _ i1 _ j1 & & + v_d _ j1 ( h_e^ * h_e^t ) _ i-2,k ( g_ihr^s^0s^+s^-)_jk&= & _ m ( h_e^*)_mk ( _ i1 _ j2 +
j1 ) + ( a_e^*)_i-2,k _ j1 & & + ( h_e^*)_i-2,k _ j2 ( g_ill^s^0s^+s^-)_jk&= & ( g^2-g^2 ) u_m _ i m _ jk - g^2 v_m ( _ i-2,j _ mk + _ i-2,k _ mj ) & & -(h_e^*h_e^t)_jk v_d _ i1 ( g_ilr^s^0s^+s^-)_jk= - _ i1 ( a_e^*)_jk + _ i2 ( h_e^*)_jk ( g_irr^s^0s^+s^-)_jk&= & g^2 u_m _ i m _
jk -v_d _ i1 ( h_e^th_e^*))_jk & & - v_m with the definition = g_ijk^s^0^ * s^0_i _j ^*_k + we get g_ijk^s^0^*= ( ) where g_ill^s^0^ * & = & u_m _ i m ( - g^2 + g^2 ) i -v_u ( h_u h_u^ ) _ i2 g_ilr^s^0^
* & = & - _ i2 a_u + h_u _ i1 - h_u _ m _ i-2,m g_irl^s^0^ * & = & g_ilr^h^0^ * g_irr^s^0^ * & = & - u_m _ i m g^2 i -v_u ( h_u^t h_u^ * ) _ i2 where @xmath323 is the unit @xmath15 matrix . with the definition = g_ijk^s^0^ * s^0_i _j ^*_k + we get g_ijk^s^0^*= ( ) where g_ill^s^0^ * & = & u_m _ i m
( g^2 + g^2 ) i -v_d ( h_d h_d^ ) _
i1 g_ilr^s^0^ * & = & - _ i1 a_d + h_d _
i2 g_irl^s^0^ * & = & g_ilr^s^0^ * g_irr^s^0^ * & = & u_m _ i m g^2 i -v_d ( h_d^t h_d^ * ) _ i1 with the definition = g_i^s^0w^+w^- s^0_i w^+ w^- + we get g_i^s^0w^+w^-= g ( v_d _ i1+v_u _
i-2,m ) [ wcoupl ] where v= with the definition = g_i^s^0z^0z^0 s^0_i z^0 z^0 + we get g_i^s^0z^0z^0= ( v_d _ i1+v_u _ i2 + v_m _
i-2,m ) [ zcoupl ] with the definition = g_ijk^s^0u s^0_i _ j u_k + g_ijk^s^0d s^0_i _
j d_k + we get g_ijk^s^0u=- ( h_u)_jk _ i2 and g_ijk^s^0d=- ( h_d)_jk _ i1 with the definition = ( o_lijk^cch + o_rijk^cch ) _ j^- s^0_i + ( o_lijk^nnh + o_rijk^nnh ) _
j^0 s^0_i we have o^cch_l ijk&= & - o^cch_r ijk&=&(o^cch_l jik)^ * and o^nnh_l ijk&=&_j ( -g n^*_i2 n^*_j3 + g n^*_i1 n^*_j3 -g n^*_j2 n^*_i3 + g n^*_j1 n^*_i3 ) & & ( _ k1-_k2+_k3+_k4 + _ k5 ) o^nnh_r ijk&=&(o^nnh_l jik)^ *
we will consider the gauge boson and ghost tadpoles in an arbitrary @xmath128 gauge to show that the dependence on @xmath118 cancels out . we will do it for any model .
we write down the tadpole contribution from the @xmath326 for a general theory with the coupling @xmath327 to the higgs boson .
@xmath328 ( 0,0 ) ( 0,-0.5 ) where g_^=(-i ) and the factor @xmath329 is a symmetry factor .
now we do some transformations in the second term of the @xmath326 propagator @xmath330 , g()&= & ( 1- ) & = & - and therefore we can write g_^=(-i ) then it_z&= & i g_hzz ( -i ) & = & g_hzz where we have used the definition a_0(m^2 ) as @xmath331 grows for large @xmath118 as @xmath332 we conclude that @xmath333 grows like @xmath334 .
this dependence has to cancel against other diagrams .
it is easy to realize that the goldstone of the @xmath326 will not do it because , although its mass depend on @xmath118 , its contribution to the tadpole will only grow like @xmath118 because its coupling to @xmath335 does not depend on @xmath118 .
but the ghost coupling to @xmath335 does depend @xmath118 as we will see .
let us then calculate the tadpole of the ghost of the @xmath326 .
we have @xmath336 ( 0,0 ) ( 0,-0.5 ) where the factor @xmath337 is because of the anti - commutative properties of the ghosts . using the definition of @xmath181
we get i t_c_z= g_hc_z_z a_0(m_z^2 ) adding the two contributions together we obtain it_z + it_c_z= we see that for the @xmath118 dependence to cancel one must have g_hzz + g_hc_z_z = 0 [ zcancel ] as we will show below this is true for the sm , mssm and also for the bilinear r - parity model .
then the contribution from the @xmath326 and neutral ghost tadpoles is , for any model , gauge independent and given by it_z + it_c_z= g_hzz a_0(m_z^2 ) [ tadz ] the calculation for the @xmath338 boson and charged ghosts is very similar .
the main differences are that the @xmath27 tadpole does not have a factor @xmath329 and that there are _ two _ ghosts for the @xmath338
. therefore we have it_w + it_c^+_w+it_c^-_w & = & we see that the @xmath118 dependence will cancel out if g_hww + g_hc^+_w_w + g_hc^-_w_w = 0 [ wcancel ] we will show below that this is true in general . then the contribution from the @xmath338 and charged ghost tadpoles is , for any model , gauge independent and given by it_w + it_c^+_w+it_c^-_w = 3 g_hww a_0(m_w^2 ) [ tadw ] now lets us see us the cancellation occurs in the standard model ( sm ) .
the relevant couplings for the @xmath326 are g_hzz&= & m_z g_hc_z_z&= & - m_z and we immediatly see that eq.([zcancel ] ) is verified . for the @xmath338 we have g_hww&= & g m_w g_hc^+_w_w&= & - m_w g_hc^-_w_w&= & - m_w satisfying eq.([wcancel ] ) . in the bilinear r - parity model
the relevant couplings are g_i^s^0c^+_w_w&=&- ( v_d _ i1+v_u _ i2 + v_m _ i-2,m ) g_i^s^0c^-_w_w&=&- ( v_d _ i1+v_u _ i2 + v_m _ i-2,m ) [ wghost ] and g_i^s^0c_z_z=- ( v_d _ i1+v_u _ i2 + v_m _ i-2,m ) [ zghost ] then using eqs.([wcoupl],[zcoupl],[wghost],[zghost ] ) in eqs.([wcancel],[zcancel ] ) we see that the same cancellation occurs .
after showing the gauge invariance of the gauge boson tadpoles together with their ghosts we give now the general tadpole in a compact form .
we will write them for the unrotated neutral higgs @xmath340 because that is what is needed for substitution into eq.([eq : tadpoles ] ) .
the general form can be written as ( @xmath341 ) , t_h_i^0^x= p^x_i where p_i^w -2mm&=&-2 mm 3 g_i^s^0w^+w^- a_0(m_w^2 ) p_i^z -2mm&=&-2 mm g_i^s^0z^0z^0 a_0(m_z^2 ) p_i^s^ -2mm&=&-2 mm - _
k=1 ^ 8^ g_ikk^s^0s^+s^- a_0(m_k^2 ) p_i^s^0 -2mm&=&-2 mm - _
k=1 ^ 5 g_ikk^s^0s^0s^0 a_0(m_k^2 ) p_i^p^0 -2mm&=&-2 mm - _
k=1 ^ 5^ g_ikk^s^0p^0p^0 a_0(m_k^2 ) p_i^ -2mm&=&-2 mm - _
k=1 ^ 6 3 g_ikk^s^0^ * a_0(m_k^2 ) p_i^ -2mm&=&-2 mm - _
k=1 ^ 6 3 g_ikk^s^0^ * a_0(m_k^2 ) p_i^^ -2mm&=&-2 mm - _
k=1 ^ 5 ( -1 ) 4 m_k a_0(m_k^2 ) p_i^^0 -2mm&=&-2 mm - _
k=1 ^ 7 ( - ) 4 m_k a_0(m_k^2 ) p_i^u -2mm&=&-2 mm - _
k=1 ^ 3 ( -3 ) g_ikk^h^0u 4 m_k a_0(m_k^2 ) p_i^d -2mm&=&-2 mm - _
k=1 ^ 3 ( -3 ) g_ikk^h^0d 4 m_k a_0(m_k^2 ) where @xmath342 means that we sum over all fields _ except _ for the goldstone boson .
as explained in section [ gauinv ] the contribution of the goldstones is added to the self energies to achieve gauge invariance .
in this section we write down the contribution of the several self energy diagrams in the @xmath127 gauge . ^v_ij&= & - _ k f_ijk^x b_1(p^2,m^2_k , m^2_x ) ^v_ij&= & - _ k g_ijk^x m_k b_0(p^2,m^2_k , m^2_x ) with f^w_ijk&= & 2 ( o^ncw_l jk o^cnw_l ki + o^ncw_r jk o^cnw_r ki ) g^w_ijk&= & -4 ( o^ncw_l jk o^cnw_r ki + o^ncw_r jk o^cnw_l ki ) and f^z_ijk&= & ( o^nnz_l jk o^nnz_l ki + o^nnz_r jk o^nnz_r ki ) g^z_ijk&= & -2 ( o^nnz_l jk o^nnz_r ki + o^nnz_r jk o^nnz_l ki ) ^v_ij&= & - _ r _ k f^x_ijkr b_1(p^2,m^2_k , m^2_r ) ^v_ij&= & - _ r _ k
g^x_ijkr m_k b_0(p^2,m^2_k , m^2_r ) with f^s^_ijkr-1mm&= & -1 mm ( o^ncs_r jkr o^cns_l kir + o^ncs_l jkr o^cns_r kir ) g^s^_ijkr-1mm&= & -1 mm ( o^ncs_l jkr o^cns_l kir + o^ncs_r jkr o^cns_r kir ) f^s^0_ijkr-1mm&= & -1 mm ( o^nnh_l jkr o^nnh_l kir + o^nnh_r jkr o^nnh_r kir ) g^s^0_ijkr-1mm&= & -1 mm ( o^nnh_l jkr o^nnh_r kir + o^nnh_r jkr o^nnh_l kir ) f^p^0_ijkr-1mm&= & -1 mm - ( o^nna_r jkr o^nna_l kir + o^nna_l jkr o^nna_r kir ) g^p^0_ijkr-1mm&= & -1 mm -(o^nna_l jkr o^nna_l kir + o^nna_r jkr o^nna_r kir ) f^_ijkr-1mm&= & -1 mm ( o^nus_r jkr o^uns_l kir + o^nus_l jkr o^uns_r kir ) g^_ijkr-1mm&= & -1 mm ( o^nus_l jkr o^uns_l kir + o^nus_r jkr o^uns_r kir ) f^_ijkr-1mm&= & -1 mm ( o^nds_r jkr o^dns_l kir + o^nds_l jkr o^dns_r kir ) g^_ijkr-1mm&= & -1 mm ( o^nds_l jkr o^dns_l kir + o^nds_r jkr o^dns_r kir ) both solar and atmospheric neutrino data may be accounted for by massless fc neutrino interactions .
for example the atmospheric data have been recently considered in n. fornengo , m. c. gonzalez - garcia and j. w. f. valle , hep - ph/9906539 [ to be published in jhep ] ; m. c. gonzalez - garcia _ et al .
_ , phys .
* 82 * , 3202 ( 1999 ) [ hep - ph/9809531 ] .
neutrino decay may also play a role in the explanation of atmospheric neutrinos , e.g. v. barger , j. g. learned , p. lipari , m. lusignoli , s. pakvasa and t. j. weiler , phys .
* b462 * , 109 ( 1999 ) [ hep - ph/9907421 ] for recent global fits of atmospheric neutrino data see , for example , n. fornengo , m. c. gonzalez - garcia and j. w. f. valle , hep - ph/0002147 [ to be published in nucl .
* b * ] and m.c .
gonzalez - garcia , h. nunokawa , o.l .
peres and j. w. f. valle , nucl .
* b543 * , 3 ( 1999 ) ; g. l. fogli , e. lisi , a. marrone and g. scioscia , `` combined analysis of atmospheric neutrino results , '' _ in * venice 1999 , neutrino telescopes , vol . 1 * 275 - 282 _ ; r. foot , r. r. volkas and o. yasuda , phys .
* d58 * , 013006 ( 1998 ) the global msw discussion of the solar neutrino data is given in m.c .
gonzalez - garcia , p.c .
de holanda , c. pea - garay and j. w. f. valle , nucl .
* b573 * , 3 ( 2000 ) [ hep - ph/9906469 ] .
for another recent discussion see also s. goswami , d. majumdar and a. raychaudhuri , hep - ph/0003163 . for prospects see j. n. bahcall , p. i. krastev and a. y. smirnov , hep - ph/0002293 and g. l. fogli , e. lisi , d. montanino and a. palazzo , phys .
rev . * d61 * , 073009 ( 2000 ) .
see , for example , s. lola and j.d .
vergados , prog . part .
phys . * 40 * , 71 ( 1998 ) ; g. altarelli and f. feruglio , phys . lett . * b439 * , 112 ( 1998 ) ; g. altarelli and f. feruglio , phys . lett .
* b451 * ( 1999 ) 388 hep - ph/9812475 ; s. lola and g.g .
ross , hep - ph/9902283 ; r. barbieri , l.j .
hall and a. strumia , phys .
b445 * , 407 ( 1999 ) , hep - ph/9808333 ; m.e .
gomez , g.k .
leontaris , s. lola and j.d .
vergados , phys . rev . *
d59 * , 116009 ( 1999 ) , hep - ph/9810291 ; g.k .
leontaris , s. lola , c. scheich and j.d .
vergados , phys . rev .
* d53 * , 6381 ( 1996 ) ; c. h. albright , k. s. babu and s. m. barr , phys .
lett . * 81 * , 1167 ( 1998 ) [ hep - ph/9802314 ] .
daz , j.c .
romo and j. w. f. valle , nucl .
phys . * b524 * , 23 ( 1998 ) [ hep - ph/9706315 ]
. for reviews see , e.g. j. w. f. valle , `` super - gravity unification with bilinear r - parity violation , '' proceedings of 6th international symposium on particles , strings and cosmology ( pascos 98 ) , boston , ma , 22 - 27 mar 1998 [ hep - ph/9808292 ] ed .
p. nath ; m.a .
daz , talk given at _
international europhysics conference on high energy physics _ , jerusalem , israel , 19 - 26 aug 1997 , f. de campos , m.a .
garca - jareo , a.s .
joshipura , j. rosiek , and j. w. f. valle , ; t. banks , y. grossman , e. nardi , and y. nir , ; a. s. joshipura and m.nowakowski , ; h. p. nilles and n. polonsky , ; b. de carlos , p. l. white , ; s. roy and b. mukhopadhyaya , j. schechter and j. w. f. valle , phys .
rev . * d22 * , 2227 ( 1980 ) . g. g. ross and j. w. f. valle , phys
. lett .
* b151 * , 375 ( 1985 ) ; j. ellis , g. gelmini , c. jarlskog , g. g. ross and j. w. f. valle , phys .
* b150 * , 142 ( 1985 ) ; c. s. aulakh and r. n. mohapatra , phys .
* b121 * , 14 ( 1983 ) for some recent references on bilinear r - parity violation , see f. takayama and m. yamaguchi , phys .
lett . * b476 * , 116 ( 2000 ) [ hep - ph/9910320 ] ; d. e. kaplan and a. e. nelson , jhep * 0001 * , 033 ( 2000 ) ; e. j. chun and s. k. kang , phys .
rev . * d61 * , 075012 ( 2000 ) ; a. datta , b. mukhopadhyaya and s. roy , phys . rev . * d61 * , 055006 ( 2000 ) ; a. s. joshipura and s. k. vempati , phys .
* d60 * , 111303 ( 1999 ) ; e.j .
chun , s. k. kang , c. w. kim and u. w. lee , nucl .
* b544 * , 89 ( 1999 ) [ hep - ph/9807327 ] ; e. nardi , phys .
* d55 * , 5772 ( 1997 ) [ hep - ph/9610540 ] ; f. m. borzumati , y. grossman , e. nardi and y. nir , phys .
* b384 * , 123 ( 1996 ) [ hep - ph/9606251 ] .
a. masiero and j. w. f. valle , ; j.c .
romo , c.a .
santos , j. w. f. valle , ; j. c. romao , a. ioannisian and j. w. valle , phys .
* d55 * , 427 ( 1997 ) [ hep - ph/9607401 ] ; for a review see j.c .
romo , proceedings of _ international workshop on physics beyond the standard model : from theory to experiment _ , valencia , 13 - 17 oct 1997 , ed .
v. berezinsy , g. raffelt , j. w. f. valle , nuclear physics b , proceedings supplements , 2000,vol . 81 ,
issn 0920 - 5632 for papers discussing r - parity violation phenomenology see , e.g. f. de campos , o. j. eboli , m. a. garcia - jareno and j. w. valle , nucl .
* b546 * , 33 ( 1999 ) [ hep - ph/9710545 ] ; m. nowakowski and a. pilaftsis , nucl .
phys . *
b461 * , 19 ( 1996 ) [ hep - ph/9508271 ] ; l. navarro , w. porod and j. w. valle , phys . lett .
* b459 * , 615 ( 1999 ) [ hep - ph/9903474 ] ; m. a. diaz , d. a. restrepo and j. w. valle , hep - ph/9908286 [ to be published in nucl .
* b * ] ; b. allanach _ et al .
_ , hep - ph/9906224 ; a. faessler , s. kovalenko and f. simkovic , phys .
rev . * d58 * , 055004 ( 1998 ) [ hep - ph/9712535 ] ; s. davidson , m. losada and n. rius , hep - ph/9911317 ; m. bisset , o. c. kong , c. macesanu and l. h. orr , hep - ph/9811498 ; m. a. diaz , e. torrente - lujan and j. w. f. valle , nucl .
* b551 * , 78 ( 1999 ) [ hep - ph/9808412 ] ; | the simplest unified extension of the minimal supersymmetric standard model with bi - linear r parity violation naturally predicts a hierarchical neutrino mass spectrum , in which one neutrino acquires mass by mixing with neutralinos , while the other two get mass radiatively . we have performed a full one - loop calculation of the neutralino - neutrino mass matrix in the bi - linear mssm , taking special care to achieve a manifestly gauge invariant calculation .
moreover we have performed the renormalization of the heaviest neutrino , needed in order to get meaningful results .
the atmospheric mass scale and maximal mixing angle arise from tree - level physics , while solar neutrino scale and oscillations follow from calculable one - loop corrections .
if universal supergravity assumptions are made on the soft - supersymmetry breaking terms then the atmospheric scale is calculable as a function of a single violating parameter by the renormalization group evolution due to the non - zero bottom quark yukawa coupling .
the solar neutrino problem must be accounted for by the small mixing angle ( sma ) msw solution .
if these assumptions are relaxed then one can implement large mixing angle solutions , either msw or just - so .
the theory predicts the lightest supersymmetic particle ( lsp ) decay to be observable at high - energy colliders , despite the smallness of neutrino masses indicated by experiment .
this provides an independent way to test this solution of the atmospheric and solar neutrino anomalies .
# 1#2 # 1fig .
( [ # 1 ] ) # 1#2 -2 cm 24.4 cm /#1#1 / # 1eq .
( [ # 1 ] ) # 1eq . ( [ # 1 ] ) # 1fig .
( [ # 1 ] ) # 1to # 1 pt # 1#2#3_eur .
phys .
j. _ * c # 1 * ( # 2 ) # 3 # 1#2#3_nucl .
phys . _
* b # 1 * ( # 2 ) # 3 # 1#2#3_phys .
lett . _
* b # 1 * ( # 2 ) # 3 # 1#2#3_phys .
rev . _
* d # 1 * ( # 2 ) # 3 # 1#2#3_phys .
rep . _ * # 1 * ( # 2 ) # 3 # 1#2#3_phys .
rev .
lett . _ * # 1 * ( # 2 ) # 3 # 1#2#3_mod .
phys .
lett . _ * a # 1 * ( # 2 ) # 3 # 1#2#3_sov .
j. nucl .
phys . _ * # 1 * ( # 2 ) # 3 # 1#2#3_sov .
phys .
jetp lett .
_ * # 1 * ( # 2 ) # 3 # 1#2#3_riv .
nuovo cimento _ * # 1 * ( # 2 ) # 3 # 1#2#3_yad . fiz .
_ * # 1 * ( # 2 ) # 3 # 1hep - ph/#1 1@xmath0 # 10= -.025em0 - 0 .05em0 - 0 -.025em.0433em0 hep - ph/0004115 + ftuv-00 - 0412 + ific/00 - 24 + * neutrino masses and mixings from supersymmetry with bilinear r
parity violation : * * a theory for solar and atmospheric neutrino oscillations * + m. hirsch@xmath1 , m. a. daz@xmath2 , w. porod@xmath1 , j. c. romo@xmath3 and j. w. f. valle@xmath1 + _ @xmath4 departament de fsica terica , ific
csic , universitat de valncia + 46100 burjassot , valncia , spain + _ _ @xmath5 facultad de fsica , universidad catlica de chile + av .
vicua mackenna 4860 , santiago , chile + _ _ @xmath6 departamento de fsica , instituto superior tcnico + av .
rovisco pais 1 , @xmath7 1049 - 001 lisboa , portugal + _ |
recently , kharzeev , pisarski and tytgat @xcite presented the following idea , based on the nontrivial topological structure of the qcd - vacuum @xcite : metastable states which act like regions with a non - vanishing qcd vacuum angle @xmath4 may be excited when matter undergoes the deconfining phase transition , provided it is of second order . in these false vacua , parity ( and also cp ) is spontaneously broken , a quality that could lead to experimental signatures like parity odd correlations of the produced particles . in the present work
, we investigate the production of @xmath0-particles during the decay of the cp - odd metastable states , concentrating on the amplification of the low momentum modes by parametric resonance when the zero mode rolls down from the false to the true vacuum .
this mechanism plays an important role for particle production in inflationary cosmology @xcite and has also been investigated for the formation of disoriented chiral condensates @xcite and for axions @xcite .
here we discuss the efficiency of the amplification mechanism for the production of @xmath5 from the decaying metastable bubble . as a main result we derive the corresponding momentum spectrum showing non - thermal behaviour .
the integrated density of particles produced is estimated to be @xmath6 .
the idea is based on the effective lagrangian of the witten - divecchia - veneziano model @xcite @xmath7,\end{aligned}\ ] ] which describes the low - energy dynamics of the pseudoscalar mesons in the large @xmath8-limit of qcd . in the following ,
we consider the `` real world''-vacuum angle @xmath9 .
formation and decay of a non - zero @xmath4-vacuum and its signatures are investigated in @xcite .
the @xmath10-matrix @xmath11 in eq .
( [ lagr ] ) describes the meson fields , and the term containing the mass matrix @xmath12 provides the explicit soft breaking of chiral symmetry due to the quark masses .
@xmath12 can be written as @xmath13 , where the diagonal elements are chosen as @xmath14 and @xmath15 . with the parametrization @xmath16
, the matrix @xmath17 representing the singlet and the octet meson fields reads @xmath18 the last term in the effective lagrangian reflects the @xmath19-anomaly by giving a mass to the singlet , even in the chiral limit of vanishing quark masses .
the parameter @xmath20 represents the topological susceptibility .
finite temperature is introduced into the model via the temperature dependence of the parameters : for @xmath21 , the value @xmath22 can be determined from experimental results @xcite , and @xmath23 .
the corresponding value of the pion decay constant is @xmath24 .
when the temperature approaches the temperature @xmath25 of the phase transition , @xmath26 goes to zero , indicating the effective restoration of the @xmath19-symmetry for @xmath27 with enhanced @xmath5 - and @xmath0 - production as a possible signature @xcite .
we use the temperature of the deconfinement phase transition , @xmath25 , following the assumption of @xcite that any other phase transition of this model also occurs at @xmath25 . according to mean field estimates @xcite , @xmath28 and @xmath29 near the phase transition . in order to get an analytically tractable approximation that allows some insight into the physical processes , we consider only the singlet , which is the main component of the @xmath0-meson . using @xmath30
, the effective lagrangian simplifies to @xmath31 the singlet effective potential @xmath32 where we have dropped the index @xmath33 in the notation of the singlet and introduced @xmath34 for convenience , can be considered as a projection of the full effective potential .
it takes three qualitatively different shapes , depending on the temperature , more specifically , on the value of @xmath35 . for _
large _ values of @xmath36 , corresponding to _ small _
@xmath37 , the potential is a more or less deformed parabola ( see fig . [
fig : pot]a ) . for @xmath38 ,
i.e. @xmath27 , the potential is periodic with minima at @xmath39 , which are all equivalent true vacua ( see fig . [
fig : pot]d ) . and
for a _ small , nonzero _ value of @xmath36 , corresponding to a temperature just below the phase transition , metastable states appear in the potential which are distinct from the true vacuum ( see fig . [
fig : pot]c ) .
one recognizes that the false vacua are odd under parity transformation , but even under charge conjugation , and therefore odd under cp @xcite .
one can show that the effective potential exhibits metastable states also in the case of nonvanishing @xmath40fields , provided that @xmath41 @xcite .
but only the singlet is responsible for the formation of these states , in the sense that it is the only field which appears in the term of the effective potential which includes the topological susceptibility .
applying the model to the description of a heavy ion collision , we suppose that during the phase transition , the potential changes its shape from the parabola ( fig .
[ fig : pot]a ) , corresponding to the low temperature phase , to that of the high temperature phase ( fig .
[ fig : pot]d ) , and back to the low temperature phase , but this time including metastable states that have been created at a temperature just below the phase transition .
after the phase transition , but still at high temperature , the @xmath42field may be trapped in a metastable state , forming a `` cp - odd bubble inside the hadronic phase '' in the language of @xcite .
considering the effective potential as a function of temperature , one recognizes that there exists a value @xmath43 for which the local minimum turns into a saddle point ( see fig . [
fig : pot]b ) . for the lowest false vacuum
, this happens for @xmath44 at the value @xmath45 . assuming the proportionality @xmath46 to hold even far from @xmath25 , the temperature corresponding to this special case of the potential can be estimated to be @xmath47 .
once the saddle point temperature is reached and the barrier has disappeared , the @xmath42field starts rolling down into the true vacuum and oscillating around it , while energy is transferred from the mean value of the field into its quantum fluctuations .
the resulting growth of the occupation numbers of the quantum fluctuations is interpreted as particle production with respect to the true , cp even , vacuum , whereas the zero mode is damped .
this process is investigated in more detail in the following .
a crucial assumption for this picture is that at least after the saddle point temperature has been reached , the effective potential changes slowly compared to the motion of the field , so that the oscillations can be considered as taking place in a static potential , frozen at @xmath48 .
accordingly , the ratio @xmath49 is kept fixed during the evolution in time , and the value of @xmath50 as well .
the solutions we find ( see below ) show that the zero mode reaches the true vacuum for the first time after @xmath51 , whereas the relevant time scale to be considered for the particle production is of the order of @xmath52 . to study the dynamics of the decay process , we decompose the field into its expectation value @xmath53 and fluctuations @xmath54 about it , @xmath55 with @xmath56 , so that the `` physical '' fluctuations are given by the usual expression @xmath57 . with this decomposition , following the standard method as described in , e.g. , @xcite , we obtain two evolution equations from the effective lagrangian ( [ ourl ] ) . the effect of the quantum fluctuations at lowest order is taken into account with a hartree - type approximation @xcite that consists of the factorization @xmath58 and @xmath59 and the self - consistency condition represented by eq .
( [ expval ] ) given below ; terms of higher order in @xmath54 are neglected .
taking the expectation value @xmath60 , we arrive at the equation for the zero mode @xmath61 written for the dimensionless mean field @xmath62 , where the dimensionless time variable @xmath63 is introduced .
the expectation value of the fluctuations @xmath64 is calculated from the mode functions @xmath65 which are known from the fourier expansion of the field .
the integration is done over the interval of momenta for which the quantum fluctuations are enhanced .
we use the definition ( [ expval ] ) for the back reaction , although it does not fulfill @xmath66 , because it reflects the background of the calculation , the closed - time - path technique which is usually applied for the description of out - of - equilibrium dynamics in field theory . in this framework
, the fluctuations are defined as the two - point correlation functions or equal - time green s functions @xcite @xmath67 the mode functions satisfy the mode equation @xmath68 written for the dimensionless mode functions @xmath69 , where @xmath70 is the time - dependent frequency squared with @xmath71 .
they are used for the calculation of the particle production : the induced particle number density @xmath72 is obtained from the spectral particle number density @xmath73 which can be calculated from the mode functions .
since the unstable solutions of the mode equations contribute to the resonant particle production , the integration in ( [ num ] ) is carried out over the momenta corresponding to these amplified solutions , as in eq .
( [ expval ] ) . for practical purposes ,
we use @xmath74 in definition ( [ numk ] ) , because it keeps the particle number positive definite for all values of @xmath75 and minimizes the fluctuations in the particle number .
accordingly , the initial values of the mode functions are chosen as @xmath76 and @xmath77 such that @xmath78 . however , the results depend only weakly on the choice of @xmath79 both in the initial conditions and in the definition of the particle number . the choice for the initial conditions of the zero mode is suggested by the model : @xmath17 is `` released '' right below the saddle point @xmath80 with zero velocity , i.e. @xmath81 . for the numerical results shown below , we used the starting value @xmath82 .
considering the coupled system ( [ zero ] ) and ( [ modeeq ] ) with ( [ omega ] ) , one recognizes that the expectation value transfers energy to the mode functions via a time dependent frequency , and the mode functions in turn modify the equation of motion for the zero mode .
a natural first step in the solution of this system is to neglect the fluctuation terms in both equations , obtaining the `` classical '' evolution equations with @xmath83 where the mode functions do not react back on the zero mode .
if the variable frequency @xmath84 in the mode equation depends _ periodically _ on time , according to floquet s theorem @xcite it has quasiperiodic solutions of the form @xmath85 where the characteristic exponent @xmath86 depends on @xmath87 , and @xmath88 is a periodic function of the same period as @xmath87 , usually normalized to have unit amplitude . for certain values of the parameter @xmath87
, there exist so called resonance bands that contain the exponentially growing ( or resonant ) solutions with a real characteristic exponent @xmath86 .
these resonance bands are known , for example , for the solutions of the mathieu- or the lam - equation .
an analytic solution of the classical system can be found with a sine - gordon - equation @xcite , which consists essentially of setting @xmath89 in our case .
the corresponding approximate zero mode equation is given by @xmath90 where @xmath91 represents the radius of the potential and is needed to adjust it to the original potential .
it has the solution @xmath92 with @xmath93 , and @xmath94 is the square of the modulus of the jacobian elliptic function @xmath95 .
the resulting mode equation is a lam - equation with the resonance band @xcite @xmath96 .
the resulting @xmath86 is shown in fig .
[ fig : mu ] in comparison with the result of the numerical calculation of @xmath86 for the solution of the classical mode equation . in the range of the single resonance band of the lam - equation , @xmath97 , the numerical calculation yields _ three _ resonance bands , and at least two for @xmath98 .
( this structure reminds of the mathieu - equation which has more than one resonance band . ) the maximum value @xmath99 is found around @xmath100 , and the maximum value for the sine - gordon - solution , @xmath101 , yields an acceptable approximation . the momentum spectrum of the produced particles reflects the structure of the resonance bands .
when the general expression for the mode functions ( [ floq ] ) is inserted into eq .
( [ numk ] ) , it is obvious that the particle production is essentially determined by @xmath86 : @xmath102 in the classical approximation discussed above , the back reaction of the fluctuations on the zero mode ( and also on the fluctuations themselves ) is ignored .
this implies that the approximation is not energy conserving , which can immediately be seen from the zero mode solution which is not damped although it is supposed to transfer its energy to the mode functions , which indeed grow exponentially .
since the mode functions are not decoupled from the zero mode , at least the other direction of energy transfer is reflected by the approximation : the larger the initial value @xmath103 , the more energy can be transferred , and the larger the mode functions and the particle number grow .
the evolution equations including the back reaction are solved numerically . for the evaluation ,
we use @xmath104 and @xmath105 . because the value of @xmath106 in our approximation is only fixed by the value of @xmath107 , which depends on our restriction to the singlet case and
may not correspond to the real physical situation , we also try @xmath108 as the other limit of the range of possible values for @xmath109 .
although the inclusion of the back reaction destroys the parametric resonance , it does not suppress the particle production completely .
the particle number ( fig .
[ fig : ntau ] ) reaches its asymptotic value for @xmath110 . at about the same time , the value of the back reaction term @xmath111 grows larger than 1 , indicating that the approximations made become inappropriate for the description of the dynamics for larger times ( cf .
( [ zero ] ) ) .
one recognizes from fig .
[ fig : ntau ] that the choice of @xmath109 has some influence on @xmath112 ; the asymptotic number density of the produced particles lies between @xmath6 and @xmath113 .
the absolute number of particles can be calculated from this density if the size of the cp - odd bubbles is known .
for example , for a domain with radius @xmath114 , as suggested in @xcite , we estimate a yield of 90 - 100 particles from our model .
a measure for the size of correlated domains of particles is the correlation length @xmath115 , which can be calculated from the two - point correlation function @xmath116 where @xmath117 denotes the dimensionless length . for large @xmath118 ,
@xmath119 from which we estimate the correlation length for @xmath112 .
as a lower limit for the domain radius , we obtain @xmath120 for @xmath121 and @xmath122 for @xmath123 . the momentum spectrum of the produced particles at @xmath124 is shown in fig . [ fig : spec ] .
its maximum is located at about the value of @xmath125 as the maximum of @xmath86 in fig .
[ fig : mu ] , which had been calculated for the system _ without _ back reaction .
the spectrum deviates significantly from a thermal bose - einstein distribution .
thanks the e. schrdinger institute , vienna , for support during the workshop `` quantization , generalized brs cohomologies and anomalies '' , where this work has been encouraged from discussions with m.h.g .
we thank especially d. kharzeev for useful remarks and suggestions .
this research is supported in part by dfg , contract ka 1198/4 - 1 .
+ 99 d. kharzeev , r. d. pisarski , m. h. g. tytgat , phys .
lett . 81
( 1998 ) 512 ; d. kharzeev , r. d. pisarski , m. h. g. tytgat , in proc .
: 3rd workshop on continuous advances in qcd ( qcd 98 ) , minneapolis , to appear , hep - ph/9808366 .
for `` more on history of p , cp , t violation in strong interactions '' see d. kharzeev , r. pisarski , m. tytgat , physics today , february 2000 , 76 ( letter to the editor ) and p. d. morley , i. schmidt , z. phys .
c 26 ( 1985 ) 627 .
see , e.g. , j. f. donoghue , e. golowich , b. r. holstein , dynamics of the standard model ( cambridge university press , 1992 ) . j. traschen , r. brandenberger , phys . rev .
d 42 ( 1990 ) 2491 ; y. shtanov , j. traschen , r. brandenberger , phys .
d 51 ( 1995 ) 5438 .
d. boyanovsky , h. j. de vega , phys .
d 47 ( 1993 ) 2343 ; d. boyanovsky , h. j. de vega , r. holman , j. f. j. salgado , phys .
d 54 ( 1996 ) 7570 ; d. boyanovsky , h. j. de vega , proceedings of the vieme colloque cosmologie , paris , 16 - 18 june 1999 , hep - ph/9909372 , and references therein . l. kofman , a. d. linde , a. a. starobinsky , phys .
lett . 73 ( 1994 ) 3195 ; l. kofman , a. d. linde , a. a. starobinsky , phys .
d 56 ( 1997 ) 3258 ; p. b. green , l. kofman , a. d. linde , a. starobinski , phys .
d 56 ( 1997 ) 6175 .
p. b. greene , l. kofman , a. a. starobinsky , nucl .
b 543 ( 1999 ) 423 .
a. singh , _ field theoretical methods in cosmology _ , phd - thesis , hep - ph/9509241 .
d. i. kaiser , phys .
d 56 ( 1997 ) 706 ; d. i. kaiser , phys .
d 57 ( 1998 ) 702 .
m. shifman , prog .
39 ( 1997 ) 1 ; t. fugleberg , i. halperin and a. zhitnitsky , phys .
d59 ( 1999 ) 074023 .
d. cormier , _ non - equilibrium field theory dynamics in inflationary cosmology _
, phd thesis , hep - ph/9804449 . k. rajagopal , in quark - gluon plasma 2 , ed .
r. hwa , world scientific , singapore , 1995 , p. 484 .
d. boyanovsky , h. j. de vega , r. holman , phys .
d 51 ( 1995 ) 734 ; phys .
d 54 ( 1996 ) 1784 .
s. mrwczyski , b. mller , phys .
b 363 ( 1995 ) 1 .
d. i. kaiser , phys .
d 59 ( 1999 ) 117901 .
e. w. kolb , a. singh , m. srednicki , phys .
d 58 ( 1998 ) 105004 .
g. veneziano , nucl .
b 159 ( 1979 ) 213 .
e. witten , nucl .
b 156 ( 1979 ) 269 ; ann .
phys . 128 ( 1980 ) 363 ; phys .
lett . 81 ( 1998 ) 2862 .
di vecchia , g. veneziano , nucl .
b 171 ( 1980 ) 253 ; p.
di vecchia , f. nicodemi , r. pettorino , g. veneziano , nucl .
b 181 ( 1981 ) 318 .
k. buckley , t. fugleberg , a. zhitnitsky , hep - ph/9910229 ; a. zhitnitsky , hep - ph/0003191 .
j. kapusta , d. kharzeev , l. mclerran , phys .
d 53 ( 1996 ) 5028 .
z. huang , x .-
wang , phys .
d 53 ( 1996 ) 5034 .
d. kharzeev , private communication .
see , e.g. , e. l. ince , ordinary differential equations , ( dover publications , new york , 1956 ) .
d. kharzeev , r. d. pisarski , hep - ph/9906401 . | we investigate the decay of metastable states with broken cp - symmetry which have recently been proposed by kharzeev , pisarski and tytgat to form in hot hadronic matter .
we consider the efficiency of the amplification of the @xmath0-field via parametric resonance , taking the backreaction into account . for times of the order @xmath1
, we find a particle density of about @xmath2 and a correlation length of @xmath3 .
the corresponding momentum spectra show a non - thermal behaviour
. hot hadronic matter , metastable cp - odd bubbles , parametric resonance + _ pacs _ : 11.10.wx , 12.38.mh , 12.39.fe |
this supplementary material contains four parts . in the first two parts
, we prove that the three - loop statistics obeys two relations , eq .
( [ linear3 ] ) and eq .
( [ cyclic ] ) , from the main text . in the third part , we review the basics of group cohomology and dijkgraaf - witten models . in the last part , we derive eq .
( [ statformula ] ) from the main text , which gives an explicit formula for the three - loop statistics in the dijkgraaf - witten models with group @xmath84 . ) .
we use @xmath119 , @xmath120 , and @xmath121 to denote the surfaces swept out by @xmath122 , @xmath123 and @xmath124 , respectively .
the dashed lines denote the vortex loops @xmath125 , @xmath126 and @xmath39 which are enclosed by these surfaces . ] in this section , we prove eq .
( [ linear3 ] ) , which we reprint below for convenience : @xmath127 this relation states that @xmath28 is linear under one of the two types of fusion processes for loops .
we will not present the proof of the other linearity relation , eq .
( [ linear2 ] ) , since eq .
( [ linear2 ] ) can be established in the same way as the familiar result that 2d abelian mutual statistics @xmath128 is linear under fusion of quasiparticles : @xmath129 .
our focus is on the `` more 3d '' relation ( [ linear3 ] ) . to prove eq .
( [ linear3 ] ) , it is enough to show that the braiding process associated with one side of eq .
( [ linear3 ] ) can be smoothly deformed to the process associated with the other side of eq .
( [ linear3 ] ) . then , since the statistical phase is invariant under `` smooth '' deformations@xcite of the braiding path , the relation will follow immediately .
the desired deformation is shown in fig .
[ fig_topoequiv ] .
the lower - left panel of fig .
[ fig_topoequiv ] shows the torus @xmath130 which is swept out by the loop @xmath124 in the braiding process associated with @xmath131 .
similarly , the upper - left panel of fig . [ fig_topoequiv ] shows the two tori @xmath119 and @xmath120 swept out by the loops @xmath122 and @xmath123 in the braiding processes associated with @xmath132 and @xmath133 .
the two right panels show the middle steps of the deformation .
in this section , we prove eq .
( [ cyclic ] ) , which we reprint below for convenience : @xmath138 to prove eq .
( [ cyclic ] ) , we consider a series of thought experiments , shown in figs .
[ fig_firstthought ] , [ fig_secondthought ] and [ fig_thirdthought ] . in the first thought experiment , we create @xmath6 identical links @xmath139 , carrying flux @xmath47 , @xmath48 and @xmath140 respectively ( fig .
[ fig_firstthought ] ) .
we then imagine fusing the @xmath6 links together to form a single link made up of three loops , @xmath141 where @xmath142 , @xmath143 and @xmath144 .
after the fusion , we imagine braiding @xmath134 around @xmath61 with the base @xmath145 .
given the linearity of @xmath28 ( [ linear3 ] ) , the statistical phase associated with this braiding is @xmath146 at the same time , since @xmath147 , the braiding between @xmath148 is no different from two - loop braiding .
therefore , we can use ( [ twostringstat ] ) to deduce @xmath149 .
we conclude that @xmath150 we now derive similar expressions for @xmath151 and @xmath152 . in the case of @xmath151
, we imagine another thought experiment ( fig .
[ fig_secondthought ] ) , where we fuse the @xmath0 loops together to form @xmath134 , but we do nt fuse the @xmath1 or @xmath2 loops .
we then shrink the @xmath1 loops and fuse the @xmath1 onto @xmath2 to form composite loops @xmath153 .
next we fuse the @xmath154 loops to form a new loop @xmath155 .
finally , at the end , we create a vortex - antivortex pair of loops @xmath76 and @xmath156 , with @xmath76 carrying unit flux @xmath157 , and we imagine braiding @xmath76 around @xmath158 . given the linearity of @xmath28 ( [ linear2 ] ) , the statistical phase associated with this braiding process is @xmath159 at the same time , since @xmath160 , we see that @xmath158 is a pure charge so we can use ( [ particlestat ] ) to write @xmath161 .
hence @xmath162 finally , to compute @xmath163 , we consider a third thought experiment ( fig . [ fig_thirdthought ] ) , where we fuse the @xmath1 loops together to form @xmath61 , but we do nt fuse the @xmath0 or @xmath2 .
the composite loops @xmath164 are then fused together to form @xmath165 . at the end
, we create a vortex - antivortex pair of loops @xmath74 and @xmath166 , with @xmath74 carrying unit flux @xmath167 , and we imagine braiding @xmath168 around @xmath74 . by the same reasoning as above , we have @xmath169 to complete the derivation , we note that @xmath170 here the first relation follows from the observation that @xmath171 is the total charge on the @xmath6 identical links and is therefore divisible by @xmath6 ; the second relation follows by the same reasoning .
adding together ( [ abc ] ) , ( [ bca ] ) , ( [ cab ] ) , and using ( [ neutrality ] ) , we derive the cyclic relation ( [ cyclic ] ) . in this section ,
we give a brief review of the basics of group cohomology and dijkgraaf - witten models .
the purpose of this review is to provide the reader with the necessary background to understand the next section where we will compute the three - loop statistics in ( 3 + 1)d dijkgraaf - witten models .
let @xmath117 be a finite group .
the basic objects that group cohomology studies are _ @xmath173-cochains_. an @xmath173-cochain is a @xmath174 valued function @xmath175 : @xmath176 the collection of @xmath173-cochains form an abelian group @xmath177 , where the group operation is defined by @xmath178 the _ coboundary operator _
@xmath179 is a map @xmath180 , defined by @xmath181^{(-1)^{i}}. \label{delta}\end{aligned}\ ] ] it is easy to check that the coboundary operator satisfies @xmath182 .
more importantly , one can check that @xmath179 is nilpotent : @xmath183 . with the help of the coboundary operator
, we can now define _ @xmath173-cocycles _ and _ @xmath173-coboundaries_. an @xmath173-cocycle is an @xmath173-cochain @xmath91 that satisfies @xmath184 .
likewise , an @xmath173-coboundary is an @xmath173-cochain @xmath88 that can be written as @xmath185 where @xmath186 .
the nilpotence of @xmath179 implies that a coboundary must also be a cocycle .
this allows us to define an equivalence relation for the cocycles : two @xmath173-cocycles @xmath187 and @xmath188 are said to be _ cohomologically equivalent _ if and only if @xmath189 , for some @xmath190 . the equivalence classes of the @xmath173-cocycles form an abelian group , called the @xmath173th _ cohomology group _ , which is denoted by @xmath191 $ ] .
the dijkgraaf - witten models@xcite are exactly soluble lattice models that realize different types of lattice gauge theories with finite gauge group . here
we review the space - time path integral formulation of these models . for a hamiltonian formulation of these models in ( 2 + 1)d ,
see e.g. ref .
@xcite .
the basic input needed to construct a @xmath192-dimensional dijkgraaf - witten model with gauge group @xmath117 is ( 1 ) a @xmath192-cocycle @xmath91 and ( 2 ) a triangulation of @xmath192-dimensional euclidean space - time . to build the associated dijkgraaf - witten model
, we label the vertices of the triangulation in an ordered sequence @xmath193 .
the degrees of freedom in the model are group elements @xmath194 which live on the edges @xmath195 $ ] of the triangulation , and can be thought of as gauge fields . for each gauge field configuration @xmath196 ,
the corresponding action @xmath197 is defined by the following recipe .
first , one needs to determine if the configuration @xmath196 is _ flat _ , that is @xmath198 for every @xmath199-simplex @xmath200 $ ] . if @xmath196 is not flat , then @xmath201 . on the other hand ,
if @xmath196 is flat , then @xmath197 is given by a product of complex weights , one for every @xmath192-dimensional simplex in the triangulation .
for example , the action for a ( 3 + 1)d dijkgraaf - witten model with @xmath85-cocycle @xmath91 is given by @xmath202 } \left[\omega(h_{pq } , h_{qr } , h_{rs } , h_{st})\right]^{\sigma_{pqrst}}.\ ] ] here @xmath203 $ ] ( @xmath204 ) runs over the @xmath85-simplices in the triangulation and @xmath205 is a chirality factor which is determined by the ordering of the vertices in each @xmath85-simplex .
the action for dijkgraaf - witten models in other dimensions is similar . to obtain the partition function
, we sum over gauge field configurations and multiply by a normalization factor of @xmath206 where @xmath207 is the number of group elements in @xmath117 and @xmath208 is the number of vertices in the triangulation .
for example , in the ( 3 + 1)d case , the partition function is @xmath209 } \left[\omega(h_{pq } , h_{qr } , h_{rs } , h_{st})\right]^{\sigma_{pqrst}},\ \ \label{supp - dw}\end{aligned}\ ] ] where the summation @xmath210 is taken over flat gauge field configurations only , since @xmath201 for the other configurations .
it can be shown that if @xmath87 belong to the same cohomological equivalence class , then the corresponding dijkgraaf - witten models are equivalent , i.e. share the same partition function .
thus , the inequivalent @xmath192-dimensional dijkgraaf - witten models are classified by the cohomology group @xmath211 $ ] .
[ [ loop - statistics - of-31d - dijkgraaf - witten - modelsgauged - spt - models ] ] 4 .
loop statistics of ( 3 + 1)d dijkgraaf - witten models / gauged spt models ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ in this section , we derive the formula ( [ statformula ] ) from the main text .
this formula gives the three - loop statistics of ( 3 + 1)d gauged spt models with symmetry group @xmath84 .
alternatively the formula ( [ statformula ] ) can be thought of as giving the three - loop statistics of dijkgraaf - witten models with @xmath84 , since gauged spt models are exactly equivalent to dijkgraaf - witten models .
( see ref . for a discussion of this equivalence in ( 2 + 1 ) dimensions ) . for notational reasons ,
we find it more convenient use the dijkgraaf - witten language , so in what follows we will phrase our calculation in terms of the three - loop braiding statistics of dijkgraaf - witten models .
our derivation of ( [ statformula ] ) proceeds in three steps . in the first step , we derive a `` dimensional reduction '' formula that relates the vortex loop statistics of ( 3 + 1)d dijkgraaf - witten models to the vortex statistics of ( 2 + 1)d dijkgraaf - witten models . in the second step , we review some previously known results on vortex statistics in ( 2 + 1)d dijkgraaf - witten models . in the final step , we put everything together and we derive the formula ( [ statformula ] ) . in this section , we derive a `` dimensional reduction '' formula ( [ supp - dimred ] ) that relates the vortex loop statistics of ( 3 + 1)d dijkgraaf - witten models to the vortex statistics of ( 2 + 1)d dijkgraaf - witten models . as in the main text , we assume @xmath114 and we restrict our analysis to models with abelian three - loop statistics . the starting point for our derivation is eq .
( [ 3dtorus ] ) from the main text , which we repeat below for convenience : @xmath212 this relation allows us to compute the three - loop statistics @xmath24 using a 3d spatial torus geometry . here
, the quantity @xmath213 is the mutual statistics between two noncontractible vortex lines @xmath74 and @xmath76 oriented along the @xmath73 direction of a 3d torus , in the presence of a gauge flux @xmath80 through the `` @xmath73-hole '' of the 3d torus ( fig .
[ fig_dimreduction ] ) .
similarly , the quantity @xmath214 is the mutual statistics between noncontractible vortex lines @xmath75 and @xmath77 in the absence of a gauge flux through the `` @xmath73-hole '' of the 3d torus .
equation ( [ supp-3dtorus ] ) is especially useful for analyzing dijkgraaf - witten models since these models have a vanishing correlation length , and hence we can use a 3d torus of any size and triangulation that we like . here
we find it convenient to choose a `` thin torus '' , i.e. a 3d spatial torus with a thickness of only one unit cell in the @xmath73 direction and arbitrary dimensions in the @xmath215 and @xmath216 directions ( see fig .
[ fig_dimreduc_dw ] for an analogous geometry in one lower dimension ) .
then , since our ( 3 + 1)d system has finite thickness in the @xmath73 direction , we can think of it as a ( 2 + 1)d lattice model .
in fact , we will show below that if we fix the gauge flux through the @xmath73-hole of the torus to be some group element @xmath217 , then this ( 2 + 1)d lattice model is exactly the ( 2 + 1)d dijkgraaf - witten model associated with the 3-cocycle @xmath218 where @xmath91 is the 4-cocycle associated with the ( 3 + 1)d dijkgraaf - witten model . given the above result , we can rewrite eq .
( [ supp-3dtorus ] ) as @xmath219 where @xmath220 denotes the mutual statistics of two vortex excitations @xmath54 in the ( 2 + 1)d dijkgraaf - witten model ( [ supp - c ] ) with 3-cocycle @xmath221 , while @xmath222 denotes the braiding statistics of two vortex excitations @xmath78 in the ( 2 + 1)d dijkgraaf - witten model ( [ supp - c ] ) with 3-cocycle @xmath223 .
equation ( [ supp - dimred ] ) is the desired dimensional reduction formula , and the main result of this section .
direction contains one lattice constant only . each triangular prism @xmath224 $ ] contains three tetrahedra @xmath225 $ ] , @xmath226 $ ] and @xmath227 $ ] .
the ( 2 + 1)d weight associated with the prism @xmath224 $ ] can be rewritten as a ( 1 + 1)d weight of the corresponding triangle @xmath200 $ ] . ] to complete our derivation , we now prove our earlier claim that the ( 3 + 1)d dijkgraaf - witten model in the `` thin torus '' geometry is equivalent to a ( 2 + 1)d dijkgraaf - witten model with cocycle @xmath228 . in the first step
, we consider a ( 3 + 1)d space - time geometry which is periodic in all spatial dimensions .
we triangulate the ( 3 + 1)d space - time using a collection of 4d triangular prisms with a thickness of one lattice constant along the @xmath73 spatial dimension ; each 4d prism is further triangulated into four 4-simplices .
( an analogous triangulation in one lower dimension is shown in fig .
[ fig_dimreduc_dw ] ) . in this triangulation
, the action @xmath197 in ( [ supp - dw ] ) can be written as a product of the local complex weights @xmath229 associated with each 4d prism @xmath230 $ ] . here
, we use the notation that all the links parallel to the @xmath73-axis are labeled by @xmath231 , [ qq'],\dots$ ] , similarly to that in fig .
[ fig_dimreduc_dw ] . each 4d prism @xmath230 $ ]
then contains four 4-simplices , @xmath232 $ ] , @xmath233 $ ] , @xmath234 $ ] and @xmath235 $ ] , so the local weight @xmath229 is given by @xmath236^{\sigma_{pqrs } } \label{suppw}\end{aligned}\ ] ] where @xmath237 is the chirality of the 3-simplex @xmath238 $ ] .
we next fix the gauge fields living on all @xmath73-links @xmath231 $ ] to be @xmath239 .
different choices of @xmath240 will correspond to different gauge fluxes through the `` @xmath73-hole '' of the 3d torus . substituting @xmath241 into ( [ suppw ] ) and remembering that periodicity along the @xmath73 dimension identifies the field @xmath242 with @xmath243
, we obtain @xmath244^{\sigma_{pqrs}}$ ] where @xmath245 one may check that @xmath228 is a 3-cocycle for each @xmath240 .
the ( 3 + 1)d dijkgraaf - witten partition function ( [ supp - dw ] ) then reduces to @xmath246 } \left[\chi_h(h_{pq } , h_{qr } , h_{rs})\right]^{\sigma_{pqrs}}\label{supp - dw-2d},\ ] ] which is exactly the partition function of a ( 2 + 1)d dijkgraaf - witten model with cocycle @xmath228 . here , we summarize some previously known results regarding the braiding statistics of ( 2 + 1)d dijkgraaf - witten models .
we list these results without proof ; readers who are interested in how to obtain these results may consult refs . . for simplicity ,
we focus on dijkgraaf - witten models with gauge group @xmath247 and abelian quasiparticle statistics .
it is known that the most general model with these properties can be constructed from a 3-cocycle of the form@xcite @xmath248 ) } , \label{supp-3cocyl}\ ] ] where @xmath249 is an arbitrary integer @xmath250 matrix . here
we parameterize the different elements of @xmath117 using integer @xmath5-component vectors @xmath251 with @xmath252 , and the square bracket @xmath253 $ ] is defined to be @xmath254 with values taken in the range @xmath255 .
it is easy to check that the above expression for @xmath91 is indeed a 3-cocycle for any choice of @xmath249 .
the most general quasiparticle excitations in the above models carry both flux and charge . in order to simplify our notation
, we will call all of these excitations `` vortices '' ; pure charges will be regarded as special kinds of vortices carrying zero flux .
general vortices can be labeled by ordered pairs @xmath256 where @xmath257 are @xmath5-component integer vectors , @xmath258 and @xmath259 with @xmath260 .
the label @xmath261 can be thought of as the amount of flux @xmath262 carried by the vortex @xmath0 while the label @xmath263 describes the amount of charge attached to @xmath0 .
one subtlety is that the above labeling scheme is only well - defined once we choose some conventions in particular , we have to pick a convention for which vortices are labeled as `` pure '' fluxes @xmath264 . for concreteness
, we will use the labeling convention defined in ref . throughout our discussion .
the braiding statistics of the vortices are known and can be written down explicitly in terms of the matrix @xmath249 .
specifically , the mutual statistics between two vortices @xmath256 and @xmath265 is given by @xmath266 while the exchange statistics of @xmath256 is given by @xmath267 for a derivation of the above formulas , see e.g. refs . .
we now derive the formula , eq .
( [ statformula ] ) , from the main text .
this formula gives the three - loop statistics for any @xmath268 dijkgraaf - witten model with gauge group @xmath84 and with a @xmath85-cocycle of the form @xmath269 ) } , \label{supp-4cocycle}\end{aligned}\ ] ] where @xmath102 is a three - index integer tensor .
( the reason that we focus on the above class of @xmath85-cocycles is that we believe that they are the most general cocycles such that the corresponding dijkgraaf - witten models have abelian loop statistics ) . in the first step ,
we compute the `` dimensionally reduced '' @xmath270-cocycle @xmath271 ( [ supp - c ] ) corresponding to @xmath91 . inserting the expression ( [ supp-4cocycle ] ) into ( [ supp - c ] )
, we find @xmath272)},\ ] ] where @xmath273 is an integer matrix whose elements are given by @xmath274 next , we note that the above 3-cocycle @xmath228 falls into the form ( [ supp-3cocyl ] ) , so we can immediately write down the braiding statistics of the vortices in the ( 2 + 1)d dijkgraaf - witten models with 3-cocycle @xmath228 . in particular , according to ( [ supp-2dmut ] ) , the mutual statistics between two vortices @xmath256 and @xmath265 is given by @xmath275 in the final step , we use the dimensional reduction formula ( [ supp - dimred ] ) to relate the three - loop statistics in the ( 3 + 1)d dijkgraaf - witten model with cocycle ( [ supp-4cocycle ] ) to the vortex statistics in the ( 2 + 1)d dijkgraaf - witten models with cocycles @xmath221 and
@xmath223 : @xmath276 examining fig . [ fig_dimreduction ] , it is clear that @xmath277 carry the same flux as one another , while @xmath75 carries opposite flux .
the same is true for @xmath278 .
therefore , we can label these vortices as @xmath279 substituting these expressions into ( [ supp-2dmut2 ] ) and using ( [ supp - eq2d ] ) , we derive @xmath280 to get the unit flux statistics @xmath44 ( [ unitfluxtensor ] ) , we specialize to the case where @xmath15 and @xmath26 carry unit flux , that is @xmath281 , @xmath282 , @xmath283 , and we multiply eq .
( [ supp-3loop ] ) by @xmath6 . after these manipulations ,
the second term on the right hand side drops out and we obtain @xmath284 we conclude that @xmath285 the exchange statistics @xmath99 can be computed in a similar way , and is given by @xmath286 | while it is well known that three dimensional quantum many - body systems can support non - trivial braiding statistics between particle - like and loop - like excitations , or between two loop - like excitations , we argue that a more fundamental quantity is the statistical phase associated with braiding one loop @xmath0 around another loop @xmath1 , while both are linked to a third loop @xmath2 .
we study this three - loop braiding in the context of @xmath3 gauge theories which are obtained by gauging a gapped , short - range entangled lattice boson model with @xmath3 symmetry .
we find that different short - range entangled bosonic states with the same @xmath3 symmetry ( i.e. different symmetry - protected topological phases ) can be distinguished by their three - loop braiding statistics . .
( b ) a top view of the braiding process within the plane that @xmath2 lies in . ( c )
a torus @xmath4 is swept out by @xmath0 during the braiding .
loop @xmath1 ( dashed circle ) is enclosed by @xmath4 . ]
_ introduction .
_ a powerful way to characterize the topological properties of two dimensional gapped quantum many - body systems is to examine their quasiparticle braiding statistics @xcite .
it is thus natural to wonder : what is the analogous quantity that characterizes three dimensional ( 3d ) systems ?
the simplest candidate 3d quasiparticle statistics is of limited use since 3d systems can only support bosonic and fermionic quasiparticles . on the other hand
, 3d systems can support much richer braiding statistics between particle - like excitations and loop - like excitations@xcite or between two loop - like excitations@xcite .
thus , one might guess that particle - loop and loop - loop braiding statistics are the natural generalizations of quasiparticle statistics to three dimensions . in this paper
, we argue that this guess is incorrect : particle - loop and loop - loop braiding statistics do not fully capture the topological structure of 3d many - body systems . instead , more complete information can be obtained by considering a _ three - loop _ braiding process in which a loop @xmath0 is braided around another loop @xmath1 , while both are linked with a third loop @xmath2 ( fig . [ fig_braiding ] ) .
we believe that three - loop braiding statistics is one of the basic pieces of topological data that describe 3d gapped many - body systems , and much of this work is devoted to understanding the general properties of this quantity . also ,
as an application , we show that three - loop statistics can be used to distinguish different short - range entangled many - body states with the same ( unitary ) symmetry i.e. different _ symmetry - protected topological ( spt ) phases_@xcite .
the latter result shows that the braiding statistics approach to spt phases , outlined in ref .
, can be extended to three dimensions . _ discrete gauge theories .
_ for concreteness , we focus our analysis on a simple 3d system with loop - like excitations , namely lattice @xmath3 gauge theory@xcite .
more specifically , we consider a 3d lattice boson model built out of @xmath5 different species of bosons , where the number of bosons in each species is conserved _ modulo _
@xmath6 so that the system has a @xmath7 symmetry .
we suppose that the ground state of the boson model is gapped and short - range entangled that is , it can be transformed into a product state by a local unitary transformation@xcite .
we then imagine coupling such a lattice boson model to a @xmath7 lattice gauge field@xcite
. in general these gauge theories contain two types of excitations : point - like `` charge '' excitations which carry gauge charge , and string - like `` vortex loop '' excitations which carry gauge flux .
the most general charge excitations can carry gauge charge @xmath8 where each component @xmath9 is an integer defined modulo @xmath6 .
the most general vortex loop can carry gauge flux @xmath10 where @xmath11 is a multiple of @xmath12 .
in fact , since we can always attach a charge to a vortex loop to obtain another vortex loop , a general vortex loop excitation carries _ both _ flux and charge .
let us try to understand the braiding statistics of these excitations .
in general , there are three types of braiding processes we can consider : processes involving two charges , processes involving a charge and a loop , and processes involving multiple loops .
clearly , the first type of process can not give any statistical phase since the charges are excitations of the short - range entangled boson model and therefore must be bosons .
on the other hand , the second kind of process , involving charges and loops , _ can _ give a nontrivial phase .
more specifically , if we braid a charge @xmath8 around a vortex loop with gauge flux @xmath10 , the resulting statistical phase is given by the aharonov - bohm formula @xmath13 where `` @xmath14 '' denotes the vector dot product . .
( b ) if @xmath15 are neutral , the two - loop process can be smoothly deformed into a process in which @xmath0 is braided around the vacuum.,width=316 ] all that remains is to examine the braiding statistics of loops . the simplest process one can consider@xcite involves braiding a loop @xmath0 around another loop @xmath1 as shown in fig .
[ fig_twoloop](a ) . to analyze this process , we use two facts about _ unlinked _ vortex loops : first , a subset of vortex loops , which we call `` neutral '' loops ,
can be shrunk to a point and annihilated by local gauge invariant operators .
second , all other vortex loops can be obtained from neutral loops by attaching an appropriate amount of charge . with these facts in mind , let us first suppose that both @xmath15 are neutral . in this case , it follows from general principles that the statistical phase @xmath16 , since we can `` smoothly''@xcite deform the two - loop braiding process into another process in which @xmath0 is braided around the vacuum [ fig .
[ fig_twoloop](b ) ] .
now consider the general case where @xmath15 carry charge . in this case
, @xmath15 can be thought of as neutral loops with some attached charge .
it then follows from the aharonov - bohm formula ( [ particlestat ] ) that the berry phase associated with braiding @xmath0 around @xmath1 is @xmath17 where @xmath18 and @xmath19 denote the charge and flux carried by @xmath15 respectively . to see this , note that during the two - loop braiding , the charge @xmath20 is braided around the flux @xmath21 and the flux @xmath22 is braided around the charge @xmath23 .
while the above calculations show that 3d gauge theories can exhibit nonvanishing braiding statistics , we can see that these statistical phases are the same for all gauge theories with gauge group @xmath3 , independent of the properties of the bosonic matter .
yet , we expect that the bosonic matter should be important : if two lattice boson models realize different short - range entangled phases with the same symmetry ( i.e. different spt phases@xcite ) , then presumably the corresponding gauge theories belong to distinct phases as well , by analogy with the 2d case@xcite .
clearly , if we want to distinguish these different types of 3d gauge theories , we must consider braiding processes with more than two loops .
_ three - loop braiding statistics .
_ for these reasons , we are naturally led to consider a braiding process involving two loops @xmath15 which are linked with a third `` base '' loop @xmath2 ( fig . [ fig_braiding ] ) .
when the loop @xmath0 sweeps around @xmath1 in a right - handed manner , it can acquire a statistical berry phase which we will denote by @xmath24 where @xmath25 with @xmath26 being an integer vector .
we use the notation @xmath24 , rather than @xmath27 because @xmath28 is insensitive to the charge attached to @xmath2 and depends only on its flux @xmath29 .
similarly , we will also consider an exchange or half - braiding process in which two identical loops @xmath0 , which are linked with a base loop with flux @xmath30 , are braided through one another and exchange places .
the statistical phase associated with this exchange will be denoted by @xmath31 .
note that throughout this paper we assume the loops have abelian statistics .
these three - loop braiding processes are fundamentally different from the two - loop case because in the three - loop topology , the base loop @xmath2 may prevent us from shrinking @xmath0 and @xmath1 to a point and annihilating them locally . as a result ,
the above argument that vortex loop statistics follow the aharonov - bohm law ( [ twostringstat ] ) is no longer valid .
thus , the three - loop braiding statistics are less constrained than the two - loop case .
_ constraints on @xmath24 and @xmath31 .
_ we now discuss the basic physical constraints on the three - loop braiding statistics .
one of the simplest constraints is that @xmath32 . to derive this property
, we note that a process in which @xmath0 winds around @xmath1 can be smoothly deformed into one in which @xmath1 winds around @xmath0 .
therefore , since the statistical phase is invariant under smooth deformations of the braiding path , @xmath24 must be symmetric in @xmath0 and @xmath1 .
another obvious constraint is that @xmath33 .
this relation is clear since a full braiding is equivalent to performing two exchanges in series .
even more powerful constraints on @xmath28 can be obtained by thinking about `` fusion '' of vortex loops .
more specifically , there are two distinct ways to fuse loops together . in the first type of fusion process [ fig .
[ fig_fusion](a ) ] , two loops @xmath34 that are linked to the same loop @xmath2 can be fused to form a new loop ` @xmath35 ' that is also linked to @xmath2 . in the second type of fusion process [ fig .
[ fig_fusion](b ) ] , two loops @xmath34 that share the same flux @xmath36 but are linked with two different loops @xmath37 and @xmath38 , can be fused to form a loop ` @xmath39 ' , which is linked to both @xmath37 and @xmath38 .
it is not hard to see that @xmath24 must be _ linear _ under both fusion processes@xcite ( fig .
[ fig_linearity ] ) : @xmath40 to derive these identities , it suffices to show that the processes defined by the left hand sides of fig .
[ fig_linearity]a - b can be smoothly deformed into the processes corresponding to the right hand sides of fig .
[ fig_linearity]a - b .
these deformations are described in the supplementary material@xcite . )
[ panel ( a ) ] and ( [ linear3 ] ) [ panel ( b ) ] . here ,
@xmath41 , while @xmath42 and @xmath43
. ] one implication of the linearity of @xmath28 ( [ linear2]-[linear3 ] ) is that we can reconstruct all the three - loop statistics from the statistics of vortex loops with _ unit _ flux .
the statistics of these unit fluxes can in turn be summarized by two tensors @xmath44 and @xmath45 .
these tensors are defined by @xmath46 where @xmath15 are _ any _ two loops carrying unit flux @xmath47 and @xmath48 respectively , and where @xmath49 with a @xmath50 in the @xmath51th entry and @xmath52 everywhere else . to see why the tensor @xmath44 is well - defined modulo @xmath53 , note that if we choose another set of loops @xmath54 with the same flux ,
then the only possible topological difference between @xmath54 and @xmath15 is that they may have different amounts of charge attached to them .
but from the aharonov - bohm formula ( [ particlestat ] ) , we see that attaching charge to @xmath0 and @xmath1 can only shift the value of @xmath55 by multiples of @xmath56 and hence can only shift @xmath44 by multiples of @xmath53 .
similar reasoning applies in the case of @xmath45 .
given that the @xmath44 and @xmath45 effectively summarize all the three - loop statistics , all that remains is to find the physical constraints on these two quantities .
we now argue that these constraints are as follows : @xmath57 the first two constraints ( [ mutex ] ) are obvious , since they are special cases of the more general relations discussed above .
the quantization conditions ( [ quant ] ) are also easy to derive : for example , to prove the first equation in ( [ quant ] ) , consider a thought experiment in which a loop @xmath0 carrying flux @xmath58 , together with @xmath6 identical loops @xmath1 carrying flux @xmath59 , are all linked to a common base loop with flux @xmath60 .
now imagine we fuse the @xmath1 loops together to form a new loop @xmath61 and then we braid @xmath0 around @xmath61 . by the linearity of @xmath28 ( [ linear2 ] ) ,
the resulting statistical phase is @xmath62 at the same time , we can see that @xmath63 so @xmath61 is a pure charge .
it then follows from the aharonov - bohm formula ( [ particlestat ] ) that @xmath64 , which is a multiple of @xmath56 . combining these two observations
we deduce that @xmath44 is a multiple of @xmath56 .
the proof of the second equation in ( [ quant ] ) is similar .
equations ( [ cyclic]-[cyclicex ] ) are the most interesting constraints on @xmath65 as these relations have no analogues in the theory of 2d braiding statistics .
we call eq .
( [ cyclic ] ) the _ cyclic _ relation .
a physical derivation of the cyclic relation is given in the supplementary material@xcite .
the first equation in ( [ cyclicex ] ) can be proved in a similar manner . on the other hand
, we do not have a physical derivation of @xmath66 , so this constraint on @xmath65 is simply a conjecture .
this conjecture is supported by two pieces of evidence : first , all the microscopic models constructed below obey this relation .
second , we can prove the weaker , but closely related relation @xmath67 ( mod @xmath53 ) using the second equation in ( [ mutex ] ) together with the first equation in ( [ cyclicex ] ) .
_ dimensional reduction .
_ we now derive a formula for the three - loop statistics that will be useful in analyzing the microscopic models discussed later .
this formula is obtained by considering our system in an @xmath68 torus geometry i.e. a geometry with periodic boundary conditions in all three directions .
let @xmath15 be two loops linked with a base loop @xmath2 carrying flux @xmath69 [ fig .
[ fig_dimreduction](a ) ] . for concreteness ,
suppose that @xmath2 lies in the @xmath70 plane while @xmath15 lie in the @xmath71 plane .
when @xmath0 sweeps around @xmath1 , it gives rise to a statistical phase @xmath72 which we wish to compute . to this end
, we stretch @xmath0 in the @xmath73 direction until it wraps all the way around the periodic @xmath73 direction .
we can then fuse @xmath0 with itself , thereby splitting @xmath0 into two noncontractible loops @xmath74 and @xmath75 [ fig [ fig_dimreduction](b ) ] .
similarly , we can stretch @xmath1 in the @xmath73 direction and fuse it with itself so that it splits into @xmath76 and @xmath77 .
it is clear that the braiding process involving @xmath0 and @xmath1 can now be decomposed into two separate processes in which @xmath74 is braided around @xmath76 and @xmath75 is braided around @xmath77 .
since these two processes are separate , we can think of them as taking place in two separate systems [ fig .
[ fig_dimreduction](c ) ] . furthermore , for the process involving @xmath54 we can stretch @xmath2 in the @xmath70 plane so that it fuses and annihilates with itself .
this effectively leaves a gauge flux @xmath69 through one of the three holes of the 3d torus more precisely , the hole bounded by a noncontractible cycle along the @xmath73 direction ( the `` @xmath73-hole '' ) . likewise , for the process involving @xmath78 we can shrink @xmath2 in the @xmath70 plane until it fuses and annihilates with itself , leaving no gauge flux through the `` @xmath73-hole . '' in this way , we see that @xmath24 can be expressed as @xmath79 where the quantities on the right hand side are statistical phases associated with braiding two vortex lines around one another in the @xmath70 plane . in the first term , this braiding takes place in the presence of a gauge flux @xmath80 through the `` @xmath73-hole '' , while in the second term there is no such gauge flux .
the relative sign comes from the fact that the two pairs are braided in opposite directions .
the formula ( [ 3dtorus ] ) is useful because each of the terms on the right hand side can be thought of as braiding statistics of a 2d system if we take the thermodynamic limit @xmath81 , while keeping @xmath82 finite but larger than the correlation length .
an analogous formula can be derived for the exchange statistics @xmath83 .
_ microscopic models .
_ to obtain examples of systems with nontrivial three - loop statistics , we consider `` gauged '' spt models that is , we take the exactly soluble lattice boson models of ref . which realize different spt phases , and we couple them to a gauge field .
these gauged spt models can be equivalently@xcite thought of as dijkgraaf - witten models@xcite . as above
, we focus on the case where the symmetry group is @xmath84 . as discussed in ref .
, the basic input for constructing a 3d gauged spt model is a _ @xmath85-cocycle _
@xmath86 . if two cocycles @xmath87 differ by a _ @xmath85-coboundary _ @xmath88 , i.e. @xmath89 , then the corresponding models belong to the same spt phase . thus , inequivalent models are classified by elements of the cohomology group @xmath90 $ ] . here
we focus on @xmath85-cocycles @xmath91 of the form @xmath92 ) } , \label{omega}\end{aligned}\ ] ] where @xmath93 is an integer tensor and we parameterize the different group elements of @xmath84 with integer vectors @xmath94 with @xmath95 .
the square bracket @xmath96 $ ] is defined to be @xmath97 ( mod @xmath6 ) with values taken in the range @xmath98 .
our task is to compute the three - loop statistics @xmath44 and @xmath99 of the gauged spt model with cocycle @xmath91 .
the details of this calculation , which is based on the formula ( [ 3dtorus ] ) , can be found in the supplementary material@xcite .
the end result is @xmath100 as a consistency check , one can easily verify that these expressions satisfy conditions ( [ mutex]-[quant ] ) .
conversely , it is a straightforward mathematical exercise to check that every @xmath44 and @xmath99 that obeys ( [ mutex]-[quant ] ) can be written in the form ( [ statformula ] ) for some @xmath93 .
hence , every solution to ( [ mutex]-[quant ] ) can be physically realized as a gauged spt model .
_ examples . _
the simplest example is @xmath101 . in this case
, @xmath102 has only one component @xmath103 , and ( [ statformula ] ) gives trivial loop statistics , @xmath104 , for any choice of @xmath102 .
this is a reasonable result since @xmath105 = 0 $ ] , so all the spt models with @xmath106 are equivalent to product states@xcite .
the simplest nontrivial example is given by @xmath107 .
in this case , if we choose @xmath108 , @xmath109 and all other components vanishing , we obtain the three - loop statistics shown in table [ tab ] .
we can see that there are @xmath110 distinct types of statistics that can be realized by the gauged spt models with @xmath111 .
this is also a reasonable result since @xmath112 = ( \mathbb{z}_n)^2 $ ] so the spt models realize @xmath110 distinct phases@xcite .
evidently , each phase is associated with a different type of three - loop statistics .
.@xmath44 for the spt models with @xmath113 symmetry . [ cols="^,^,^,^,^,^",options="header " , ] _ discussion . _
the above examples show that the gauged spt phases with @xmath106 and @xmath111 are uniquely characterized by their three - loop statistics .
more generally , we find it plausible that every 3d spt phase with unitary symmetries is uniquely characterized by its three - loop statistics similarly to what has been proposed in the 2d case@xcite .
one subtlety in checking this conjecture for more general @xmath114 is that when @xmath115 , the cocycles ( [ omega ] ) do not exhaust all elements of @xmath116 $ ] .
furthermore , the remaining elements of @xmath116 $ ] can lead to non - abelian three - loop statistics ( see ref . for examples of this phenomenon in the 2d case ) .
thus , a theory of non - abelian loop statistics may be necessary to proceed further in this direction .
is three - loop statistics measurable ? in principle , three - loop statistics could be measured experimentally by performing interferometry on loop - like excitations ; in practice , such an experiment would be challenging .
a more straightforward application is to numerical simulations , where three - loop statistics could be directly extracted from an appropriate berry phase computation . after submitting this paper for publication
, we became aware of an independent work@xcite containing related results on three - loop statistics .
other recent work on this topic includes refs.@xcite .
we thank m. cheng , c .- h .
lin and a. vishwanath for helpful discussions .
this work is supported by the alfred p. sloan foundation and nsf under grant no .
. 100 j. m. leinaas and j. myrheim , nuovo cimento * 37b * , 1 ( 1977 ) ; f. wilczek , phys . rev . lett . * 48 * , 1144 ( 1982 ) ; d. arovas , j. r. schrieffer and f. wilczek , phys . rev . lett . *
53 * , 722 ( 1984 ) .
y. aharonov and d. bohm , phys . rev . *
115 * , 485 ( 1959 ) .
m. g. alford and f. wilczek , phys .
rev .
lett . * 62 * , 1071 ( 1989 ) ; r. rohm , princeton university ph.d .
thesis ( unpublished ) ( 1985 ) . l. m. krauss and f. wilczek , phys .
rev .
lett .
* 62 * , 1221 ( 1989 ) ; j. preskill and l. m. krauss , nucl .
phys .
b * 341 * , 50 ( 1990 ) . c. aneziris , a. p. balachandran , l. kauffman , and a. m. srivastava , int . j. mod .
phys . a * 6 * , 2519 ( 1991 )
. m. g. alford , k .-
m .
lee , j. march - russell , and j. preskill , nucl .
phys .
b * 384 * , 251 ( 1992 ) .
j. c. baez , d. k. wise , and a. s. crans , adv .
theor .
math .
phys . * 11 * , 707 ( 2007 ) .
x. chen , z .- c .
gu , z .- x .
liu , and x .-
g .
wen , phys .
rev .
b * 87 * , 155114 ( 2013 ) . l. fidkowski and a. kitaev , phys . rev .
b * 83 * , 075103 ( 2011 ) ; f. pollmann , a. m. turner , e. berg , and m. oshikawa , phys , rev .
b * 81 * , 064439 ( 2010 ) ; x. chen , z .- c .
gu , and x .-
g .
wen , phys .
rev .
b * 83 * , 035107 ( 2011 ) ; x. chen , z .- c .
gu , and x .-
g .
wen , phys .
rev .
b * 84 * , 235128 ( 2011 ) ; n. schuch , d. perez - garcia , and i. cirac , phys . rev .
b * 84 * , 165139 ( 2011 ) .
m. z. hasan and c. l. kane , rev .
mod .
phys .
* 82 * , 3045 ( 2010 ) .
m. levin and z .- c .
gu , phys .
rev .
b * 86 * , 115109 ( 2012 ) .
our analysis can be straightforwardly extended to gauge theories with any finite abelian gauge group @xmath117 ; we focus on @xmath114 primarily to simplify the discussion .
f. verstraete , j. i. cirac , j. i. latorre , e. rico , and m. m. wolf , phys .
rev .
lett .
* 94 * , 140601 ( 2005 ) ; g. vidal , phys .
rev .
lett * 99 * , 220405 ( 2007 ) ; x. chen , z .- c .
gu , and x .-
g .
wen , phys .
rev .
b * 82 * , 155138 ( 2010 ) .
j. b. kogut , rev .
mod .
phys . * 51 * , 659 ( 1979 ) .
there is also another two - loop braiding process in which a small loop @xmath0 braids around a large loop @xmath1 , in the same way that a particle braids around a loop .
the statistical phase for this process is similar to ( [ particlestat ] ) : @xmath118 .
we say that a multi - loop braiding process can be `` smoothly '' deformed into another process if we can interpolate between the two processes via a sequence of local changes to the worldsheets swept out by the loops .
this sequence of local changes can be arbitrary except that the worldsheets of any two loops that are being braided around one another must stay far apart during every step .
m. cheng and z .- c .
gu , phys .
rev .
lett . * 112 * , 141602(2014 ) .
the exchange statistics @xmath31 also satisfies linearity relations which are analogues of ( [ linear2]-[linear3 ] ) , but we will not list them here . see supplementary material for details .
r. dijkgraaf and e. witten , commun .
math .
phys . * 129 * , 393 ( 1990 ) .
x. chen , y .- m .
lu , and a. vishwanath , nat .
commun . * 5 * , 3507 ( 2014 ) .
m. de wild propitius , _ topological interactions in broken gauge theories _ , phd dissertation ( 1995 ) , available at arxiv : hep - th/9511195 .
s. jiang , a. mesaros , and y. ran , arxiv:1404.1062 j. wang and x .- g .
wen , arxiv:1404.7854 .
c .-
m .
jian and x .-
l .
qi , arxiv:1405.6688 . k. s. brown , _ cohomology of groups _ , ( springer , 1982 ) . y. hu , y. wan , and y .- s .
wu , phys .
rev .
b * 87 * , 125114 ( 2013 ) .
c .- h .
lin and m. levin , phys .
rev .
b * 89 * , 195130 ( 2014 ) . |
the coherent states @xmath0 introduced by glauber @xcite with respect to problems of quantum optics are eigenstates of the photon annihilation operator @xmath6 ( @xmath7 ) , their eigenvalues covering the entire complex plane .
since the heisenberg uncertainty relation @xcite is minimized in coherent states , the latter are interpreted as the `` classical '' ones .
the photon distribution function ( pdf ) in the coherent state is the usual poisson distribution , which is a smooth function with a single maximum near the mean photon number . in nonclassical gaussian states of light
the pdfs possess a number of specific properties related to the effect of quadrature squeezing @xcite .
the pdf for one - mode gaussian states of electromagnetic field was discussed in @xcite . in @xcite the single - mode _ correlated _
state of light was introduced , which minimizes the schrdinger - robertson uncertainity relation @xcite .
the oscillatory character of pdf for squeezed and correlated gaussian light was discussed in @xcite .
different nonclassical states of light might play an essential role in nonlinear processes @xcite . by using different finite or infinite sums of coherent states for a one - mode quantum harmonic oscillator various non - gaussian nonclassical states of light may be constructed .
recently , there have been much interest in the properties and in the generation of so - called `` schrdinger cats '' @xcite in the context of quantum optics @xcite-@xcite .
these quantum superpositions of the finite or infinite number of coherent states have various nonclassical characteristics emerging due to the quantum interference between summands .
the first example of such kind was constructed in @xcite , where the concept of the even and odd coherent states @xmath8 was introduced .
these states satisfy the equation @xmath9 .
moreover , they manifest such remarkable nonclassical properties as oscillations in photon number distribution , sub - poissonian statistics , and the quadrature squeezing
. the methods of generating and detecting the even and odd states in quantum optics were discussed , e.g. , in @xcite .
the aim of the present article is to study the photon statistics in another representative of nonclassical states : the `` real '' ( rcs ) and `` imaginary '' ( ics ) coherent states ( or `` real '' and `` imaginary cats '' ) . @xmath10
the normalization constant being given by ( @xmath11 ) @xmath12 similar superpositions , @xmath13 , were considered from the group - theoretical point of view in @xcite , where they were called `` charged schrdinger cats '' .
all the states of this family belong to the class of superposition states discussed in @xcite and @xcite .
the interference between @xmath0 and @xmath1 leads again to the non - poissonian ( modulated poissonian ) pdf .
evidently , the statistical properties of real and imaginary cat states , as well as the degree of squeezing , depend essentially on both the modulus @xmath5 and the phase difference @xmath14 between @xmath15 and @xmath16 .
it is interesting , however , that the correlation coefficient between the quadrature components identically equals zero .
we obtain the generating function @xmath17 of factorial moments @xmath18 , and use it to find the analytical expressions for the photon means , dispersions , and mandel s @xmath19-parameter .
besides , we calculate the cumulants @xmath20 of the pdf and demonstrate the oscillatory behaviour ( essentially nonclassical ) of the ratio @xmath21 .
the parameter @xmath22 was shown in @xcite to play an essential role in analysing the particle multiplicity distributions in high energy physics , since in experiments it demonstrates an oscillatory behaviour , which is sensitive to the detailes of the state of particles created in high energy collisions @xcite .
for the squeezed and correlated states of the electromagnetic field this parameter was discussed in @xcite . in the present article we concentrate on studying the behaviour of @xmath22 for the schrdinger real and imaginary cat states
using the well known expansion @xmath23 it is easy to derive the corresponding expansions for rcs and ics : @xmath24 @xmath25 where @xmath26 , and @xmath27 is the vector of the fock space .
the average values of the ladder operators @xmath6 , @xmath28 in real and imaginary cat states are invariant with respect to the hermitian conjugation : @xmath29 consequently , the quadrature operators @xmath30 , @xmath31 ( we assume @xmath32 ) @xmath33 possess the following mean values : @xmath34 for the elements of the quadrature covariance matrix we get @xmath35 @xmath36 @xmath37 the last equality shows that both real and imaginary schrdinger cat states have no correlation between the quadratures .
being rewritten in terms of @xmath5 and @xmath38 , the quadrature variances read @xmath39 @xmath40 the variance of the quadrature @xmath30 always exceeds @xmath41 . as to the quadrature @xmath31
, it is squeezed provided the inequality @xmath42 holds . assuming @xmath43 to be small
, one can expand the summands in powers of @xmath43 .
for the real cat state the limit of the inequality is @xmath44 , for any phase @xmath38 .
the analogous limit for the imaginary cat state is less than zero for arbitrary phase .
thus , there is no quadrature squeezing in the ics when @xmath43 tends to zero . on the contrary
, the squeezing in rcs exists for any @xmath38 up to @xmath45 .
using expansions ( [ 5 ] ) and ( [ 6 ] ) one gets the photon distribution functions @xmath46 it is obvious that @xmath47 , being a modulated poissonian distribution , is the sum of three poissonian distributions with complex means , which became the usual poissonian distribution with @xmath48 at real @xmath15 .
oscillations of pdf are demonstrated in figs . 1 and 2 both for rcs ( @xmath49 ) and ics ( @xmath50 ) in the representative cases of @xmath51 ( fig . 1 ) and @xmath52 ( fig .
2 ) , @xmath53 . the simplest way to derive the statistical characteristics of the field ( in particular , @xmath54 and @xmath55 ) is to differentiate the generating function @xmath56 with respect to the auxiliary real parameter @xmath57 .
then the factorial moments read @xmath58 the factorial moments up to the @xmath59-th order are plotted in fig . 3 for rcs ( @xmath49 ) at @xmath51 and @xmath53 .
the plot of @xmath18 for ics at the same @xmath15 seems similar to fig .
it is clear that in the case of @xmath60 expression ( [ 17 ] ) tends to the classical ( poissonian ) limit @xmath61 .
nevertheless , the pdfs in both rcs and ics dramatically differ from the poissonian distribution for all values of @xmath5 .
this means , in particular , that the knowledge of a few of the lowest factorial moments is not sufficient to decide whether the statistics is poissonian or not . in the limit @xmath62
the difference becomes clear only for the moments having the orders not less than @xmath63 . in the case of @xmath52 and @xmath53
we have @xmath64 .
the mean photon number equals @xmath65 whereas mandel s
@xmath19-parameter @xmath66 is equal to @xmath67 however , in the case under study this parameter is not suitable for the analysis of the photon statistics , since for large values of @xmath68 it is close to zero , although statistics remains essentially non - poissonian .
actually the equality @xmath69 is only necessary , but not sufficient condition of the poissonian statistics .
more informative characteristics of the photon statistics are the cumulants @xmath70 to find the relation between the moments and cumulants note that @xmath71 thus , @xmath72 putting @xmath73 in ( [ 21 ] ) one gets the recursive relation @xmath74 followed by another one , which is useful for the numerical calculation of the cumulant of the @xmath75-th order through the cumulants of preceding orders and moments up to the @xmath75-th order , @xmath76 in particular , @xmath77 k_2&=&f_2-f_1 ^ 2,\\[3 mm ] k_3&=&f_3 - 3f_2f_1 + 2f_1 ^ 3,\\[3 mm ] k_4&=&f_4 - 4f_3f_1 - 3f_2 ^ 2 + 12f_1 ^ 2f_2 - 6f_1 ^ 4,\\[3 mm ] k_5&=&f_5 - 10f_2f_3 - 5f_1f_4 + 20f_1 ^ 2f_3 + 30f_2 ^ 2f_1 -60f_1 ^ 3f_2 + 24f_1 ^ 5,\\[3 mm ] k_6&=&f_6 - 6f_1f_5 - 15f_2f_4 + 120f_1f_2f_3 + 30f_1 ^ 2f_4 - 120f_1 ^ 3f_3\\ & & -270f_2 ^ 2f_1 ^ 2 + 360f_1 ^ 4f_2 - 10f_3 ^ 2 + 30f_2 ^ 3 - 120f_1 ^ 6 .
\end{aligned}\ ] ] the cumulants of the photon number up to @xmath78 corresponding to @xmath79-state with @xmath51 , @xmath53 are plotted in fig .
4 ( @xmath49 ) to demonstrate the fast ( factorial - like ) divergence of a series of @xmath20 . since the cat states with @xmath80
are inherently nonclassical , the cumulants differ from zero even for small values of @xmath75 . their oscillating behavior is demonstrated in fig .
4 ( @xmath50 ) . the ratio @xmath21 may be used instead of @xmath19-parameter to evaluate how `` nonclassical '' is a quantum state @xcite .
its oscillatory behavior in rcs and ics with @xmath51 and @xmath52 , @xmath53 is seen from fig .
the strong decrease at higher @xmath75 is also seen : the factorial - like divergence of @xmath20 surpasses the quasiexponential growth of @xmath18 . the high amplitude oscillations of @xmath21 for cat states with @xmath52 admit a simple explanation : while @xmath18 equals approximately the classical value @xmath61 , the cumulants @xmath81 .
nevertheless , for large @xmath75 the ratio @xmath82 damps even for @xmath52 .
note that the plot of @xmath22 for the imaginary cat state with @xmath52 is symmetrical to the analogous plot for the real cat state with respect to @xmath75-axis .
we have shown that both real and imaginary schrdinger cat states manifest quadrature squeezing and non - poissonian statistics .
the factorial moments of the photon distribution function have proven to be monotonically growing , while the oscillatory factorial divergence to infinity is observed for the cumulants @xmath20 when @xmath83 .
the ratio @xmath82 has been found to be strongly oscillating for small @xmath75 and damping when @xmath75 grows .
these properties confirm the nonclassical nature of the schrdinger real and imaginary cat states .
if one considers the time evolution of quantum states in the case of a harmonic oscillator with unit frequency , then rcs and ics will be transformed as @xmath84 ( with the same complex exponentials in both terms ) .
thus in the time - dependent case the complex parameters of the coherent states determining the superpositions do not preserve the property to be mutually complex conjugated .
however , the phase difference @xmath14 between these parameters is time invariant . since the pdf is also time invariant , it is convenient and natural to choose the initial moment in such a way that the centers of the two coherent states wave packets would be located symmetrically with respect to the real axis
. just this natural choice has been made in the paper .
the multimode generalization of the real and imaginary cat states may be constructed analogously to the multimode generalization of the even and odd coherent states done in @xcite .
we are greateful to the esprit br project 6934 quintec for the support . 99 r.j . glauber , phys .
131 ( 1963 ) 2766 .
w. heisenberg , z. phys .
43 ( 1927 ) 172 .
d. stoler , phys .
d 1 ( 1970 ) 3217 .
walls , nature 306 ( 1983 ) 141 .
agarwal and g. adam , phys .
rev . a 39 ( 1989 ) 6259 .
dodonov , e.v .
kurmyshev and v.i .
manko , phys .
lett . a 79 ( 1980 ) 150 .
e. schrdinger , ber.kgl.acad.wiss .
berlin 24 ( 1930 ) 296 .
robertson , phys.rev .
35 ( 1930 ) 667 .
dodonov , o.v .
manko , v.i .
manko and l. rosa , phys .
lett . a 185 ( 1994 ) 231 .
dodonov , o.v .
manko and v.i .
manko , phys .
a 49 ( 1994 ) 2993 .
lugiato and a.gatti , phys .
70 ( 1993 ) 3868 .
e. schrdinger , naturwissenschaften 23 ( 1935 ) 844 .
b. yurke and d. stoler , phys .
57 ( 1986 ) 13 .
hach iii and c.c .
gerry , phys .
a 49 ( 1994 ) 490 .
p. adam , i. fldesi and j. janszky , phys .
a 49 ( 1994 ) 1281 .
dodonov , i.a .
malkin and v.i .
manko , physica 72 ( 1974 ) 597 .
m. brune , s. haroche , j.m .
raimond , l. davidovich and n. zagury , phys .
a 45 ( 1992 ) 5193 .
reid and l. krippner , phys .
a 47 ( 1993 ) 552 .
o. castans , r. lpez and v.i .
manko , in proceedings of the yamada international conference on group theoretical methods in physics , osaka , japan , 1994 ( to be published by world scientific ) ; journal of russian laser research ( to be submitted ) .
w. schleich , j.p .
dowling , r.j .
horowicz and s. varro , in new frontiers in quantum optics and quantum electrodynamics , edited by a. barut ( plenum , new york , 1990 ) .
dremin , modern physics letters a 8 ( 1993 ) 2747 .
dremin , uspekhi fiz .
nauk 164 ( 1994 ) 785 .
dodonov , i.m .
dremin , p.g .
polynkin and v.i .
manko , phys .
lett . a 193 ( 1994 ) 209 .
ansari and v.i .
manko , phys . rev . a 50 ( 1994 ) 1942 . | we study the photon statistics in the superpositions of coherent states @xmath0 and @xmath1 named `` schrdinger real and imaginary cat states '' .
the oscillatory character of the photon distribution function ( pdf ) emerging due to the quantum interference between the two components is shown , and the phenomenon of the quadrature squeezing is observed for the moderate values of @xmath2 . despite the quantity @xmath3 tends to the unit value ( like in the poissonian pdf ) at @xmath4 , the photon statistics is essentially non - poissonian for all values of @xmath5 .
the factorial moments and cumulants of the pdf are calculated , and the oscillations of their ratio are demonstrated .
= -1 cm |
one milestone in black hole accretion is the standard thin disk model ( shakura & sunyaev 1973 ; pringle 1981 ) .
this model applies below the eddington accretion rate defined as @xmath10 ( @xmath11 is the eddington luminosity ) . in this model ,
all the viscously dissipated energy is immediately radiated away .
when the accretion rate is above the eddington rate , advection begins to become important and the accretion model is described by the `` slim disk '' ( abramowicz et al .
1988 ; see also begelman & meier 1982 ) . in a slim disk ,
the radiative efficiency is lower than that of the standard thin disk because of energy advection and photons trapping effects .
the energy dissipated in the disk is advected with the accreting matter , since the radiative diffusion timescale is longer than the accretion timescale .
note that the slim - disk model can not correctly treat the photon trapping , because photon trapping is basically a multi - dimensional effect ( ohsuga et al . 2002 ; 2003 ) .
the potential applications of slim - disks include narrow - line seyfert galaxies ( mineshige et al .
2000 ) and ultraluminous x - ray sources ( watarai et al . 2001 ; vierdayanti et al .
2006 ) .
the above - mentioned pioneer works on slim disk are all one dimensional and analytical .
multi - dimensional and time - dependent numerical simulations obviously can reveal important additional information about the dynamics of the accretion flow .
many radiation - hydrodynamic ( rhd ; eggum et al .
1987 , 1988 ; okuda 2002 ; okuda et al . 2005 ; ohsuga et al .
2005 ) and radiation magnetohydrodynamic ( mhd ; ohsuga et al . 2009 ; ohsuga & mineshige 2011 ) numerical simulation works on slim disk have been performed . among these works , ohsuga et al .
( 2005 ) obtain a quasi - steady structure of the supercritical accretion flows and outflows by a two - dimensional global simulation .
their results broadly confirm the main properties predicted by the analytical slim - disk model .
moreover , they show that the accretion flow is composed of the disk region around the equatorial plane and the outflow region above and below the disk .
ohsuga & mineshige ( 2007 ) further identify that the supercritical accretion is feasible because a very large radiation energy density actually produces a small radiative flux as well as a force , because of the large optical depth and photon trapping effects .
recently , there are some works studying the thermal stability of radiation pressure dominated thin disk using shearing box mhd numerical simulation with radiative transfer ( hirose et al .
2009 ; jiang et al .
hirose et al.(2009 ) found that the disk is thermally stable , while jiang et al .
( 2013 ) found that it is unstable .
the reason for the discrepancy is discussed in the latter work . in the rhd simulation of ohsuga
( 2005 ) , an anomalous shear stress is included to mimic the angular momentum transfer . however , in reality , we expect maxwell stresses associated with mhd turbulence driven by the magneto - rotational instability ( mri ) to provide angular momentum transport in accretion flows ( balbus & hawley 1998 ) . since mri is driven only by shear associated with the orbital dynamics , when an anomalous shear stress is adopted we should only set the two azimuthal components ( @xmath12 and @xmath13 ) of the stress tensor to be non - zero ( stone et al.1999 )
. this is confirmed by the local shearing box simulations , which indicate that the azimuthal components of the maxwell stress are one order of magnitude larger than the poloidal components ( e.g. , hawley et al.1995 , 1996 ; stone et al .
1996 ) . in ohsuga
( 2005 ) only the @xmath12 component of the viscous stress is included , while the @xmath13 component is neglected .
so our first aim in this paper is to examine the effect of including the @xmath13 component on the dynamics of the slim disk .
it should play an important role in the angular momentum transport between different latitudes and might suppress the kelvin
helmholtz instability on the boundary between the outflow regions and the disk .
another aim is to study the effect of the magnitude of @xmath1 .
ohsuga et al.(2005 ) only consider one value of @xmath1 .
previous hd numerical simulations of non - radiative accretion flows have shown that the structure and dynamics of accretion flow depend on the value of @xmath1 ( igumenshchev & abramowicz 1999 ; stone et al .
1999 ; yuan et al .
the radial profile of inflow rate becomes significantly flatter when @xmath1 becomes larger ( yuan et al . 2012 ) .
different than the non - radiative accretion flow , a slim disk is dominated by radiation .
it is thus unclear whether or how the dynamics of accretion flow depend on the value of @xmath1 .
yet another aim of the present work , which is perhaps more interesting , is to examine the radial profiles of inflow rate . in almost all analytical models of accretion disks ,
the mass accretion rate is assumed to be constant with radius .
the validity of this assumption , however , has never been proved .
for the hot accretion flows , both hd and mhd numerical simulations have found that the mass inflow rate ( refer to equation ( 4 ) for definition ) decreases inward ( e.g. , stone et al .
1999 ; stone & pringle 2001 ; yuan et al .
2012 , and references therein ) .
this is one of the most important findings of global simulation of accretion flows , because this result supplies an important clue to revealing the dynamics of accretion flow .
various models have been proposed to explain this result such as adiabatic inflow
outflow solution ( blandford & begelman 1999 , 2004 ; begelman 2012 ) and convection - dominated accretion flow ( cdaf ; narayan et al .
2000 ; quataert & gruzinov 2000 ) . in the former ,
it is _ assumed _ that the inward decrease of inflow rate is because of mass loss in the outflow , while in the latter the flow is assumed to be convectively unstable , and a convective envelope solution is constructed that can also explain the simulations .
no outflow is needed in cdafs .
recent numerical simulations have shown that mhd accretion flows are convectively stable ( narayan et al .
2012 ; yuan et al .
moreover , by comparing the properties of inflow and outflow on the base of their hd and mhd numerical simulation data , yuan et al .
( 2012 ) argue that mass outflow should be significant .
they propose that inward decrease of accretion rate is due to outflow .
( 2013 ) obtained a similar conclusion . in most of the published papers , the inflow and outflow rates ( equation(4 ) and ( 5 ) )
are computed at each instant of time by using instantaneous velocities , and then the time - averaged .
narayan et al .
( 2012 ; see also sadowski et al .
2013 for the case of spin black holes ) , however , found that the outflow rate is much lower since they think the outflow rate should be calculated by doing the time - average first ( see yuan et al .
2012 for more discussions on the discrepancy of the two approaches of calculating the outflow rate ) . while it is agreed that outflow exists , so far the strength of outflow is still an open question .
f.yuan et al .
( 2013 , in preparation ) have also studied the mechanism of producing outflow .
it was found that in the hd case it is buoyancy associated with the convective instability , while it is mainly magnetic centrifugal force in the case of mhd accretion flow .
it is obviously interesting to see whether the radial profile of the inflow rate of a slim disk behaves like a hot accretion flow , and if it does , whether the outflow is produced by the buoyancy associated with the convective instability .
ccccccccccccccccc models & run&@xmath1 & @xmath14 & @xmath15 & stress tensor & @xmath16 & t@xmath17(orbits)&t@xmath18 ( orbits ) & @xmath19 & l + a & 1a & 0.1 & 1000 & 0@xmath20 @xmath21 & @xmath22 & 96@xmath23225&11.5&46.1&184.5&2.4 + b & 1b & 0.1 & 1000 & 0@xmath20 @xmath24 & @xmath22,@xmath0&96@xmath23225 & 11.5&46.1 & 160.0 & 2.2 + & 2b & 0.1 & 3000 & 0@xmath20 @xmath24 & @xmath22,@xmath0&96@xmath23225 & 11.5&46.1 & 374.5 & 3.2 + & 3b & 0.05 & 3000 & 0@xmath20 @xmath24 & @xmath22,@xmath0&96@xmath23225 & 11.5&46.1 & 382.8 & 3.6 + & 4b & 0.01 & 3000 & 0@xmath20 @xmath24 & @xmath22,@xmath0&96@xmath23225 & 63.5&98.1 & 240.3 & 4.2 + & 5b & 0.005 & 3000 & 0@xmath20 @xmath24 & @xmath22,@xmath0&96@xmath23225 & 150.0&184.6 & 125.3&3.6 + the paper is organized as follows .
we describe our models and numerical method in section 2 .
the simulation results are presented in section 3.section 4 is devoted to a summary and discussions .
the rhd equations are the same as those in ohsuga et al .
( 2005 ) . in the rhd equations ,
the flux - limited diffusion ( fld ) approximation developed by levermore & pomraning ( 1981 ) is adopted .
we neglect the self - gravity of the disk and use the pseudo - newtonian potential to mimic the general relativistic effects , @xmath25 ( paczysky & wiita 1980 ) , where @xmath26 is the gravitational constant , @xmath27 is the mass of the black hole , @xmath28 the schwarzschild radius , and @xmath29 is the radius .
we assume that the gas is in local thermodynamic equilibrium and neglect the frequency dependence of the opacities . as we state in section 1
, we adopt a stress tensor @xmath30 to mimic the shear stress , which is in reality should be replaced by the magnetic stress associated with mhd turbulence driven by the mri . in most cases , following stone
et al . ( 1999 ) , we assume that the only non - zero components of @xmath30 are the azimuthal components : @xmath31 @xmath32 here , the dynamical viscosity coefficient @xmath33 is described as a function of the pressure @xmath34 , where @xmath35 is the keplerian angular speed , @xmath36 is the gas pressure , @xmath37 is the radiation energy density , and @xmath38 is the flux limiter ( levermore & pomraning 1981 ) .
the viscous dissipative function is given by @xmath39 .
all of our models are calculated in spherical coordinates ( @xmath29,@xmath15,@xmath40 ) .
the origin is set at a central black hole of @xmath41 .
the size of the computational domain is @xmath42 and @xmath43 or @xmath44 .
the inner boundary ( @xmath45 ) must be smaller than the sonic point of the accretion flow , which ensures that the inner boundary conditions do not affect the simulation results .
abramowicz et al . ( 1988 ) show that for a slim disk the location of the sonic point depends on the accretion rate and the viscosity parameter @xmath1 . for small @xmath1 and a high accretion rate ,
the sonic point locates in the range of @xmath46 .
so we set @xmath47 .
the computational domain is divided into @xmath48 grid cells .
a non - uniform grid is employed in the @xmath29 direction .
the grid points in the @xmath29 direction are equally distributed logarithmically , i.e. , @xmath49 constant . in the @xmath15 direction , in order to better resolve the flow at the equator and to not lose the resolution at the axis , we adopt the mixed grid .
twenty grids are uniformly distributed within the @xmath50 from the axis , i.e. , @xmath51 .
other grids are distributed in the angular range of @xmath52 or @xmath53 in such a way that @xmath54 or @xmath55
. the outflow boundary condition is adopted at the inner radial boundary , i.e. the values of physical variables in the ghost zones are set to the values at the inner radial boundary .
the outer radial boundary condition is the same as that employed by ohsuga et al .
( 2005 ) , who suggested that the matter , having a specific angular momentum corresponding to the keplerian angular momentum at @xmath56 , is continuously injected into the computational domain from the outer boundary near the equator .
the injected gas distributes within 0.05@xmath24 from the equator in the gaussian function ; their radial velocity is set according to equation(3.61 ) in kato et al .
( 1998 ) , while their poloidal velocity is set to be zero . in the angular direction , we employ axisymmetry relative to the axis and reflection symmetry relative to the equator ( when @xmath57 ) , respectively . the computational domain is initially filled with a hot , rarefied , and optically thin atmosphere .
the numerical approach can be found in ohsuga et al .
( 2005 ) .
[ 0.90 ] [ 0.90 ] the properties of all of the simulations are listed in table 1 , where columns 1 and 2 give the classification of our models and their number ; column 3 gives the value of @xmath1 ; column 4 gives the mass injection rate ( @xmath58 ) in units of the critical mass accretion rate , @xmath59 ; columns 5 , 6 , and 7 give the computational domain in the @xmath15 direction , the components of stress tensor , and the quantity of grid cells , respectively ; column 8 gives the approximate dynamical time @xmath60 ( all times in this paper are reported in units of the orbital time at @xmath29=100@xmath61 ) after which the accretion flows become quasi - steady ; column 9 gives the final time @xmath62 ; and columns 10 and 11 give the mass accretion rate onto the black hole ( @xmath63 ) in the units of the critical mass accretion rate and the luminosity ( @xmath64 ) in the units of @xmath65 , respectively .
according to the component of stress tensor and the angular range of computational domain , the simulations are divided into two groups , namely models a and b. models a contains only the @xmath12 component of stress tensor , while model b contains both the @xmath12 and @xmath13 components .
figure [ fig 0 ] displays a two - dimensional density distribution overlaid with velocity vectors .
the top and bottom panels are for @xmath66=92.541 orbits of run 2b and 254.085 orbits of run 5b , respectively .
[ 0.65 ] [ 0.65 ] [ 0.65 ] [ 0.65 ] [ 0.65 ] [ 0.65 ] [ 0.65 ] [ 0.65 ] [ 0.65 ]
to investigate the effect of @xmath0 on the dynamics of accretion flow , we compare models a and b. figure [ fig 1 ] shows the angular structure of time - averaged variables at @xmath67 and 30 @xmath68 for run 1a and run 1b with @xmath69 . in this paper ,
all the time - averaged quantities are obtained by averaging 100 data files within 23 orbits after the accretion flows have achieved the quasi - steady state . in figure
[ fig 1 ] , the black solid ( @xmath70 ) and dashed ( @xmath71 ) lines correspond to run 1a while the red solid ( @xmath70 ) and dashed ( @xmath71 ) lines correspond to run 1b .
the figure shows obvious differences of the flow structure , especially at the high - latitude region .
the angular profiles of angular velocity ( @xmath72 ) show that the flow in run 1a rotates faster than that of run 1b at the high - latitude region .
close to the equator , the rotation of run 1a is slightly slower than run 1b . because @xmath0 transports angular momentum between different latitudes , for run 1a the angular momentum between different latitudes can not be transported , although the high - latitude flow rotates faster than the low - latitude one .
therefore , it is seen that the angular velocity increases from the equator to almost the axis ( although the angular velocity at the rotating axis is set to be zero ) .
when @xmath73 is included in run 1b , the angular momentum can be transported from the quickly rotating flows at high latitude to the slowly rotating flows at low latitude .
this is why the discrepancy of the rotation velocity between the equator and the high - latitude region is smaller in run 1b than in run 1a .
the discrepancy of the rotation velocity at the equator between run 1a and run 1b is much smaller than in the high - latitude region , since the density at the high - latitude region is much smaller than in the equator .
the angular profiles of radial velocity ( @xmath74 ) show obvious differences between run 1a and run 1b , especially at the high - latitude region .
the positive value of @xmath74 indicates that the flows are outflows , while the negative value indicates the inflow nature of flows . at large radii such as @xmath75 ,
the high - latitude flow ( the flow within @xmath76 from the axis ) is mainly outflow .
the speed of the high - latitude outflow of run 1a is higher than that of run 1b . at small radii such as @xmath77 ,
the flow is inflow at all latitudes for both models .
the angular profiles of radial velocity agree with the angular profiles of radial forces , as shown in figure [ fig 1 ] .
the radial radiation force dominates the radial component of net force at the high - latitude region at @xmath78 ; therefore , it is the dominant force changing the angular distribution of the radial velocity there . at small radii
, gravity is the dominant force .
the angular profiles of density ( @xmath79 ) of run 1a and run 2b are similar , with the maximum density located at the equator .
compared with run 1a , the density of the high - latitude region of run 1b is higher , while at the equator is slightly lower .
the angular profiles of gas temperature ( @xmath80 ) show that the disk of run 1a and run 1b has nearly the same gas temperature at @xmath81 . for run 1b the gas temperature of high - latitude outflow
is lower than that of run 1a . for the high - latitude outflow ,
the lower the gas temperature , the higher the density .
[ 0.67 ] [ 0.67 ] [ 0.67 ] [ 0.67 ] [ 0.67 ] [ 0.67 ] [ 0.66 ] [ 0.66 ] [ 0.66 ] [ 0.66 ] the angular profiles of the radiation " temperature ( @xmath82 , where @xmath83 is the radiation constant ) show that compared with run 1a , the radiation temperature for run 1b is lower near the equator and higher near the rotating axis . for the super - critical accretion flow , the radiation force ( including also the gradient of the radiation pressure ) is the dominant force driving the high - latitude outflow .
the radiation force acting on unit mass is given by @xmath84 , where the flux - mean opacity @xmath85 is the sum of components due to absorption and scattering .
the scattering in the high - latitude outflow is the dominant factor for the opacity so that @xmath86 and @xmath87 .
here we employ the approximation and have @xmath88 . in the optically thin limit , @xmath89 . in general , the high - latitude outflow has high gas temperature and low density , so the flow is optically thin . for run 1a ,
the opacity of outflow near the rotating axis is close to the optically thin limit , while for run 1b the opacity of outflow is away from the optically thin limit because of the higher density .
therefore , although run 1a has lower radiation energy density at the high - latitude region than run 1b , the radiation flux of run 1a is larger than that of run 1b , which is shown by the angular profiles of radial radiation flux in figure [ fig 1 ] .
therefore , at the high - latitude region the radiative force of run 1a is larger , exceeding that of run 1b .
compared with run 1a , the density of high - latitude outflow for run 1b is higher so that the driving force of high - latitude outflow is weaker . for the strict steady state and for inviscid hydrodynamic flow , the bernoulli parameter be ( be @xmath90 ) is conserved along the streamline .
therefore , the positive sign of be is often used to be the necessary condition for the outflow to escape to infinity . however , in our case , these conditions are not satisfied , thus be is no longer conserved ( e.g. , yuan et al .
in fact , the initial result from our ongoing work indicates that be can increase along the trajectory of outflow elements ( f.yuan et al .
2013 , in preparation ) .
this means that an outflow with a negative be can also potentially escape to infinity . despite these uncertainties ,
we still show in figure [ fig 2 ] the angular distribution of the time - averaged bernoulli parameter be because its value may still play some role in determining the properties of outflow .
we can see that at the region close to the axis ( @xmath91 ) and @xmath92 , the value of be in run 1b is about one order of magnitude lower than that in run 1a . from figure
[ fig 1 ] , we can see that including @xmath0 for the polar outflows , rotational speed , radial speed and temperature decrease can decrease the specific be of outflow . in general , the value of be is larger close to the axis . when compared with
the hydro and mhd numerical simulations of hot accretion flow presented in yuan et al .
( 2012 ) , it is interesting to note that the angular distribution of be is more similar to the mhd case ( their model d in yuan et al .
( 2012 ) ) rather than to the hydro case ( their models a , b , and c ) .
this is perhaps because the radiation acceleration in the @xmath29 direction is much stronger in the region close to the axis than in other directions . in the case of models
a , b , c in yuan et al .
( 2012 ) , there is no such force .
but in the mhd case , the magnetic pressure force plays a similar role with the radiation force here .
figure [ fig 3 ] shows the angular distribution of the time - averaged flow for model b ( with @xmath93 ; i.e. , run 2b and 5b ) at @xmath94 and @xmath95 , respectively .
we can see that the viscosity parameter @xmath1 obviously affects the angular distribution of accretion flow .
the angular profiles of angular velocity ( @xmath96 ) ) ) show a large difference close to the axis because of the different @xmath1 .
when @xmath1 is smaller , the flow rotates faster . a smaller @xmath1 results in a smaller @xmath73 ; thus , the angular momentum can not be efficiently transported from high to low latitudes .
the angular profile of the radial velocity ( @xmath97 ) is different for different @xmath1 .
the speed of the high - latitude outflow in run 15b is higher than that of run 2b . at @xmath95 ,
the high - speed outflow is constrained to be within @xmath98 from the axis . with the decrease of @xmath1 , the range of @xmath15 within which high - latitude outflow takes place becomes larger toward the equator .
at @xmath94 , the radial velocity is negative over most @xmath15 .
the plots of angular profiles of density , gas temperature , and `` radiation '' temperature show that these three quantities change rapidly with @xmath15 for run 5b . with the decrease of @xmath1
, the density and radiation energy density concentrate toward the equator , and the flow can be divided into two areas with different temperature , i.e.,a low - temperature area near the equator and a high - temperature area near the axis . [ 1.4 ] [ 1.4 ] [ 1.4 ] [ 1.3 ] [ 0.65 ] [ 0.65 ] [ 0.65 ] [ 0.65 ] in order to understand the angular profiles of radial velocity , we plot in figuer .
[ fig 4 ] the angular distribution of radial forces acting on unit mass for run 5b at @xmath99 ( top panel ) and @xmath100 ( bottom panel ) , respectively .
we can see that the angular distribution of net force ( solid lines ) is similar to the angular distribution of @xmath74 shown in figure [ fig 3 ] .
in addition , figure [ fig 4 ] shows that within @xmath101 the radial component of centrifugal force efficiently counteracts the gravity and even exceeds the gravity at some degree .
the top panel shows that the flows within @xmath98 from the equator are super - keplerian at @xmath102 . at @xmath103 ,
the flows are super - keplerian within @xmath104 from the equator . however , within @xmath105 from the axis , the radial component of centrifugal force rapidly decreases and can not efficiently counteract the gravity . hence , the equivalent potential well is deeper near the axis than near the equator .
on the other hand , the radial component of the gas - pressure gradient force is negligible within @xmath106 of the equator , while near the axis this component is not negligible but is not the dominant force .
the radial net force within @xmath106 of the equator is dominated by the radial radiation force , the gravity , and the radial centrifugal force . in the inner region ,
the radial net force near the axis is dominated by the gravity , so that we see that there is inflow near the axis . in the outer region ,
the radial net force near the axis is dominated by the radial radiation force , so that we can see that there is strong outflow near the axis .
hence , the radial radiation force plays an important role in maintaining the radial equilibrium of flows near the equator and driving the high - latitude outflow in the outer region .
figure [ fig 5 ] shows the angular distribution of angular forces .
it is seen that the angular component of net force is nearly zero near the equator , which indicates that flows are in force equilibrium near the equator .
this is because that the angular component of gas - pressure gradient force is also negligible , and the radiation force balances the centrifugal force in the @xmath15-direction .
the angular motion of flows near the axis is controlled by the centrifugal force and gas - pressure gradient force , while the angular component of radiation force is negligible near the axis .
[ 0.65 ] [ 0.65 ] [ 0.65 ] [ 0.65 ] the radial structure of a slim disk has been solved using a vertical - integrated one - dimensional method ( e.g. , abramowicz et al . 1988 ) .
however , the solution is one dimensional , thus the method can not treat more viscous components besides @xmath22 and can not address outflow . here ,
on the basis of the solution of two - dimensional simulation , we plot the radial structure of the time - averaged flow near the equator in figure [ fig 6 ] .
the solution is averaged over the angle between @xmath107 and @xmath108 . in figure
[ fig 6 ] , the left panel is for run 2b ( red line ) and run 3b ( blue line ) , while the right panel is for run 4b ( red line ) and run 5b ( blue line ) .
panel ( a ) of figure [ fig 6 ] shows the radial profiles of density .
the dot - dashed lines indicate the self - similar solution of the slim - disk model ( wang & zhou 1999 ) .
we find that the density profile of different models can be well described by a power law function , @xmath6 with @xmath109 .
this result is much flatter than the self - similar solution of the slim disk where @xmath110 ( wang & zhou 1999 ) , but it is very similar to the case of hot accretion flows ( stone et al .
1999 ; yuan et al .
2012 , and references therein ) .
the reason for the discrepancy is that in the self - similar solution the mass accretion rate @xmath5 is assumed to be a constant of radius , while as we will show in section 3.4 , the accretion rate actually decreases inward because of the mass loss in outflows .
moreover , it is interesting to note that the value of power law index of the density profile , @xmath7 , is quite `` universal '' for different models , although these models have different @xmath1 and different radial profiles of inflow rate ( refer to section[outflow ] ) .
( 2013 ) studied the effects of initial and boundary conditions in simulations of accretion flow .
they find a similar result , namely the density profile is more converged compared with the diverse radial profile of inflow rate .
our result is apparently different from figure 11 in ohsuga et al .
( 2005 ) , in which they find @xmath110 . for comparison with ohsuga et al .
( 2005 ) , we have checked the models in ohsuga et al .
( 2005 ) and found that only when @xmath111 and 1000 , and @xmath8 , the density profile can be described by @xmath110 .
all other models have @xmath112 .
moreover , in the former case , the @xmath113 density profile only holds in the range of @xmath114 . beyond this radius ,
the profile becomes much flatter .
but in the latter case , the @xmath115 density profile holds until @xmath116 , as shown by figure 6 .
the reason for the discrepancy is unclear .
[ 0.5 ] [ 0.5 ] panel ( b ) of figure [ fig 6 ] shows the radial profiles of gas temperature @xmath80 ( red and blue dashed lines ) and `` radiation '' temperature @xmath117 ( red and blue solid lines ) .
@xmath117 can be approximately described by a radial power law function of @xmath118 , as the black solid line shows .
this is again flatter than the self - similar solution of the slim - disk model , which has @xmath119 as shown by the dot - dashed line .
we think the reason is because the compression work becomes weaker because of the presence of outflow . in the inner region close to the black hole
, @xmath120 becomes flatter .
we can find that @xmath121 and @xmath120 are nearly equal in the outer region , while the gas temperature is higher than @xmath120 in the inner region .
the radius where the two temperatures deviate depends on the mass - injection rate and viscous parameter @xmath1 . the higher the mass - injected rate and the smaller the @xmath1 , the smaller the `` deviation '' radius .
however , when @xmath122 and 0.005 , the `` deviation radius '' is located around the inner boundary .
the discrepancy of the two temperatures is because of the inefficient coupling between gas and radiation .
the energy transfer between gas and radiation is controlled by the absorption opacity @xmath123 ( refer to the term @xmath124 ( @xmath125 is the blackbody radiation intensity ) of equations ( 7 ) and ( 8) in ohsuga et al .
the absorption opacity @xmath126 is due to free - free absorption and bound - free absorption .
we find @xmath127 , i.e. , @xmath123 decreases inward .
therefore , at a small radius , the coupling between radiation and gas is weak , and the temperature equilibrium between the gas and the radiation field is not achieved before the gas falls onto the black hole .
smaller @xmath1 gives rise to smaller radial velocity , which provides more time to transfer the energy of the gas to the radiation field .
so , the `` deviation radius '' moves inward with the decrease of @xmath1 .
panel ( d ) in figure [ fig 6 ] shows the radial distribution of the specific angular momentum .
we can see that @xmath1 can affect the angular momentum distribution , especially in the vicinity of the black hole ( @xmath12810@xmath61 ) . in the case of @xmath1=0.01 and 0.005 ,
the specific angular momentum becomes super - keplerian in the range of 3@xmath1296 @xmath68 .
@xmath14 also affects the angular momentum distribution .
the models with smaller @xmath1 and lower @xmath14 have slightly flatter distribution
. this result does not agree with that of abramowicz et al .
abramowicz et al . ( 1988 ) identified the tendency of the specific angular momentum distribution to become flatter with the increase of the accretion rate when the accretion rate is less than @xmath130 .
they could not study the higher accretion rate , because their method fails for @xmath131 .
the reason for the discrepancy is unclear . to analyze the energetics of all the models , we define the advection factor ( @xmath132 ) of accretion flow as follows : @xmath133 where @xmath134 is the gas and radiation energy advection rate ,
@xmath135 is the viscous dissipation rate , and @xmath136 is the radiation cooling rate . here
@xmath137 is defined as the @xmath15 component of @xmath138 ( @xmath139 is the radiation flux ) .
both @xmath140 and @xmath141 are obtained first by time - averaging and then by averaging the quantity over an angle between @xmath107 and @xmath142 .
figure [ fig 7 ] shows the time - averaged advection factor near the equator for the models of @xmath93 ( i.e. , runs 2b , 3b , 4b , and 5b ) .
we can see that @xmath143 is approximately close to a constant in the range of @xmath144 .
the value of @xmath143 deceases inward when @xmath145 , especially for the models with small @xmath1 .
since the opacity of accretion flow is dominated by scattering opacity and the half - thickness of accretion decreases inward , the vertical opacity of accretion flow decreases inward .
therefore , the photon trapping effect becomes weaker and the radiation becomes stronger .
@xmath146 can even be larger than @xmath141 , so @xmath143 is negative , as shown by figure [ fig 7 ] .
this is consistent with abramowicz et al .
this indicates that advection plays a heating rather than cooling role , similar to the case of luminous hot accretion flows ( yuan 2001 ) .
[ 0.5 ] [ 0.5 ] following stone et al .
( 1999 ) , we define the mass inflow ( @xmath147 ) and outflow rates ( @xmath148 ) , in units of the critical mass accretion rate @xmath149 , as the following time - averaged and angle - integrated quantities : @xmath150 @xmath151 the net mass accretion rate , @xmath152 , is the accretion rate that finally falls onto the black hole . it is noted that the above rates are obtained by time - averaging the integral rather than integrating the time - averages . according to equation ( 5 ) , we know that the outflow rate is not a correct measurement of real outflow , although it can provide some important information .
this is because it includes the contribution of turbulent motion .
figure [ fig 8 ] shows the radial distribution of inflow , outflow , and net rates of model b ( @xmath153 ; i.e. , runs 2b , 3b , 4b , and 5b ) .
we can see that the inflow and outflow rates decrease inward , as in the case of hot accretion flows ( stone , pringle et al .
1999 ; yuan et al .
2012 , and references therein ) .
the radial profiles of @xmath154 and @xmath155 can be described by a power law function of radius . using @xmath156 to fit the inflow rate in the range of @xmath157
, we find that the value of @xmath3 is not sensitive to @xmath14 but mainly determined by the value of @xmath1 .
a large @xmath1 corresponds to a small @xmath3 .
we have calculated models with different @xmath58 and derived the average values of @xmath3 .
they are @xmath158 , and @xmath159 for models with @xmath160 , and @xmath9 , respectively .
the dependence of @xmath3 on @xmath1 can be approximately described by a power law function @xmath161 .
this result is again similar to the case of hot accretion flows .
yuan et al .
( 2012 ) find that @xmath162 and @xmath163 for @xmath122 and @xmath164 , respectively .
note that the definition of @xmath1 is different in the two works .
we find that the value of @xmath1 in this paper corresponds to a smaller @xmath1 in yuan et al .
the quantitative comparison of the value of @xmath1 is difficult . as argued in yuan et al .
( 2012 ) and f. yuan et al .
( 2013 , in preparation ) , in the case of hot accretion flow , the inward decrease of inflow rate is due to mass loss via outflows .
we believe this is also the case for the slim disk .
the similar slope in the slim disk and hot accretion flow implies that the strength of outflow is similar in the two cases .
we will discuss the possible origin of outflow below . in order to analyze the angular distribution of the mass accretion rate , we time - average the following inflow and outflow rates as a function of @xmath15 : @xmath165
@xmath166 similar to equation ( 5 ) , equation ( 7 ) also does not correctly measure the angular distribution of real outflow .
the results are shown in figure [ fig 9 ] for model b. we can see that their angular distributions are nearly symmetric to the equator and become broader with the increase of radius and/or @xmath1 .
the distribution is different than the case of hot accretion flows ( refer to figures 2 and 3 in yuan et al .
firstly , in the case of slim disks , the inflow and outflow rates are `` synchronous '' , i.e , they reach their maximum at the same @xmath15 angle .
but in the case of hot accretion flows , they are not ( compare figure [ fig 9 ] in the present paper and figures 2 and 3 in yuan et al .
2012 ) . secondly , the outflow rate centers around the equatorial plane in the present case , while in the case of hot accretion flows , the maximum of outflow rate is located roughly at the surface of the disk ( their model b and model c ) . in that case , the inflow and outflow rates are not symmetric to the equatorial plane .
the reason for the discrepancy is unclear .
the outflow becomes stronger with the increase of radius .
this is consistent with figure [ fig 8 ] and is easy to understand .
compared with the region around the equator , the inflow and outflow rates are almost negligible at the high - latitude region , where the radial velocity of outflow is very high .
this is because the density of high - latitude flows is very low .
the high - latitude outflow is likely driven by radiation , but its contribution to outflow rate seems to be negligible .
the outflow rate is dominated by the low - latitude outflow .
the nature of the low - latitude outflow , namely whether they are real systematic outflow or simply turbulence , is an important question and needs to be studied in the future .
another question is whether the origins of outflow are real . in the case of hot accretion flows , yuan et al .
( 2012 ) identify the mechanism of producing the outflow by buoyancy when magnetic field is absent .
this is because a hydro accretion flow is convectively unstable . in section 3.5 of the present paper
, we analyze the convective stability of slim disks .
following yuan et al .
( 2012 ) , we analyze some properties of inflow and outflow , such as the bernoulli parameter , gas temperature , radial velocity , and angular momentum . the motivation is to study the mechanism of producing outflow .
we calculate the flux - weighted quantities ( refer to equations ( 8) and ( 9 ) in yuan et al .
2012 ) and then time - average the quantities .
figure [ fig 10 ] shows the radial distribution of flux - weighted quantities ( @xmath167 in units of @xmath168 , where @xmath169 is the keplerian velocity ; @xmath80 in units of the virial temperature @xmath170 , where @xmath171 is the photon mass and @xmath172 the boltzmann constant;and @xmath173 in unit of @xmath169 ) of inflow and outflow for run 2b and run 5b , respectively . in both models ,
be is negative . the value of be is in general smaller than that of hot accretion flows ( yuan et al .
2012 ) , which is of course because the energy loss in slim disk is stronger .
the top right panel in figure [ fig 10 ] shows that the temperature of outflow is higher than that of inflow .
this seems to suggest that the mechanism of outflow production is because of buoyancy , like in the case of hydrodynamical hot accretion flow .
moreover , the discrepancy of the two temperatures is larger when @xmath1 is larger ( run 2b ) .
the convective stability of slim disk is analyzed in section [ convection ] . the bottom left panel in figure [ fig 10 ] shows that the radial velocity of outflow can be well described by @xmath174 const .
this is similar to hot accretion flows ( yuan et al .
it is interesting to note that when @xmath1 is smaller , the radial velocities of both inflow and outflow are smaller . for inflow , this is easy to understand . for outflow , the discrepancy of the radial velocity may be related to the discrepancy of the temperature between inflow and outflow . as we can see from the top right panel , when @xmath1 is smaller , the discrepancy is smaller , and thus the buoyancy may be weaker .
the reason why the temperature discrepancy in the case of small @xmath1 is smaller is perhaps related to the magnitude of the convective energy flux .
the convective energy flux transports energy along surfaces of constant @xmath175 ( quataert & gruzinov 2000 ) .
the magnitude of this flux is proportional to the convective diffusion coefficient @xmath176 , and a smaller @xmath1 corresponds to a smaller @xmath176 ( narayan et al .
therefore , when @xmath1 is larger , the fluids at the surface are more strongly heated by the stronger convective energy flux . at last , the bottom right panel shows that the angular momenta of inflow and outflow are almost identical .
this is similar to the case of hydrodynamical hot accretion flows ( yuan et al .
2012 ) . in the case of hydrodynamical hot accretion flow
, it is shown that the inward decrease of inflow rate is because of mass loss in outflow produced by the buoyancy associated with convective instability ( yuan et al . 2012 ) . in this section ,
we analyze the convective stability of the slim disk on the basis of our simulation data .
the energy equation of accretion flows can be written as @xmath177 where @xmath178 ( @xmath179 is surface density , and @xmath180 is the specific entropy ) . in general , @xmath181 so that @xmath182 and @xmath183 .
the inward increase of entropy is thought of as a necessary condition of convective instability for rotating flow .
a series of simulation studies ( igumenshchev & abramowicz 1999 ; stone et al .
1999 ; igumenshchev & abramowicz 2000 ; yuan & bu 2010 ) verified that hot accretion flows are convectively unstable , confirming the prediction of narayan & yi ( 1994 ) .
compared with hot accretion flows , the physics of a slim disk is different . in a slim disk ,
the gas particles can efficiently radiate , but the photons are trapped and hence can not efficiently escape from the system because of the large scattering optical depth .
in addition , a slim disk is supported by the radiation pressure , and hence the specific entropy is dominated by the radiation photons .
sadowski et al .
( 2009 , 2011 ) found that a radiation pressure - supported disk is convectively unstable .
gu ( 2012 ) revisited this problem , taking into account the local energy balance between the viscous heating and the advective and radiative cooling , and found that a slim disk is convectively stable .
gu ( 2012 ) thought that the significant difference in the results between their work and sadowski et al ( 2009 ) .
is probably related to the different approaches for describing the vertical structure .
but in gu ( 2012 ) , a self - similar solution of the radius is adopted , the consequence of which is unknown .
here we analyze the convective stability of a slim disk on the basis of our simulation data .
we assume that the hiland criterion ( e.g. , begelman & meier 1982 ) is applicable to the slim disk .
the condition in cylindrical coordinates ( @xmath184 ) for convective stability in a rotating accretion flow ( tassoul 1978 ) is @xmath185 where @xmath186 is defined as an effective frequency , @xmath187 is the epicyclic frequency which is calculated by @xmath188 ( @xmath189 is the specific angular momentum ) , and @xmath190 and @xmath191 is the @xmath192 and @xmath193 component of the well - known brunt visl frequency , respectively , which can be calculated by @xmath194 and @xmath195 where @xmath196 , @xmath197 is the total pressure , and @xmath198 is the adiabatic index . here @xmath199 and @xmath200 are employed for the flows of low - density ( @xmath201 g/@xmath202 ) and high - gas temperature ( @xmath203 k ) ( i.e. , outflow or corona region ) , while @xmath204 and @xmath205 are employed for other radiation - dominated flows ( i.e. , the disk body ) where radiation and gas are effectively coupled because of the large scattering opacity .
the results are shown in figure [ fig 11 ] .
the left panel is for run 2b at @xmath206 orbits , while the right panel for run 5b at @xmath207 orbits .
the dark region denotes the unstable region , i.e. , @xmath208 .
we can see that the results are somewhat subtle . for run 2b nearly half of the region
is convectively unstable , while for run 5b the fraction of the unstable region is much less . given this result , we can perhaps say that run 2b is marginally convectively unstable and run 5b is stable . in both cases ,
compared with the case of a hydrodynamical hot accretion flow ( refer to figure 6 in yuan & bu 2010 ) , the unstable region becomes significantly smaller .
we speculate that the difference between the present work and a hydrodynamical hot accretion flow ( i.e. , figure 6 in yuan & bu 2010 ) is due to radiation .
recall that in an magnetohydrdynamical hot accretion flow , the magnetic field plays a role of stabilizing the convection ( balbus & hawley 2002 ; narayan et al .
2002 ) . here
radiation plays an effectively similar role although the underlying physics may be different .
physically , when radiation is important as in a slim disk , radiation can directly transport energy , thus the `` need '' for convection to transport energy becomes weaker .
why is the unstable region in run 2b larger than that in run 5b ?
the only difference between run 2b and run 5b is the value of @xmath1 , which is 0.1 and 0.005 , respectively .
compared with the case of a large @xmath1 , when @xmath1 is smaller , the convective energy flux is smaller , thus the role of radiation to stabilize the flow becomes relatively more effective .
this explains the difference between run 2b and run 5b .
this may also explain why the temperature discrepancy between inflow and outflow in run 2b is larger than in run 5b ( refer to figure [ fig 10 ] ) . as an initial step of the application of our numerical simulation , in this section
we compare the emitting spectra from two - dimensional numerical simulations and one - dimensional analytical global solutions . for simplicity , we assume a multi - color blackbody spectrum on the basis of effective temperatures . in reality , however , compton scattering plays an important role in the emitted spectrum ( kawashima et al .
we identify the radiation temperature @xmath117 at the photosphere ( where the effective optical depth equals 1 ) as the effective temperature .
the effective optical depth is calculated by @xmath209 , where @xmath210 and @xmath211 are the absorption and thomson scattering optical depth integrated from the outer boundary along the _ z_-direction .
we adopt the free free absorption opacity ( @xmath212 ) .
taking run 5b as an example , we have calculated the emitted spectrum based on the simulation data .
only the range of @xmath213 is considered .
the reason is that only the region of @xmath214 has achieved steady state while the region of @xmath215 is effectively optically thin due to small absorption optical depth caused by high gas temperatures .
we also calculate the global solutions of one - dimensional model corresponding to the parameters of run 5b(refer to abramowicz et al .
1988 and watarai et al .
2000 for the calculation approach ) and the corresponding spectrum .
figure [ fig 12 ] shows the calculation results .
we note that the calculated luminosity based on our two - dimensional numerical simulation agrees well with that obtained by kawashima et al .
significant differences of temperature and subsequently emitted spectrum between the two models can be found .
this result can give us some initial idea of the difference between simple one - dimensional calculation and the more realistic two - dimensional simulation , which should be useful when applying the slim disk theory to observations .
the detailed calculation of the spectrum can be found in kawashima et al .
( 2012 ) and beyond the scope of the present work .
in this paper , we have performed a two - dimensional radiative hydrodynamical numerical simulation of slim disks .
the technical differences between the present work and ohsuga et al .
( 2005 ) are that we include an additional component of viscous stress , i.e. , @xmath0 , and consider various values of the viscous parameter @xmath1 .
we find that the component @xmath0 plays an important role in transporting the angular momentum between different latitudes . as a result , compared with the case of no @xmath0 component ( ohsuga et al .
2005 ) , the high - latitude outflow ( within @xmath76 from the axis ) rotates slower , while the flow close to the equatorial plane rotates faster .
in addition , we find that the high - latitude outflow has higher density , lower speed , and a smaller bernoulli parameter . for the effect of the magnitude of @xmath1
, we find that the models with different @xmath1 have similar radial structure but different angular structure .
the value of @xmath1 strongly affects the radial velocity and the value of @xmath167 of outflow at a high latitude .
we have paid more attention in the present work to studying the physics of slim disks .
we have calculated the radial profiles of inflow and outflow rates defined by equations ( 4 ) and ( 5 ) .
we have found that both of them decrease inward .
specifically , the inflow rate can be well described by a power law form , @xmath216 .
the value of @xmath3 is not sensitive to the accretion rate but is mainly dependent on the value of @xmath1 . for @xmath217 , and @xmath9 , @xmath218 , and @xmath159 , respectively ( figure [ fig 8 ] ) .
correspondingly , the radial profile of density becomes flatter compared with the case of a constant @xmath5 .
the density profile can be described by @xmath6 .
it is interesting to note that the value of @xmath7 is within a narrow range , @xmath109 for @xmath219 ( figure [ fig 6 ] ) .
these results are very similar to a hydrodynamical hot accretion flow . in that case ,
yuan et al .
( 2012 ) show that the inward decrease of inflow rate is because of the mass loss in outflow .
we believe this is also the case for the present slim disk . in the case of hot accretion flows ,
the mechanism of producing outflow is identified to be buoyancy associated with the convective instability of the accretion flow .
to investigate the origin of outflow in the slim disk , we first calculate and compare the properties of inflow and outflow .
we have found that the temperature of inflow is lower than that of outflow .
the discrepancy is larger when @xmath1 is larger ( figure [ fig 10 ] ) .
this suggests the existence of convective instability at some level , especially when @xmath1 is large .
we then analyze the convective stability of the accretion flow on the basis of our simulation data .
the result is somewhat subtle . when @xmath8 , about half of the region of the accretion flow is convectively unstable , but when @xmath220 , less than half of the region is unstable ( figure [ fig 11 ] ) .
recall that a non - radiative accretion flow is convectively unstable , and we speculate that radiation can stabilize the convection .
physically this is because radiation can also take away energy , like convection .
the effectiveness of this stabilizing seems to depend on the magnitude of @xmath1 .
when @xmath1 is smaller , it is more effective , i.e. , the accretion flow tends to be more convectively stable ( section [ convection ] ) .
returning to the issue of the radial profile of inflow rate , two questions arise .
the first question is what is the mechanism of producing outflow , especially if the slim disk is roughly convectively stable when @xmath1 is small ?
we speculate that the outflow may be produced by radiation force . or more precisely speaking , both convection and radiation force can produce outflow .
their relative importance may depend on @xmath1 .
when @xmath1 is small , the convective energy flux is weaker , thus radiation force will be the dominant mechanism of producing outflow . the force analysis presented in figure
[ fig 4 ] already suggests the importance of radiation force .
we will study this problem in more detail in a future work .
if this speculation is correct , the mechanism of producing outflow in slim disks and hot accretion flows is quite different .
we note in this context that the angular distribution of the outflow and inflow rates in our simulation are quite different from the case of hot accretion flows ( refer to section 3.4 )
. this may be evidence for the different origin of outflow in slim disk and hot accretion flows . combing the cases of slim disks and the hydrodynamical hot accretion flow , we find that the slope of the radial profile of inflow rate is quite similar , although the mechanisms of producing outflow in the two cases are likely different , as we state above .
in fact , the slope is even similar to the case of magnetohydrodynamical hot accretion flow as well . in that case
, the mechanism of producing outflow is identified to be a lorentz force such as magnetocentrifugal force ( yuan et al .
2012 ; f. yuan et al .
2013 , in preparation ) .
what is the reason for the same slope in spite of different mechanisms ?
begelman ( 2012 ) may provide an answer to the question of why the hydrodynamical and magnetohydrodynamical hot accretion flows have the same radial profile of inflow rate , see also the summary presented in yuan et al .
( 2012 ) . now
the similarity among the three cases seems to indicate that the analysis in begelman ( 2012 ) also applies to radiation - dominated slim disks .
finally , our whole investigation presented in this paper is based on the assumption that the mass accretion rate can be super - eddington . however
, observations of a large sample of agns with 407 sources show that almost all active galactic nuclei ( agns ) are radiating below @xmath11 ( kollmeier et al .
later , steinhardt & elvis ( 2010 ) extended this study to a much larger sample consisting of 62,185 quasars from the sloan digital sky survey , and they got a similar conclusion ( but see kelly & shen ( 2013 ) for a different opinion ) .
liu et al .
2013 most recently address this sub - eddington puzzle .
the basic idea is that because of radiative feedback , the mass accretion rate at the inner accretion flow can be self - regulated and thus can not be super - eddington . unfortunately they find that this mechanism can not fully solve this problem .
abramowicz , m. a. , czerny , b. , lasota , j. p. , & szuszkiewicz , e. 1988 , apj , 332 , 646 balbus , s. a. , & hawley , j. f. 1998,rvmp .
, 70 , 1 balbus , s. a. , & hawley , j. f. 2002 , apj , 573 , 749 begelman , m. c. 2012 , apj , 749 , 3 begelman , m. c. , & meier , d. l. 1982 , apj , 253 , 873 blandford , r. d. , & begelman , m. 1999 , mnras , 303 , 1 blandford , r. d. , & begelman , m. 2004 , mnras , 349 , 68 bu , d. , yuan , f. , wu , m. & cuadra j. 2013 , muras,434,1692 eggum , g. e. , coroniti , f. v. , & katz , j. i. 1987 , apj , 323 , 634 eggum , g. e. , coroniti , f. v. , & katz , j. i. 1988 , apj , 330 , 142 gu , w. m. 2012 , apj , 753 , 118 hawley , j. f. , gammie , c. f. , & balbus , s. a. 1995 , mnras , 272 , 828 hawley , j. f. , gammie , c. f. , & balbus , s. a. 1996 , apj , 464 , 690 hirose , s. , krolik , j. h. , & blaes , o. , 2009 , apj , 691 , 16 igumenshchev , i. v. , & abramowicz , m. a. 1999 , mnras , 303 , 309 igumenshchev , i. v. , & abramowicz , m. a. 2000 , apjs , 130 , 463 jiang , y. f. , stone , j. m. , & davis , s. w. , 2013 , apj,778,65 kato , s. , fukue , j. , & mineshige , s. 1998 , black - hole accretion disks ( kyoto , japan : kyoto univ . press ) kelly , b. , & shen , y. 2013 , apj , 764 , 45 kollmeier , j. a. , onken , c. a. , kochanek , c. s. , et al .
2006 , apj , 648 , 128 kawashima , t. , ohsuga , k. , mineshige , s. , et al .
2012 , apj , 752 , 18 levermore , c. d. , & pomraning , g. c. 1981 , apj , 248 , 321 li , j. , ostriker , j. , & sunyaev , r. 2013 , apj , 767 , 105 liu , c. , yuan , f. , ostriker , j. p. , gan , z. , & yang , x. h. 2013 , mnras 434,1721 mineshige , s. , kawaguchi , t. , & takeuchi , m. 2000 , newar . , 44 , 435 narayan , r. , igumenshchev , i. v. , & abramowicz , m. a. 2000 , apj , 539 , 798 narayan , r. , quataert , e. , igumenshchev , i. v. , & abramowicz , m. a. , 2002 , apj , 577 , 295 narayan , r. , sadowaki , a. , penna , r. , & kulkarnl , a. k. 2012 , mnras , 426 , 3241 narayan , r. , & yi , i. 1994 , apj , 428 , 13 ohsuga , k. , & mineshige , s. 2007 , apj , 670 , 1283 ohsuga , k. , & mineshige , s. , mori , m. , umemura , m. 2002 , apj , 574 , 315 ohsuga , k. , mineshige , s. , & watarai , k. 2003 , apj , 596 , 429 ohsuga , k. , mori , m. , nakamoto , t. , & mineshige , s. 2005 , apj , 628 , 368 okuda , t. 2002 , pasj , 52 , 253 okuda , t. , teresi , v. , toscano , e. , & molteni , d. 2005 , mnras , 357 , 295 paczyski , b. , & wiita , p. j. 1980 , a&a , 88 , 23 pringle , j. e. 1981 , ara&a , 19 , 137 quataert , e. , & gruzinov , a. 2000 , apj , 539 , 809 sadowski , a. , abramowicz , m. , bursa , m. , et al .
2009 , a&a , 502 , 7 sadowski , a. , abramowicz , m. , bursa , m. , et al .
2011 , a&a , 527 , a17 sadowski , a. , narayan , r. , penna , r. , & zhu , y. 2013 , arxiv:1307.1143 shakura , n. i. , & sunyaev , r. a. 1973 , a&a , 24 , 337 steinhardt , c. l. , & elvis , m. 2010 , mnras , 402 , 263 stone , j. m. , hawley , j. f. , gammie , c. f. , & balbus , s. a. 1990apj , 463 , 656 stone , j , m. , & pringle , j. e. 2001 , mnras , 322 , 461 stone , j. m. , pringle , j. e. , & begelman , m. c. 1999 , apj , 310 , 1002 tassoul , j. l. , 1978 , theory of rotating stars ( princeton , nj : princeton univ . press ) vierdayanti , k. , mineshige , s. , ebisawa , k. , & kawaguchi , t. 2006 , pasj , 58 , 915 wang , j. m. , & zhou , y. y. 1999 , apj , 516 , 420 watarai , k. , fukue , j. , takeuchi , m. & mineshige , s. 2000 , pasj , 52 , 133 watarai , k. , mizuno , t. , & mineshige , s. 2001 , apjl , 549 , l77 yuan , f. 2001 , mnras , 324 , 119 yuan , f. , & bu , d. 2010 , mnras , 408 , 1051 yuan , f. , bu , d. , & wu , m. 2012 , apj , 761 , 130 yuan , f. , wu , m. , & bu , d. 2012 , apj , 761 , 129 | we study the dynamics of super - eddington accretion flows by performing two - dimensional radiation - hydrodynamic simulations . compared with previous works , in this paper
we include the @xmath0 component of the viscous stress and consider various values of the viscous parameter @xmath1 .
we find that when @xmath0 is included , the rotational speed of the high - latitude flow decreases , while the density increases and decreases at the high and low latitudes , respectively .
we calculate the radial profiles of inflow and outflow rates .
we find that the inflow rate decreases inward , following a power law form of @xmath2 .
the value of @xmath3 depends on the magnitude of @xmath1 and is within the range of @xmath4 .
correspondingly , the radial profile of density becomes flatter compared with the case of a constant @xmath5 .
we find that the density profile can be described by @xmath6 and the value of @xmath7 is almost same for a wide range of @xmath1 ranging from @xmath8 to @xmath9 .
the inward decrease of inflow accretion rate is very similar to hot accretion flows , which is attributed to the mass loss in outflows .
to study the origin of outflow , we analyze the convective stability of the slim disk .
we find that depending on the value of @xmath1 , the flow is marginally stable ( when @xmath1 is small ) or unstable ( when @xmath1 is large ) .
this is different from the case of hydrodynamical hot accretion flow where radiation is dynamically unimportant and the flow is always convectively unstable .
we speculate that the reason for the difference is because radiation can stabilize convection .
the origin of outflow is thus likely because of the joint function of convection and radiation , but further investigation is required . |
the main goal of experiment e246 at the kek 12 gev proton synchrotron is a search for t violating transverse muon polarization ( @xmath1 ) in the decay @xmath0 ( @xmath2 ) with a sensitivity of about @xmath3 .
the standard model predictions for @xmath1 is less than @xmath4 @xcite and a contribution from the final state interactions was found to be below the @xmath5 level @xcite .
so , this experiment is able to search for additional or alternative sources of cp - violation postulated in various extensions to the standard model . in particular
it is sensitive to the cp - violation in the multi higgs doublet models , the leptoquark models and supersymmetric models with / without r parity violation , where @xmath1 can reach the measurable level of @xmath6 @xcite .
the e246 detector was commissioned in 1996 and accumulated data for about 6 years .
first results with a new limit on @xmath1 were published in @xcite . in this experiment ,
the @xmath2 decay is identified by detecting the @xmath7 as well as the @xmath8 from the decay .
the layout of the e246 set up is shown in fig . [
fig : setup ] .
a 660 mev / c kaon beam is slowed down in a beo degrader and stopped in an active target , made of 256 @xmath9 mm@xmath10 scintillating fibers . the energy and direction of the @xmath11 from the @xmath2 decay
are measured by a segmented csi(tl ) photon detector installed in the central region of the toroidal magnet .
a muon from the @xmath2 decay at rest is momentum analyzed by one of the 12 magnet gaps of the superconducting toroidal spectrometer using tracking performed by the stopping target , a scintillating ring hodoscope @xcite surrounding the target , and three mwpcs .
the muon exiting the spectrometer is stopped in a polarimeter in which the decay positron asymmetry @xmath12 is measured in order to obtain @xmath1 .
the polarimeter consists of 12 azimuthally arranged al stoppers , aligned with the magnet gaps , with a scintillator counter system located between the stoppers .
a positron from the decay of @xmath13 tends to be emitted in the same direction as the muon spin , and is detected by these plastic scintillator counters . in this detector @xmath12 appears as a difference in the counting rate between clockwise ( @xmath14 ) and counter - clockwise ( @xmath15 ) emitted positrons . by summing the @xmath14 ( @xmath16 ) and @xmath15 ( @xmath17 ) positron counts over
all 12 sectors , @xmath1 is derived from @xmath18 where @xmath19 is the analyzing power of the polarimeter and @xmath20 is an attenuation factor which reflects the fact that the decay plane of the @xmath2 event determined from the @xmath8 and @xmath7 momenta is not parallel to the median plane of the gap in which the muon momentum is measured .
the photon detector plays a crucial role in this experiment for suppression of systematic errors . using complete reconstruction of the @xmath2 kinematics we separate these events into two classes : events with the pion moving along the kaon beam ( forward direction ) and those where the pion moves in the backward direction .
the sign of p@xmath21 for the forward - going @xmath7 events is opposite to that for backward - going @xmath7 events .
this allows us to apply a double ratio between the forward- and backward - going @xmath7 events cancelling out most instrumental sources of false polarization , which are likely to be independent of the @xmath7 direction .
this method is an unique feature of e246 experiment , and is a key factor for effective suppression of most systematic errors .
the photon detector consists of 768 csi(tl ) crystals each with a pin photodiode readout .
they form a barrel structure that covers about 75% of @xmath22 with 12 symmetrically spaced holes for @xmath23s going into the magnetic spectrometer ( fig . [
fig : barrel ] ) .
the detector is capable detecting photons emerging from the target in the polar angle range from 15@xmath24 to 165@xmath24 and in 2@xmath25 of the azimuthal angle except for the muon holes .
the optically isolated csi modules point towards the center of the active target .
an individual crystal covers 7.5@xmath24 in both polar and azimuthal angles , except for 48 crystals near to the beam axis , where the azimuthal angle was doubled to 15@xmath24 , and the muon hole region .
the crystal length of 250 mm ( 13.5 r.l . ) was chosen in order to leave about 4 cm of space available radially to mount a pin - diode and preamplifier .
the overall gain stability of the calorimeter is monitored by a xe lamp light pulser system with optical fibers which distribute the light from the lamp to the crystals . to obtain uniform high light yield along the crystal
, the crystal surfaces were treated to grade the reflectivity , and a millipore white paper was chosen as a reflector ( see for details ref .
@xcite ) . as a result , for almost all modules the light yield of more than 8000 p.e./mev and the average equivalent noise level ( enl ) of about 63 kev were obtained , as shown in fig .
[ fig : enl ] . mm@xmath10 sensitive area .
( b ) distribution of the equivalent noise level .
, width=566 ] the whole detector demonstrated a low level of coherent noise of about 11 kev ( @xmath26 ) per module , which has only a small effect on the energy resolution and signal - to - noise ratio of the detector .
these parameters are the best among the large csi(tl ) calorimeters which have been built in recent time @xcite . to efficiently use the high light yield and low enl of individual crystals , the electronics for the calorimeter was specially designed .
the high counting rate environment in the calorimeter due to pion contamination in the beam was also taken into account in the design .
the calorimeter electronics is schematically shown in fig .
[ fig : electr ] and described in detail in ref .
the main amplifier incorporates a shaping amplifier with fast restoration of the base level and a timing filter amplifier which forms a timing signal for a subsequent discrimination by a constant fraction discriminator ( cfd ) .
the csi timing signals from all cfd outputs are used in the second level trigger in coincidence with the first level fast trigger .
the detector stability is monitored by a xe lamp system . during the nearly 6 year period of operation none of the 768 photodiodes and only two preamplifiers failed .
a light yield degradation of the csi modules of about 12% was observed over this time .
the in - beam calibration of the detector was carried out using muons from the dominant decay mode @xmath27 of stopped kaons .
these muons have a kinetic energy of 152.4 mev and are stopped in the csi crystals .
the calibration runs , which were carried out every beam cycle , used a reduced beam intensity and a special trigger .
it included the kaon identification with the cherenkov detector , kaon stop in the target , and the presence of a signal in the csi . in
flight kaon decays were avoided by introducing a 15ns time delay .
the events in which no energy was detected in the neighboring crystals were selected to obtain the individual gains of csi modules .
the kaon decay vertex was reconstructed from @xmath8 tracking and muon energy losses in the target were corrected using the target fiber adc information @xcite .
using individual csi calibration coefficients the decay @xmath28 is reconstructed .
[ fig : kpi2 ] photon clusters from the @xmath29 decay .
( a ) all events with peak value @xmath30 mev , @xmath31 .
the low energy tail is due to the shower leakage into muon and beam holes .
( b ) photons are detected away from muon holes .
mev , @xmath33 .
, width=529 ] shows the spectrum of the energy sum of two photon clusters from the decay @xmath34 .
the peak value 235.7 mev corresponds to about 96% of the @xmath35=245.6 mev , that is consistent with the result of monte carlo calculation for the cluster size of 5@xmath365 crystals .
the low energy tail is due to the shower leakage into the muon and beam holes , and through the rear sides of the crystals .
the resolution @xmath37 obtained for the energy sum of two photons from @xmath29 decay is 5.2% .
the energy resolution is better for clusters away from muon holes , as seen from fig .
[ fig : kpi2](b ) .
the invariant mass resolution of the @xmath7 was 6.7% for whole detector and 5.6% for events with both photons detected away from muon holes .
good time resolution is an important factor for reliable identification of photons and neutral pions from the decays of @xmath38 s and for background suppression .
the time resolution was determined in physics runs by measuring time difference between a muon signal from @xmath2 decay produced by one of the 12 fiducial plastic counters surrounding the target , and a cfd signal from a csi module . fig .
[ fig : time ] shows energy dependence of the obtained time resolution ( @xmath26 ) of the csi crystals which are the centers of the gamma clusters .
time resolution of 3.5 ns at 100 mev deteriorates to 16.8 ns in the low energy range of 1020 mev . for the @xmath2 photon energy range of 10 - 250 mev
the resolution @xmath39 nsec was obtained . to exploit the timing performance ,
all the csi cfd outputs were added to the trigger logic as the or strobe .
a 150 ns discriminator pulse width was sufficient to accept nearly all good events , while reducing the recorded trigger rate by a factor of 2 .
the slow decay time of csi(tl ) crystals can lead to losses of detected events in a high intensity beam due to pile - up of the csi pulses .
the large contamination of pions in the beam ( @xmath40 ) and an asymmeterical pion halo were the main sources which dominated the counting rate in the calorimeter .
the counting rate of a crystal depends on its position relative to the beam axis and was dramatically increased for the crystals closest to the beam , as seen in fig .
[ fig : rate ] . .
, width=453 ] taking into account the distribution of the events over the solid angle , the average counting rate per ctystal in the experiment was determined to be about 8 khz .
the tests of baseline stabilty at this counting rate using xe lamp events showed that about 5% of events were lost in physics runs due to pile up .
the limited accuracy of the detector installation can make an asymmetry in the angular distribution between the median plane of the polarimeter and @xmath2 decay plane , the normal vector @xmath41 to which is defined as @xmath42 here @xmath43 and @xmath44 are the momenta of the pion and muon , respectively .
this angular asymmetry can give rise to contamination of the in - plane polarization and can generate a spurious t - odd effect due to the opposite direction of this polarization for the forward- and backward - going pions . according to the monte carlo simulation ,
the residual value of the muon in - plane polarization in the polarimeter is 0.50.7 for both directions of the neutral pions .
the measured angular distributions of @xmath2 decay plane rotations around the beam direction ( z - axis ) and radial direction ( r - axis , perpendicular to the beam and lying in the median plane of a magnet sector , see fig . [
fig : setup](c ) ) are presented in fig .
[ fig : decay_rotation ] . decay plane rotation around r and z axes .
( a ) and ( c ) @xmath2 events with forward - going pions ; ( b ) and ( d ) @xmath2 events with backward - going pions.,width=453 ] as one can see , the detector provides good angular symmetry , and possible spurious polarization caused by detector misalignment is @xmath45 , i.e. much less than the expected statistical error . to measure the power of the double ratio , the compensation of a large imitation asymmetry in the experimental kaon stopping distribution was studied for the good forward and backward @xmath2 events .
the events accumulated in 1998 running cycle were selected , in the right - hand part of the target relative to the median plane in each magnet sector looking upstream the beam , as shown in fig .
[ fig : half_target ] . as a result , in each gap a spurious asymmetry of 26% appears , as seen in fig .
[ fig : double ] , events with forward ( backward)-going pions selected as shown in fig .
[ fig : half_target ] .
the values shown for gap no.13 are sums of forward and backward asymmetries over all polarimeter sectors . , width=415 ] and total asymmetry is @xmath46 and @xmath47 for forward and backward events , respectively .
so , the double ratio @xmath48 provides a large factor for cancellation of systematic errors in the asymmetry measurement . for different data sets this factor was determined to be in the range of 15 - 24
the csi(tl ) photon detector with pin photodiode readout is being successfully used in the e246 experiment at kek .
the overall performance of the calorimeter was not degraded significantly after being in operation for 6 years .
the csi calorimeter provides good reconstruction of the @xmath7 parameters and seems to be a key instrument for reduction of systematic uncertanties in the polarization measurement down to the level of @xmath49 that completely fulfills the requirements of the experiment .
bigi and a.i .
sanda , cp violation , cambridge university press , 2000 .
e. golowich and g. valencia , phys .
, * d40 * ( 1989 ) 112 .
zhitnitskii , yad . fiz .
* 31 * ( 1980 ) 1024 .
+ v.p.efrosinin et .
al . , phys . lett . *
b493 * ( 2000 ) 293 , hep - ph/008199 .
s. weinberg , phys .
. lett . * 37 * ( 1976 ) 657 .
+ g. blanger and c.q .
geng , phys . rev . * d44 * ( 1991 ) 2789 .
+ g .- h . wu and j.n .
b392 * ( 1997 ) 93 , hep - ph/9609314 .
+ m. fabbrichesi and f. vissani , phys
* d55 * ( 1997 ) 5334 , hep - ph/9611237 .
ivashkin et al . , nucl .
a394 * ( 1997 ) 321 .
m. abe et al . , phys .
* 83 * ( 1999 ) 4253 .
grigorev et .
al . , inst .
exp . tech . * 39 * ( 1996 ) 164 [ prib
n2 * ( 1996 ) 18 ] .
dementyev et al . , nucl .
* a440 * ( 2000 ) 151 .
y. kubota et al . , nucl .
a320 * ( 1992 ) 66 .
+ e. aker et al . , nucl .
* a321 * ( 1992 ) 69 .
+ b. shwartz , `` crystal calorimeters '' , talk at this conference , http://www.inp.nsk.su/events/confs/instr2002/day5.shtml ; k. miyabayashi , `` belle electromagnetic calorimeter '' , ibid ; b.lewandowski , `` design and performance of the babar em calorimeter '' , ibid .
kudenko , o.v .
mineev , j. imazato , nucl .
a411 * ( 1998 ) 437 .
khabibullin et al . , instr .
exp . tech .
* 43 * ( 2000 )
589 [ prib . tekhn
n5 * ( 2000 ) 9 ] . | the performance of the photon detector constructed for the search of t - violation in the decay @xmath0 is presented .
the specific features of this detector consisting of 768 csi(tl ) crystals with pin photodiode readout for high precision measurement of t - odd correlations in decays of positive kaons are considered . |
the formation , resonant or phonon - assisted , of qw excitons and their following relaxation towards a final ( quasi- ) equilibrium thermodynamic state at low lattice temperature @xmath2 are the subject of numerous experimental @xcite and theoretical @xcite studies .
recently , relaxation thermodynamics has been formulated and developed in order to analyze how bose - einstein statistics of high - density qw excitons influences the phonon - assisted thermalization processes .
@xcite the above thermodynamics assumes a strong , dominant exciton - exciton scattering and , therefore , relaxation through quasi - equilibrium thermodynamic states .
the relaxation thermodynamics has successfully been applied to model the @xmath8-dependent thermalization and photoluminescence kinetics observed in early experiments @xcite with high - density excitons ( @xmath9 @xmath10 @xmath11 @xmath10 ) in gaas qws .
@xcite the long - lived indirect excitons in high - quality gaas / algaas coupled qws provide a unique opportunity for studying quantum degeneracy in a system of two - dimensional bosons . in this case , the long radiative lifetimes of indirect excitons allows the system to cool down to temperatures where the dilute exciton gas becomes statistically degenerate .
@xcite the quality of present day gaas / algaas coupled qws has been considerably improved in comparison with those used in the pioneering experiments @xcite one decade ago .
furthermore , the very recent magneto - optical experiments @xcite clearly indicate that the in - plane momentum @xmath12 of indirect excitons is a well - defined quantum number in high - quality gaas / algaas coupled qws .
thermalization of hot photoexcited excitons down to the temperature of the cold lattice occurs mainly via scattering by thermal bulk longitudinal acoustic ( la ) phonons and is much more efficient for quasi - two - dimensional ( quasi-2d ) systems as compared to bulk semiconductors .
this follows from the relaxation of momentum conservation in the @xmath13-direction ( the qw growth direction ) for quasi-2d systems : the ground - state mode @xmath14 , i.e. , the energy state @xmath15 , couples to the continuum energy states @xmath16 , rather than to the single energy state @xmath17 ( @xmath18 is the longitudinal sound velocity and @xmath19 is the in - plane translational mass of excitons ) as occurs in bulk materials . as a result , the la - phonon assisted kinetics of qw excitons becomes dominant at @xmath20 @xmath10 : in this case exciton - exciton scattering can be neglected while bose - einstein ( be ) statistics already strongly influences the relaxation process at low temperatures .
@xcite crossover from classical to quantum statistics occurs near the degeneracy temperature @xmath21 , where @xmath22 is the spin degeneracy factor . for @xmath23 @xmath10 the degeneracy temperature of indirect excitons in gaas / algaas
coupled qws is @xmath24 k. this estimate refers to @xmath25 , which can be achieved in the @xmath26-type gaas / algaas coupled qws by applying a static magnetic field @xmath27 .
@xcite note that the very recent experiments @xcite deal with gaas / algaas coupled qws at extremely low cryostat temperature @xmath28 k. in this paper we study analytically and model numerically the acoustic - phonon - assisted relaxation kinetics of statistically - degenerate qw excitons at low densities .
the recent experiments @xcite allow us to visualize , by means of lo - phonon - assisted emission , the la - phonon - assisted kinetics of quasi-2d excitons in znse qws and , in particular , to prove that for @xmath29 @xmath10 the above kinetics indeed occurs through non - equilibrium distributions of qw excitons . in experiments
@xcite the concentration of be - degenerate indirect excitons in gaas / algaas coupled qws usually varies from @xmath30 @xmath10 at the very end of an optical excitation pulse to @xmath31 @xmath10 at large delay times @xmath32 ns after the optical excitation .
thus , thermalization of the indirect excitons at large delay times can not be described within the relaxation thermodynamics @xcite and does need a separate theoretical analysis .
the classical boltzmann kinetic equation has been generalized in order to include quantum statistics by uehling and uhlenbeck .
@xcite the relevant quantum kinetic equation for spatially - homogeneous dilute system of statistically - degenerate quasi-2d excitons coupled to bulk la - phonons is @xmath33 \delta(e_{\bf k_{\| } } - e_{{\bf k_{\|}}-{\bf q}_{\| } } - \hbar q v_s ) \nonumber\\ & + & \big [ n_{\bf k_{\| } } n^{\rm ph}_{\bf q } ( 1+n_{{\bf k_{\|}}+{\bf q}_{\| } } ) -(1+n_{\bf k_{\| } } ) ( 1+n^{\rm ph}_{\bf q } ) n_{{\bf k_{\|}}+{\bf q}_{\| } } \big ] \delta(e_{\bf k_{\| } } - e_{{\bf k_{\|}}+{\bf q}_{\| } } + \hbar q v_s ) \big\ } , \label{1.1}\end{aligned}\ ] ] where @xmath34 and @xmath35 are the occupation numbers of exciton in - plane mode @xmath36 and phonon bulk mode @xmath37 , respectively , and @xmath38 is the in - plane projection of @xmath39 .
the terms in the first and second square brackets on the right - hand - side ( r.h.s . ) of eq .
( [ 1.1 ] ) describe the stokes and anti - stokes la - phonon - assisted scattering processes , respectively .
the acoustical phonons are assumed to be in a thermal equilibrium at the bath temperature @xmath2 .
@xcite the matrix element is given by @xmath40^{1/2 } f_z(q_z l_z/2)$ ] , where @xmath41 is the crystal mass density , @xmath42 is the deformation potential of exciton - la - phonon interaction , @xmath43 is the thickness of a qw , and @xmath44 is the volume .
the form - factor @xmath45 [ e^{i \chi } / ( 1 - \chi^2/\pi^2)]$ ] refers to an infinite rectangular confinement potential .
@xcite the latter function describes the relaxation of the momentum conservation law in the @xmath13-direction and characterizes a spectral band of bulk la - phonons , which effectively interact with qw excitons .
note that eq .
( [ 1.1 ] ) is valid only for the kinetic stage of thermalization , i.e. , before a low - temperature collective state of excitons @xcite builds up .
the main aim of our work is to study the fundamental features of the acoustic phonon - assisted thermalization kinetics of qw excitons from initial strongly nonequilibrium @xmath46 towards the final equilibrium distribution with well - developed bose - einstein statistics , when @xmath47 .
our numerical simulations of the la - phonon - assisted kinetics clearly demonstrate that after the first transient , which lasts a few characteristic scattering times , a slow _ adiabatic stage _ of thermalization builds up ( see fig .
[ fig1 ] ) .
this stage is characterized by a quasi - equilibrium distribution of high - energy qw excitons with effective temperature @xmath48 and is independent of the initial distribution at @xmath49 .
the adiabatic stage lasts many characteristic scattering times and arises due to the need to populate the low - energy in - plane modes with @xmath50 in the presence of effective suppression of the stimulated scattering processes ( e.g. , an intense incoming , stokes flux of excitons into the ground - state mode @xmath51 is nearly compensated by the outgoing , anti - stokes scattering out of the state @xmath36=0 ) . in order to describe the adiabatic stage of relaxation
, we find a _ generic solution _ of the quantum kinetic eq .
( [ 1.1 ] ) . while the generic solution we calculate is different from that derived in ref . for the phonon - assisted kinetics of bulk excitons at @xmath52 ( @xmath53 is the critical temperature for bose - einstein condensation of excitons in bulk semiconductors ) , similar to this case it depends only on two control parameters of the system , @xmath54 and @xmath2 .
furthermore , the only gross information from a particular shape of the initial distribution at @xmath49 is absorbed by the start time @xmath55 of the adiabatic stage of thermalization .
this is shown in the fig .
[ fig1](a ) , where we vary the parameters of the initial gaussian distribution . in turn , the first transient depends upon initial distribution and lasts only a few scattering times @xmath56 , as illustrated in fig .
[ fig1](b ) . , where @xmath57 , and parameter @xmath58 ( dash - dotted lines ) , @xmath59 ( solid lines ) , and @xmath60 ( dashed lines ) .
the gross dependence of evolution upon initial conditions is absorbed by time @xmath61 .
( b ) the first stage of la - phonon - assisted relaxation ( @xmath62 ) for a particular distribution of indirect excitons with @xmath63 . in both plots
@xmath64 and @xmath65 . the dimensionless values , time in @xmath66 and energy / temperature in @xmath67 , can easily be rescaled to the dimensional units by using @xmath68 @xmath69ev and @xmath70 ns for gaas / algaas coupled qws , and @xmath71 @xmath69ev and @xmath72 ns relevant to znse qws . ]
because at any nonzero bath temperature @xmath73 the occupation number of the ground - state mode @xmath74 is finite , the adiabatic stage ends up with linearized kinetics .
the latter kinetics refers to the last , exponential stage of thermalization at @xmath75 and is characterized by @xmath76 . for
statistically - degenerate qw excitons the thermalization time @xmath77 reaches its smallest values at bath temperatures @xmath78 .
while the above inequality does not usually hold in the experiments with statistically - degenerate indirect excitons in gaas / algaas coupled qws , @xcite we describe a special design of gaas - based microcavities ( mcs ) for optimization of the la - phonon - assisted thermalization kinetics of low - density mc polaritons . in the proposed microcavities with a large positive detuning between the cavity and qw exciton modes , the mc polaritons have radiative lifetimes on a 100 ps - 1 ns time scale , so that the phonon - assisted relaxation towards well - developed bose - einstein statistics with large occupation numbers can optically be visualized .
thus , the mc design we discuss is an interesting alternative to the semiconductor microcavities with zero detuning between the cavity mode and qw excitons , where huge nonclassical occupation numbers of the low - energy mc polariton states have recently been observed .
@xcite in numerical evaluations we use @xmath79 , @xmath80 cm / s , and @xmath81 ev , relevant to gaas / algaas coupled qws , and @xmath82 , @xmath83 cm / s , and @xmath84 ev , relevant to single znse qws , respectively ( @xmath85 is the free electron mass ) .
note that the disorder - induced scattering and localization processes , not included in our model , are relatively strong in up - to - date znse - based qws and require a separate analysis .
@xcite however , in the very recent work @xcite the first fabrication of high - quality mgs / znse / mgs qws with less than @xmath86 ml fluctuations of the well width and , therefore , with extremely low inhomogeneous broadening has been reported . in sec .
[ sec2 ] , the boltzmann equation is adapted in order to formulate the acoustic - phonon - assisted kinetics of statistically - degenerate qw excitons .
we also discuss some approximations for the form - factor @xmath87 , which describes the relaxation of momentum conservation in scattering of qw excitons by bulk la - phonons . in sec .
[ sec3 ] , we find the generic solution of the acoustic - phonon - assisted kinetics from a strongly nonequilibrium initial distribution of qw excitons @xmath88 towards a final bose - einstein distribution @xmath89 with large occupation numbers of the low - energy states , @xmath90 , where @xmath91 is the dimensionless energy defined by @xmath57 .
the generic solution is independent of a particular shape of the initial distribution and describes the adiabatic stage of thermalization , which starts at @xmath92 in a few characteristic scattering times after @xmath49 .
we show that at @xmath93 thermalization of high - energy qw excitons is given by @xmath94 , where @xmath6 is the lowest positive eigenvalue of the relevant linearized kinetics . in sec .
[ sec4 ] , the linearized phonon - assisted kinetics of statistically - degenerate qw excitons is formulated and analyzed .
we show that the eigenvalues @xmath95 of the linear collision integral form a continuous spectrum , @xmath96 , separated from the nondegenerate eigenvalue @xmath97 , and that the corresponding eigenfunctions @xmath98 have three well - defined isolated critical points .
the dependence of the marginal eigenvalue @xmath6 on the bath ( @xmath2 ) and degeneracy ( @xmath54 ) temperatures is studied . in sec .
[ sec5 ] , straightforward numerical simulations of the phonon - assisted relaxation kinetics at @xmath99 are compared with the generic solution of the quantum boltzmann equation .
we also show that at the beginning of the adiabatic stage of thermalization , at @xmath100 , the population dynamics of the ground - state mode is given by @xmath101 , where the parameters @xmath102 and @xmath103 are calculated analytically .
furthermore , we propose a particular design of gaas - based microcavities ( mcs ) , which compromises the efficiency of la - phonon - assisted scattering ( the density of states is @xmath104 ) with the degeneracy temperature @xmath105 , i.e. , allows us to avoid the bottleneck effect in relaxation and , therefore , to optimize the thermalization kinetics of low - density qw excitons . in the appendix ,
some relationships relevant to the thermalization dynamics of quasi - equilibrated high - energy qw excitons are given .
for hot qw excitons , which are in - plane isotropically distributed at @xmath49 , the thermalization kinetics due to bulk la - phonons can be treated in one - dimensional energy space [ see eq .
( 2 ) of ref . ] . in the following we express energy @xmath106 and temperature @xmath107 ( as well as @xmath2 and @xmath54 ) in terms of @xmath108 ,
i.e. , we use the dimensionless values of @xmath109 , and @xmath110 . in order to derive and analyze the generic solution for relaxation at @xmath111 , it is convenient to rewrite the above equation in terms of the variable @xmath112 where @xmath113 and @xmath114 $ ] are the current and ( final ) equilibrium distribution functions of qw excitons , respectively , and the chemical potential @xmath69 is given by @xmath115 . in this case
the kinetic equation reduces to @xmath116 , \label{2.2}\ ] ] where the stokes ( s ) and anti - stokes ( as ) collision integrands are @xmath117 \left ( 1+n^{\rm ph}_{\varepsilon_{1}}+n^{0}_{\varepsilon - \varepsilon_{1 } } \right ) \nonumber \\ & & + \ t_b ( n^{0}_{\varepsilon - \varepsilon_{1 } } ) ' f_{\varepsilon}(t ) f_{\varepsilon - \varepsilon_{1}}(t ) , \label{2.3a } \\ { \mathcal l}_{\rm as}(\varepsilon , \varepsilon_{1},t ) & = & \left [ f_{\varepsilon}(t ) - f_{\varepsilon + \varepsilon_{1 } } ( t ) \right ] \left ( n^{\rm ph}_{\varepsilon_{1 } } - n^{0}_{\varepsilon + \varepsilon_{1 } } \right ) \nonumber \\ & & - \ t_b ( n^{0}_{\varepsilon + \varepsilon_{1 } } ) ' f_{\varepsilon}(t ) f_{\varepsilon + \varepsilon_{1}}(t ) .
\label{2.3b}\end{aligned}\ ] ] [ 2.3 ] the distribution of thermal bulk phonons is given by the planck formula , i.e. , @xmath118 .
the scattering time is defined by @xmath119 .
the functions @xmath120 , which determine the integration limits on the right - hand - side of eq .
( [ 2.2 ] ) , are @xmath121 [ 2.4 ] the functions @xmath122 in eq .
( [ 2.2 ] ) are given by @xmath123 \left[\alpha^2 - \left(d^{\rm s / as}_{\varepsilon , \varepsilon_{1}}\right)^2 \right ] \right\}^{-1/2 } d \alpha , \label{2.5}\end{aligned}\ ] ] where @xmath124^{1/2}$ ] and @xmath125^{1/2}$ ] .
the dimensionless parameter @xmath126 in the argument of the form - factor function @xmath127 is given by @xmath128 . in order to estimate
the contribution of the form - factor @xmath87 to the spectral functions @xmath129 one can analyze eq .
( [ 2.5 ] ) for @xmath130 . in this case
only the spectral function @xmath131 is relevant to the kinetic eq .
( [ 2.2 ] ) , and eq . ( [ 2.5 ] ) yields @xmath132 \right|}^2 . \label{2.6}\ ] ] for @xmath133 , the spectral width @xmath134 ( fwhm ) of @xmath135 is given by @xmath136 and determined solely by @xmath137 .
for thickness @xmath138 nm of gaas qws ( see ref . )
one gets @xmath139 and , therefore , @xmath140 .
the latter estimate can be rewritten in the dimensional energy units as @xmath141 mev .
thus the spectral band of la - phonons , which scatter a low - energy qw exciton , can be evaluated as @xmath142 , where @xmath143 is the uncertainty of the momentum in the @xmath13-direction due to the qw spatial confinement .
the above estimates show that for relatively cold qw excitons with energies @xmath144 ( the effective temperature @xmath145 k for @xmath146 nm ) the form - factor can be approximated by @xmath147 . in this case
the integral on the r.h.s . of eq .
( [ 2.5 ] ) can be calculated explicitly : @xmath148 , \label{2.7}\ ] ] where @xmath149 - k\left ( \frac{1}{\xi } \right ) \right\ } , & 0 \le \xi < 1 \\ k\left ( \frac{1}{\xi } \right ) , & \xi < 0 .
\\ \end{array } \right .
\label{2.8}\ ] ] here , @xmath150^{-1/2 } d \theta$ ] and @xmath151 are the elliptic and the complete elliptic integrals of the first kind , respectively .
in this section the thermalization kinetics at @xmath152 is described in terms of a generic solution , which weakly correlates with the initial distribution @xmath88 . in particular , we derive a thermalization law for high - energy excitons and examine nonequilibrium distribution of low - energy qw excitons .
we also specify a reference point for the generic solution , i.e. , a set of parameters , which unambiguously determines the calculated evolution .
schematic picture of the relaxation kinetics in phase space is shown in fig .
the generic solution is relevant to the times @xmath153 , where @xmath55 is the start time of the adiabatic stage of evolution . .
the reference ( start ) point @xmath154 unambiguously determines the calculated ( solid line ) evolution from the initial noequilibrium distribution @xmath88 .
the dashed line corresponds to the generic solution .
@xmath55 is the start point of adiabatic stage of evolution . for the times
@xmath155 both evolution spirals almost coincide showing a unique path towards the final equilibrium distribution @xmath156 . ] as was emphasized in sec .
[ introduction ] , the generic solution assumes a quasi - equilibrium distribution of _ high - energy _ excitons ( @xmath157 ) : @xmath158 $ ] .
this quasi - equilibrium distribution is characterized by the effective time - dependent temperature @xmath48 and chemical potential @xmath159 whose time variations are supposed to be small , so that @xmath159 can be taken the same as the chemical potential of final equilibrium distribution @xmath160 .
indeed , for @xmath161 the chemical potential is given by @xmath162 and , therefore , @xmath163 .
since at the final stage of relaxation kinetics the effective temperature of high - energy excitons approaches the bath temperature ( @xmath164 ) , starting from some moment in time one meets the condition @xmath165 ( provided that @xmath166 ) .
therefore we can linearize the quasi - equilibrium distribution function with respect to @xmath167 . in this case one
obtains @xmath168\varepsilon .
\label{3.1.1}\ ] ] on the other hand , as we show below , a _ low - energy
_ kernel ( @xmath169 ) of the distribution function characterizes the non - equilibrium qw excitons . with increasing time @xmath170 ,
the low - energy kernel shrinks in energy space , and the ratio @xmath171 decreases approaching the limit when the linearized kinetics becomes valid . in order to analyze the evolution of low - energy qw excitons within the scenario described in the previous subsection
, we substitute eq .
( [ 3.1.1 ] ) into eq .
( [ 2.2 ] ) and get the reduced kinetic equation for @xmath172 : @xmath173 f_{\varepsilon}(t ) + \eta(\varepsilon ) \delta t(t ) , \label{3.2.1}\ ] ] where @xmath174 @xmath175 @xmath176 and @xmath177 is given by eq .
( [ 2.7 ] ) .
equation ( [ 3.2.1 ] ) , which describes the population dynamics of the low - energy states during the adiabatic stage of thermalization ( @xmath178 ) , is a linear inhomogeneous differential equation for @xmath179 .
its complete solution can be written as a sum of the homogeneous and inhomogeneous contributions : @xmath180 where @xmath181 @xmath182 and @xmath183 is an arbitrary reference ( start ) time for the calculated evolution , i.e. , @xmath184 ( see fig . [ fig2 ] ) .
thus the adiabatic stage of the relaxation kinetics into the lower - energy states is completely determined by @xmath185 , provided that one knows the reference ( start ) distribution @xmath186 and thermalization law @xmath187 for high - energy qw excitons .
note that in a sharp contrast with the acoustic - phonon - assisted relaxation kinetics at @xmath52 in three - dimensional systems , @xcite the thermalization dynamics of low - energy qw excitons depends upon the homogeneous contribution @xmath188 given by eq .
( [ 3.2.6 ] ) . as will be shown in the next subsection , the thermalization law for quasi - equilibrium high - energy qw excitons ( @xmath189 ) is given by @xmath190 where @xmath191 characterizes the inverse thermalization time at @xmath192 , @xmath193 is another characteristic parameter relevant to the beginning of the adiabatic stage ( @xmath194 ) , and @xmath195 determines reference ( start ) effective temperature @xmath196 for the calculated evolution .
equation ( [ 3.2.8 ] ) is valid for @xmath197 . using the thermalization law ( [ 3.2.8 ] )
we find the integral on the r.h.s .
( [ 3.2.7 ] ) : @xmath198 - ei[-\lambda_0 ( t_1 - t_0 ) ] \right .
\nonumber\\ & & \left .
-ei[-\lambda_1 ( t - t_0 ) ] + ei[-\lambda_1 ( t_1-t_0 ) ] \right\ } , \label{3.2.9}\end{aligned}\ ] ] where @xmath199 is the exponential integral function .
the reference distribution @xmath200 , which determines the homogeneous solution ( [ 3.2.6 ] ) , is independent of the initial distribution of hot qw excitons , @xmath88 .
there is , however , an integral relationship between @xmath201 and @xmath202 , which we will discuss in subsection [ sec3.4 ] along with possible approximations for @xmath202 .
the temperature change @xmath203 characterizes the time evolution of high - energy quasi - equilibrated particles through eqs .
( [ 2.1 ] ) and ( [ 3.1.1 ] ) . in order to derive the temperature law ( [ 3.2.8 ] ) we substitute @xmath204 and @xmath205 given by eq .
( [ 3.2.5 ] ) and ( [ 3.1.1 ] ) , respectively , into the kinetic equation ( [ 2.2 ] ) : @xmath206 \delta
t(t ) + \beta \delta t^{2}(t ) + \gamma(t ) , \label{3.3.1}\ ] ] where @xmath207 and the parameters @xmath208 , @xmath209 , and functional @xmath210 $ ] are defined in appendix .
equations ( [ 3.3.1])-([3.3.2 ] ) refer to some energy @xmath211 from the high - energy domain @xmath157 . in the adiabatic stage of relaxation , when @xmath212 , one has @xmath213 so that @xmath214 and @xmath215 can be neglected on the r.h.s .
( [ 3.3.1 ] ) , because @xmath216 and @xmath217 .
at the end of the adiabatic stage ( @xmath218 ) the distribution function @xmath113 is already very close to @xmath219 even for small energies @xmath172 and , therefore , the phonon - assisted relaxation kinetics becomes exponential , i.e. , @xmath220 .
for statistically - degenerate qw excitons ( @xmath221 ) one estimates from eqs .
( [ 3.2.2 ] ) and ( [ b1 ] ) that positive @xmath222 is much less than positive @xmath208 . in this case
( [ 3.3.1 ] ) can be solved iteratively .
the first iteration , which can formally be obtained by neglecting the time derivative on the r.h.s . of eq .
( [ 3.3.1 ] ) , yields @xmath223 for the same time domain @xmath224 eq .
( [ 3.2.5 ] ) yields the following approximation : @xmath225 by substituting eq .
( [ 3.3.4 ] ) into the r.h.s . of eq .
( [ 3.3.2 ] ) one derives from eq .
( [ 3.3.3 ] ) : @xmath226 where the constant @xmath227 is defined in appendix .
we can further simplify eq .
( [ 3.3.6 ] ) taking into account that @xmath228 is a monotonously increasing function of energy and that at large @xmath229 only small vicinity of @xmath230 contributes to the integral .
finally , we end up with the asymptotic law : @xmath231 where the constant @xmath232 is given in appendix .
equation ( [ 3.3.7 ] ) is identical to eq .
( [ 3.2.8 ] ) provided that the reference ( start ) temperature @xmath233 of the high - energy quasi - equilibrated qw excitons is determined by @xmath234 while the above derivation of eqs .
( [ 3.3.3])-([3.3.8 ] ) assumes that @xmath235 , we have checked numerically that _ the temperature law _ ( [ 3.2.8 ] ) _ holds through the whole adiabatic stage _ ,
i.e. , for the time interval @xmath236 .
furthermore , the numerical evaluations also clearly indicate that the start temperature @xmath237 is practically independent of the local energy @xmath238 used in the derivation of eqs .
( [ 3.3.1])-([3.3.8 ] ) . in fig .
[ fig3 ] we plot @xmath239 and @xmath240 calculated by using the thermalization law ( [ 3.2.8 ] ) ( dashed lines ) and by direct numerical modelling of the initial kinetic eq .
( [ 2.2 ] ) ( solid lines ) , respectively . and @xmath241 ( inset ) calculated by using the thermalization law ( [ 3.2.8 ] ) ( dashed lines ) and by direct numerical modelling of the initial kinetic eq .
( [ 2.2 ] ) ( solid lines ) .
the control parameters @xmath2 and @xmath54 are the same as in fig .
[ fig1 ] . ] for the time domain @xmath242 the thermalization law ( [ 3.2.8 ] ) can be approximated by @xmath243 .
the latter dependence @xmath244 is consistent with that found for the adiabatic stage of the phonon - assisted relaxation kinetics of 3d bosons ( excitons ) at @xmath245 , when bose - einstein condensate builds up . @xcite for 2d systems at nonzero @xmath2 the occupation number of the ground - state mode @xmath246 is always final . therefore the exponential kinetics @xmath247 , which results from eq .
( [ 3.2.8 ] ) for @xmath248 , develops at the final stage of relaxation , when @xmath249 .
note that because for statistically - degenerate qw excitons , when @xmath99 and @xmath221 , one has @xmath250 ( see section [ sec4 ] ) , the two modes of behavior , @xmath251 and @xmath252 , are well - separated in time . in order to determine the relaxation kinetics of low - energy particles ,
i.e. , @xmath253 , one needs to specify the reference distribution @xmath254 and @xmath255 . according to eq .
( [ 3.2.6 ] ) , with increasing time @xmath256 only the small vicinity of @xmath257 gives contribution to the homogeneous part of the generic solution .
therefore we approximate @xmath258 .
the above approximation assumes that the reference time @xmath184 is close enough to the start time @xmath55 of the adiabatic stage of relaxation so that @xmath259 at @xmath260 .
thus , using eq .
( [ 2.1 ] ) we determine the reference distribution at @xmath261 by @xmath262 the latter expression is completely defined by the only one unknown parameter @xmath263 , the population of the ground - state mode at the reference time @xmath183 .
equation ( [ 3.4.1 ] ) allows us to find @xmath264 through the integral relationship ( [ 3.3.6 ] ) taken at @xmath261 .
furthermore , within the approximations used in the derivation of the asymptotic law ( [ 3.3.7 ] ) a simplified ( algebraic ) form of this relationship , given by eqs . (
[ 3.3.8 ] ) and ( [ b5 ] ) , is valid as well .
after the value of @xmath265 is determined , using the thermalization law ( [ 3.2.8 ] ) one can easily find the time dependence of the distribution function ( [ 3.1.1 ] ) of high - energy excitons .
therefore , the three parameters @xmath183 , @xmath263 , and @xmath265 , completely specify the reference point for the generic solution , as illustrated in fig .
if for any energy @xmath91 the distribution function @xmath266 of quantum - degenerate quasi-2d excitons is close enough to final @xmath219 so that @xmath267 and @xmath268 becomes small , the phonon - assisted kinetics can be linearized . in particular
, the adiabatic stage of thermalization at times @xmath269 refers to the linearized kinetics .
the linearized kinetics can be described in terms of the real eigenvalues @xmath270 ( @xmath271 ) and the corresponding eigenfunctions @xmath272 so that @xmath273 . the initial kinetic eq .
( [ 2.2 ] ) reduces to the linear fredholm equation of the second kind with respect to @xmath274 : @xmath275 , \qquad \label{4.1}\end{aligned}\ ] ] where @xmath276 \left ( 1+n^{\rm ph}_{\varepsilon_{1}}+n^{0}_{\varepsilon - \varepsilon_{1 } } \right ) , \label{4.2a}\ ] ] @xmath277 \left ( n^{\rm ph}_{\varepsilon_{1 } } - n^{0}_{\varepsilon + \varepsilon_{1 } } \right ) , \label{4.2b}\ ] ] [ 4.2 ] and functions @xmath278 and @xmath279 are given by eqs .
( [ 2.4a ] ) and ( [ 2.4b ] ) , respectively .
thus we replace the solution of eq .
( [ 2.2 ] ) by the eigenfunction analysis of the fredholm integral equation ( [ 4.1 ] ) .
the numerical solution of eq .
( [ 4.1 ] ) clearly shows that for a given @xmath280 all except one eigenvalues @xmath281 are nondegenerate , positive , and belong to the continuous spectrum .
this is illustrated in the bottom inset of fig .
[ fig4 ] , where the set of eigenvalues shown by the stars covers the same interval @xmath282 more dense the more discrete points in energy space are used .
the isolated non - degenerate eigenvalue @xmath283 is due to conservation of the total number of qw excitons in our model ( the only integral of motion of the system @xcite ) . .
the main part of the figure shows the first order pole , which arises at the energy region @xmath284 ( @xmath285 in this particular example ) .
the top inset illustrates eigenfunction behavior at critical points in the high energy band @xmath286 . the eigenvalue spectrum ( @xmath287 in units @xmath288 ) is shown in the bottom inset .
the control parameters @xmath2 and @xmath54 are the same as in fig .
[ fig1 ] . ]
since all @xmath270 are nondegenerate , the corresponding eigenfunctions @xmath289 form a basis in hilbert energy - space .
for the energy band @xmath290 one derives from eq .
( [ 4.1 ] ) : @xmath291 where @xmath292 is a smooth regular function of @xmath91 given by @xmath293 and @xmath228 is defined by eq .
( [ 3.2.2 ] ) .
thus the eigenfunction @xmath294 has an isolated singularity ( first order pole ) at @xmath295 .
the singularity is integrable in terms of principal value integration . a typical shape of the eigenfunctions at @xmath296 is shown in fig . [ fig4 ] . in the energy band
@xmath297 the eigenfunction has another singularity at the point @xmath298 .
this singularity is logarithmic , i.e. , integrable .
it arises when the singularity of @xmath299 at @xmath300 [ see eq .
( [ 4.2 ] ) ] coincides with the upper boundary of integration , @xmath301 , in the stokes collision term on the r.h.s .
( [ 4.1 ] ) .
the step - like jump at the critical point @xmath302 does not accompanied by discontinuity of the eigenfunction .
the latter critical point originates from a singular behavior of the form - factor function @xmath303 at @xmath304 .
the features of the eigenfunction @xmath274 at points @xmath305 are shown in the top inset of fig .
[ fig4 ] .
the marginal point @xmath306 of the continuous spectrum of @xmath307 is indeed given by @xmath308 : for @xmath309 the singularity point @xmath310 , i.e. , approaches its lowest limit . because @xmath311 is the longest relaxation time generated by the continuum @xmath312 , the eigenavalue @xmath6 determines the phonon - assisted kinetics at @xmath313 and , in particular , yields the characteristic thermalization time in the relaxation thermodynamics developed in ref . .
the dependence of @xmath6 upon the control parameters of the system , @xmath2 and @xmath54 , is plotted in the fig .
[ fig5 ] .
( in units @xmath288 ) as a function of the control parameters @xmath2 and @xmath54 . ]
the eigenvalue @xmath6 can reach both limits , i.e. , @xmath314 and @xmath315 ( see fig .
[ fig5 ] ) .
the first limit , which corresponds to the considerable acceleration of the thermalization kinetics in comparison with that in the 3d case , is due to the relaxation of momentum conservation in qw exciton
bulk acoustic phonon scattering . the slowing down of thermalization , @xmath316 , occurs ( i ) at low bath temperatures @xmath317 , due to the exponentially decreasing number of thermal acoustic phonons with energy @xmath318 , and/or ( ii ) for well - developed quantum statistics ( @xmath319 ) , due to an effective suppression of the stimulated kinetic processes .
there is no direct phonon - mediated interaction of low - energy qw excitons @xmath320 , and at low temperatures @xmath321 the relaxation kinetics occurs by the two - step process : `` low - energy qw exciton ( @xmath260 ) @xmath322 phonon ( @xmath323 ) @xmath324 qw exciton ( @xmath325 ) @xmath324 low - energy qw exciton ( @xmath326 ) @xmath322 phonon ( @xmath327 ) '' .
the first , anti - stokes transition quenches with decreasing temperature @xmath321 and yields a temperature dependent bottleneck effect in thermalization . in turn
, the critical slowing down of the relaxation kinetics at @xmath328 arises due to mutual compensation of two stimulated fluxes , into and out the low - energy qw states @xmath329 .
for example , for the ground - state mode @xmath330 the collision integrand responsible for the stimulated kinetics is given by @xmath331 .
at @xmath332 the latter combination becomes small in spite of a large occupation number @xmath333 , because @xmath334 as a result of a very small value of the chemical potential , @xmath335 .
now we can give an alternative prove of the temperature law ( [ 3.2.8 ] ) which refers to high - energy qw excitons with @xmath336 .
namely , the function @xmath337 can be expanded over the basis @xmath338 .
then the solution of eq .
( [ 2.2 ] ) is given by @xmath339 where @xmath340 are the expansion coefficients . at large times
@xmath341 only a small vicinity of @xmath287 near @xmath6 contributes to the integral , due to the time exponent in the integrand .
if now one assumes a regular distribution of quasi-2d excitons at @xmath342 , the coefficients @xmath340 smoothly depend upon @xmath287 . as a result
, the approximation @xmath343 can be used in the integrand on the r.h.s . of eq .
( [ 4.5 ] ) .
furthermore , the eigenfunctions @xmath344 have nearly the same smooth shape at high energies , as illustrated by fig .
thus we can also put @xmath345 on the r.h.s . of eq .
( [ 4.5 ] ) . as a result ,
both @xmath340 and @xmath294 can be extracted out the integral .
using eq .
( [ 3.1.1 ] ) we immediately get @xmath346 , which coincides with eq .
( [ 3.2.8 ] ) at @xmath347 .
note that the above derivation is based on the particular spectrum ( @xmath348 @xmath322 continuum @xmath349 ) of the linearized collision integral and has no analogy in the relaxation kinetics due to particle - particle interaction . in the latter case the fivefold degenerate eigenvalue @xmath350
is separated from the continuous spectrum by a set of discrete isolated eigenvalues .
in order to test the generic solution ( [ 3.2.2])-([3.2.9 ] ) we model the phonon - assisted relaxation of excitons within the initial kinetic eq .
( [ 1.1 ] ) reduced to energy space [ eqs .
( [ 2.1])-([2.2 ] ) ] . an adaptive inhomogeneous grid with 100 - 200 points for @xmath91
is used to cover the close vicinity of the ground - state mode @xmath351 ( the maximum value of the dimensionless energy is @xmath352 ) .
equation ( [ 2.2 ] ) is evaluated by a fourth - order runge - kutta integration routine with the time step @xmath353 . in order to calculate integrals on the r.h.s .
( [ 2.2 ] ) we perform spline - approximation for @xmath268 at every iterative step . in numerical simulations we use the dimensionless temperatures @xmath2 and @xmath54 and measure time in @xmath354 .
this makes our results suitable for various qws and sets of the control parameters , provided that @xmath67 , and @xmath354 are specified . in fig .
[ fig6 ] time evolution of the distribution @xmath355 as a numerical solution of eq .
( [ 2.2 ] ) is compared with the corresponding generic solution ( [ 3.2.2])-([3.2.9 ] ) relevant to @xmath356 .
all plots demonstrate an excellent agreement between analytical and numerical solutions .
note , that at high @xmath2 , e.g. , @xmath357 ( see the top plot in fig . [ fig6 ] ) , the thermalization time @xmath358 achieves the limit @xmath359 , and the relative duration of the adiabatic stage estimated in terms of @xmath55 ( duration of the first transient ) becomes smaller than that at @xmath360 . in this case
the influence of the initial distribution @xmath88 slightly affects the calculated evolution at the beginning of the adiabatic stage , and the analytical solution fits numerically simulated data a little worse as can be seen for the distribution functions at @xmath361 . in contrast , at low @xmath2 the relaxation kinetics at the adiabatic stage is slow .
for example , at @xmath362 it lasts more than @xmath363 scattering times , where typical values of @xmath56 in gaas / algaas coupled qws are on a scale of tens nanoseconds . at the adiabatic stage calculated for various sets of the control parameters , @xmath2 and @xmath54 .
solid lines correspond to numerical evaluation of eq .
( [ 2.2 ] ) ; dashed lines are obtained using the generic solution given by eq .
( [ 3.2.5 ] ) . ] in fig . [ fig7 ]
we also compare time dependences for the ground - state mode population @xmath364 obtained numerically ( solid lines ) and analytically ( dashed lines ) .
again , this figure shows that at @xmath351 the generic solution ( [ 3.2.2])-([3.2.9 ] ) reproduces the adiabatic stage of the phonon - assisted relaxation kinetics very well . within the time interval
@xmath365 the generic solution yields the following simple approximation for the adiabatic kinetics into the ground state mode : @xmath366^{\nu}. \label{5.1}\ ] ] parameters @xmath102 and @xmath103 can be found comparing series expansions of eq .
( [ 5.1 ] ) and ( [ 3.2.5 ] ) about the point @xmath261 .
in such a way we get @xmath367 and @xmath368 , where time - independent constants @xmath369 and @xmath370 are given by @xmath371 + t_b \delta t_0 \eta(0 ) ( n^{0}_{\varepsilon=0 } ) ' , \nonumber\\ & c_{2}=&\frac{1}{2 } \biglb ( ( n_{0}^{\rm in } - n^{0}_{\varepsilon=0 } ) \nonumber\\ & & \times \ { 2 \lambda_{0}^{2 } + \delta t_0 \xi_{1}(0 ) [ 5 \lambda_{0}+\lambda_1 + 2 \delta t_0 \xi_{1}(0 ) ] \ } \nonumber\\ & & - t_b \delta t_0 \eta(0 ) ( n^{0}_{\varepsilon=0 } ) ' [ 3 \lambda_0 + \lambda_1 + 2 \delta t_0 \xi_{1}(0 ) ] \bigrb ) .
\label{5.2}\end{aligned}\ ] ] time dependences of the ground - state mode population @xmath372 calculated at @xmath373 and @xmath374 using eq .
( [ 5.1 ] ) are shown in fig .
[ fig7](a ) with dash - dotted lines . at @xmath375
the approximation ( [ 5.1 ] ) violates because the phonon - assisted relaxation kinetics starts to become exponential . calculated for different bath temperatures at @xmath376 ( a ) , and @xmath65 ( b ) . similarly to fig
. [ fig6 ] , the solid and dashed curves are calculated by using eqs .
( [ 2.2 ] ) and ( [ 3.2.5 ] ) , respectively .
dash - dotted lines correspond to time dependences @xmath372 calculated using approximate eq .
( [ 5.1 ] ) . ] as we have shown in sec .
[ sec4 ] , for a given concentration of qw excitons @xmath377 @xmath10 the thermalization kinetics slows down with decreasing @xmath378 , i.e. , with development of quantum statistics . by increasing both temperatures , @xmath54 and @xmath2 , and keeping unchanged the ratio
@xmath379 one can simultaneously avoid the above bottleneck effect in relaxation and achieve high population of the ground - state mode , @xmath380 .
however , in this case the concentration of excitons @xmath381 increases as well , so that exciton - exciton interaction eventually becomes the main mechanism of relaxation in gaas / algaas coupled or znse single qws .
in contrast , in high - quality gaas - based microcavities with a relatively large positive detuning @xmath382 between the cavity mode ( @xmath383 ) and qw exciton line ( @xmath384 ) the la - phonon - assisted kinetics remains dominant even at relatively high degeneracy temperatures @xmath385 . in these systems , by changing the detuning @xmath386 within the band @xmath387 mev one can design an effective in - plane mass @xmath19 so that the bottleneck effect in phonon - assisted scattering , due to the low density of states @xmath388 , is already relaxed , whereas the degeneracy temperature @xmath389 is still relatively high .
indeed , in such mcs the lower polariton branch gives rise to the in - plane translational mass much smaller than the mass of optically - undressed qw excitons . in the meantime
the excitonic component @xmath390 of the microcavity polaritons is already very high , @xmath391 , resulting in the long optical decay ( in the @xmath13-direction ) lifetimes on a few nanosecond time scale .
this is illustrated in fig .
[ fig8](a ) , where the detuning @xmath386 is equal to @xmath392 , @xmath393 , and @xmath394 mev .
the corresponding polariton ( exciton ) masses are given by @xmath395 , @xmath396 , and @xmath397 , respectively .
the relevant lower - branch polariton dispersions are plotted in fig .
[ fig8](b ) .
because the energy @xmath67 is only on a @xmath398 mev energy scale , the parabolic approximation of the lower - branch dispersion curves is valid for low - temperature relaxation kinetics .
thus the la - phonon - assisted thermalization of the exciton - like mc polaritons can indeed be modelled by the kinetic eqs .
( [ 2.1])-([2.2 ] ) .
time evolutions of the exciton distribution @xmath113 , which are typical for the proposed design of gaas - based mcs , are shown in fig .
[ fig6 ] ( see the plots with @xmath399 , and @xmath400 ) .
figure [ fig7](a ) illustrates the corresponding population dynamics of the ground - state mode , @xmath364 . of the mc polariton eigenstate ; and ( b ) the lower branch polariton dispersion @xmath401 .
detuning @xmath402 mev ( dashed lines ) , @xmath393 mev ( solid lines ) , and @xmath394 mev ( dash - dotted lines ) .
the energy of ground - state qw excitons is given by @xmath403 ev . ]
in this paper we have studied thermalization kinetics of statistically degenerate qw excitons coupled to thermal bulk acoustic phonons . for concentrations of qw excitons @xmath404 @xmath10 the particle - particle interaction in gaas or znse qws can be neglected in comparison with qw exciton bulk acoustic phonon scattering , and
the thermalization kinetics from an initial distribution of qw excitons at @xmath49 occurs through the nonequilibrium states .
the following conclusions summarize our results .
\(i ) for the case of well - developed bose - einstein statistics , when @xmath405 so that @xmath406 , the relaxation kinetics of qw excitons coupled to thermal bulk acoustic phonons is given by the following scheme . within a few characteristic scattering times the correlation of the distribution function @xmath113 with the initial @xmath88 disappears , and the subsequent thermalization of qw excitons is described in terms of the adiabatic stage of relaxation .
the adiabatic stage is characterized by the start time @xmath55 , which absorbs a gross information about the initial distribution @xmath88 , and by the parameter @xmath6 , which depends only upon the bath and degeneracy temperatures , @xmath2 and @xmath54 . at the beginning of the adiabatic stage ,
i.e. , for the time domain @xmath407 , one has @xmath408 , and the thermalization kinetics is strongly nonexponential , with @xmath409 and @xmath410 . at large times , when the deviation of the system from the final equilibrium state is already small ( @xmath411 ) , the adiabatic stage of the phonon - assisted thermalization becomes exponential , @xmath412 , and can be described within the linearized kinetic equation .
\(ii ) the linearized la - phonon - assisted kinetics of qw excitons is formulated in terms of the fredholm integral eq .
( [ 4.1 ] ) .
the eigenvalues @xmath307 of the collision integral are given by the continuous spectrum @xmath413 and the isolated eigenvalue @xmath350 .
the marginal eigenvalue @xmath414 determines the thermalization time at @xmath75 by @xmath415 . in dependence on the two control parameters , @xmath2 and @xmath416 , the thermalization time achieves two limits : @xmath417 and @xmath418 .
the eigenfunctions @xmath274 of the collision integral ( [ 4.1 ] ) are smooth integrable functions with three isolated critical points in energy space .
the critical points of @xmath274 give rise to a first order pole , a logarithmic singularity , and a continuous step - like jump .
\(iii ) because the la - phonon - assisted kinetics becomes dominant only at small concentrations of qw excitons , @xmath419 @xmath10 , nonclassical statistics of quasi-2d excitons in znse or gaas qws develops at very low bath temperatures @xmath420 k. the proposed design of gaas - based microcavities with a relatively large positive detuning between the cavity mode and qw exciton line , @xmath421 mev , allows us , however , to build up @xmath422 by means of qw exciton
bulk la phonon scattering in much more favorable conditions , i.e. , @xmath423 k and @xmath424 .
we appreciate valuable discussions with l.v .
butov , v. falko , and m.s .
support of this work by the epsrc ( u.k . ) is gratefully acknowledged .
in appendix we give the expressions for the parameters and functions used in eqs .
( [ 3.3.1 ] ) , ( [ 3.3.6 ] ) , and ( [ 3.3.7 ] ) .
the parameters @xmath208 , and @xmath209 arise from those terms in the collision integral which contain @xmath425 and @xmath426 , respectively .
collecting such terms together one gets @xmath427 , \label{b1}\end{aligned}\ ] ] and @xmath428 .
\label{b2}\end{aligned}\ ] ] the function @xmath429 stems from the terms proportional to @xmath430 and is given by @xmath431 the parameter @xmath227 from eq .
( [ 3.3.3 ] ) is obtained by collecting all the terms @xmath432 : @xmath433 when deriving eq .
( [ 3.3.7 ] ) from eq .
( [ 3.3.6 ] ) we first put @xmath434 everywhere in the integrand ( [ 3.3.6 ] ) except the exponent . by changing the integration variable @xmath91 to @xmath435 and putting @xmath330
we derive the asymptotic eq .
( [ 3.3.7 ] ) , valid in the limit @xmath436 . in this equation @xmath232 is a time - independent pre - factor given by @xmath437 for a review of the experimental works on relaxation and photoluminescence kinetics of qw excitons see , e.g. , j. shah , _ ultrafast spectroscopy of semiconductors and semiconductor nano - structures _ , springer series in solid - state sciences , vol .
115 ( springer , berlin , 1996 ) .
t. takagahara , phys .
b * 31 * , 6552 ( 1985 ) .
a. vinattieri , j. shah , t.c .
damen , d.s .
kim , l.n .
pfeiffer , m.z .
maialle , and l.j .
sham , phys .
b * 50 * , 10868 ( 1994 )
selbmann , m. gulia , f. rossi , e. molinari , and p. lugli , phys .
b * 54 * , 4660 ( 1996 ) .
m. gurioli , p. borri , m. colocci , m. gulia , f. rossi , e. molinari , p.e .
selbmann , and p. lugli , phys .
b * 58 * , r13403 ( 1998 )
. m. gulia , f. rossi , e. molinari , p.e .
selbmann , and p. lugli , phys .
b * 55 * , r16049 ( 1997 ) .
oh , j. singh , a. thilagam , and a.s .
vengurlekar , phys .
b * 62 * , 2045 ( 2000 ) .
a.l . ivanov , p.b .
littlewood , and h. haug , phys .
b * 59 * , 5032 ( 1999 ) .
damen , j. shah , d.y .
oberli , d.s .
chemla , j.e .
cunningham , and j.m .
kuo , phys .
b * 42 * , 7434 ( 1990 ) .
r. eccleston , r. strobel , w.w .
rhle , j. kuhl , b.f .
feuerbacher , and k. ploog , phys .
b * 44 * , 1395 ( 1991 ) ; r. eccleston , b.f .
feuerbacher , j. kuhl , w.w .
rhle , and k. ploog , phys .
b * 45 * , 11403 ( 1992 ) .
t. fukuzawa , e.e .
mendez , and j.m .
hong , phys .
lett . * 25 * , 3066 ( 1990 ) . however , the experimental results of this work were later reinterpreted .
butov , a. zrenner , g. abstreiter , g. bhm , and g. weimann , phys .
73 * , 304 ( 1994 ) .
butov and a.i .
filin , phys .
b * 58 * , 1980 ( 1998 ) .
butov , a. imamoglu , a.v .
mintsev , k.l .
campman , and a.c .
gossard , phys .
b * 59 * , 1625 ( 1999 ) . l.v .
butov , a.l .
ivanov , a. imamoglu , p.b .
littlewood , a.a .
shashkin , v.t .
dolgopolov , k.l .
campman , and a.c .
gossard , phys .
86 * , 5608 ( 2001 ) .
a.v . larionov and v.b .
timofeev , pisma zh .
eksp . teor
. fiz . * 73 * , 342 ( 2001 ) [ jetp lett . * 73 * , 301 ( 2001 ) ] .
krivolapchuk , e.s .
moskalenko , and a.l .
zhmodikov , phys .
b * 64 * , 045313 ( 2001 ) .
s. charbonneau , m. l. w. thewalt , e.s .
koteles , and b. elman , phys .
b * 38 * , 6287 ( 1988 ) .
a. alexandrou , j.a .
kash , e.e .
mendez , m. zachau , and j.m .
hong , phys .
b * 42 * , 9225 ( 1990 ) .
golub , k. kash , j.p .
harbison , and l.t .
florez , phys .
b * 41 * , 8564 ( 1990 ) .
golub , k. kash , j.p .
harbison , and l.t .
florez , phys .
b * 45 * , 9477 ( 1992 ) .
f. clrot , b. deveaud , a. chromette , a. regreny , and b. sermage , phys .
b * 41 * , 5756 ( 1990 ) .
a. parlangeli , p.c.m .
christianen , j.c .
maan , i.v .
tokatly , c.b .
soerensen , and p.e .
lindelof , phys .
b * 62 * , 15323 ( 2000 ) .
m. umlauff , j. hoffmann , h. kalt , w. langbein , j.m .
hvam , m. scholl , j. sllner , m. heuken , b. jobst , and d. hommel , phys .
b * 57 * , 1390 ( 1998 ) ; h. kalt , j. hoffmann , m. umlauff , w. langbein , and j.m .
hvam , phys .
status solidi b * 206 * , 103 ( 1998 ) .
g. bacher , r. spiegel , t. kmmell , r. weigand , a. forchel , b. jobst , d. hommel , and g. landwehr , j. cryst . growth * 184/185 * , 330 ( 1998 ) .
e.a . uehling and g.e .
uhlenbeck , phys . rev . * 108 * , 1175 ( 1932 ) .
u. bockelmann , phys .
b * 50 * , 17271 ( 1994 ) .
yu.e . lozovik and v.i .
yudson , pisma zh .
* 22 * , 556 ( 1975 ) [ jetp lett .
* 22 * , 274 ( 1975 ) ] .
x. zhu , p.b .
littlewood , m. hybertsen , and t. rice , phys .
lett . * 74 * , 1633 ( 1995 ) .
littlewood and x. zhu , physica scripta * 68 * , 56 ( 1996 ) .
y. naveh and b. laikhtman , phys .
rev . lett . * 77 * , 900 ( 1996 ) .
savvidis , j.j .
baumberg , r.m .
stevenson , m.s .
skolnick , d.m .
whittaker , and j.s .
roberts , phys .
* 84 * , 1547 ( 2000 ) .
tartakovskii , m. emam - ismail , r.m .
stevenson , m.s .
skolnick , v.n .
astratov , d.m .
whittaker , j.j .
baumberg , and j.s .
roberts , phys .
b * 62 * , r2283 ( 2000 ) . f. boeuf , r. andr , r. romestain , le si dang , e. pronne , j.f .
lampin , d. hulin , and a. alexandrou , phys .
b * 62 * , r2279 ( 2000 ) . r.m .
stevenson , v.n .
astratov , m.s .
skolnick , d.m .
whittaker , m. emam - ismail , a.i .
tartakovskii , p.g .
savvidis , j.j .
baumberg , and j.s .
roberts , phys .
lett . * 85 * , 3680 ( 2000 ) .
baumberg , p.g .
savvidis , r.m .
stevenson , a.i .
tartakovskii , m.s .
skolnick , d.m .
whittaker , and j.s .
roberts , phys .
b * 62 * , r16247 ( 2000 ) . p.g .
savvidis , c. ciuti , j.j .
baumberg , d.m .
whittaker , m.s .
skolnick , and j.s .
roberts , phys .
b * 64 * , 075311 ( 2001 ) .
s.d . baranovskii , r. eichmann , and p. thomas , phys .
b * 58 * , 13081 ( 1998 ) .
c. bradford , c. b. odonnell , b. urbaszek , c. morhain , a. balocchi , k. a. prior , and b. c. cavenett , phys .
b * 64 * , 195309 ( 2001 ) .
a.l . ivanov , c. ell , and h. haug , phys .
e * 55 * , 6363 ( 1997 ) ; phys .
status solidi b * 206 * , 235 ( 1998 ) . c. cercignani , _ the boltzmann equation and its applications _ , springer series in applied mathematical sciences vol .
67 ( springer , new york , 1988 ) , sect .
iv-6 , pp .
180 - 189 . | acoustic - phonon - assisted thermalization kinetics of excitons in quantum wells ( qws ) is developed for small concentrations of particles , @xmath0 @xmath1 , when particle - particle interaction can be neglected while bose - einstein statistics already strongly influences the relaxation processes at low temperatures . in this case
thermalization of qw excitons occurs through nonequilibrium states and is given by the following scenario . during the first transient stage , which lasts a few characteristic scattering times ,
the correlations with an initial distribution of qw excitons disappear .
the next , adiabatic stage of thermalization usually takes many characteristic scattering times , depends only upon two control parameters , the lattice temperature @xmath2 and the degeneracy temperature @xmath3 , and is characterized by a quasi - equilibrated distribution of high - energy qw excitons with effective temperature @xmath4 .
we show that the thermalization law of high - energy particles is given by @xmath5 , where @xmath6 is a marginal value of the continuous eigenvalue spectrum of the linearized kinetics . by analyzing the linearized phonon - assisted kinetics of statistically - degenerate qw excitons , we study the dependence @xmath7 .
our numerical estimates refer to high quality gaas and znse qws .
finally , we propose a special design of gaas - based microcavities , which considerably weakens the bottleneck effect in relaxation of excitons ( polaritons ) and allows us to optimize the acoustic - phonon - assisted thermalization processes . |
for more than 30 years , solar neutrino experiments @xcite have been observing fewer neutrinos than what are predicted by the detailed models @xcite of the sun .
a comparison of the predicted and observed solar neutrino fluxes for these experiments are shown in table [ tbl : solarnuexp ] .
these experiments probe different parts of the solar neutrino energy spectrum , and show an energy dependence in the observed solar neutrino flux .
these observations can be explained if the solar models are incomplete or neutrinos undergo flavor transformation while in transit to the earth .
the sudbury neutrino observatory was constructed to resolve this solar neutrino puzzle .
it is capable of making simultaneous measurements of the electron - type neutrino ( @xmath1 ) flux from @xmath2b decay in the sun and the flux of all active neutrino flavors through the following reactions : @xmath6 the charged - current ( cc ) reaction on the deuteron is sensitive exclusively to @xmath1 , and the neutral - current ( nc ) reaction has equal sensitivity to all active neutrino flavors ( @xmath7 , @xmath8 ) .
elastic scattering ( es ) on electron is also sensitive to all active flavors , but with reduced sensitivity to @xmath3 and @xmath4 .
comparison of the solar neutrino flux inferred from the reaction rates of these three reaction channels under the no - oscillation assumption can provide `` smoking gun '' evidence for flavor - changing neutrino oscillations . if @xmath1 s from the sun transform into another active flavor , then the solar neutrino flux deduced from the cc reaction rate ( @xmath9(@xmath1 ) )
must be less than those deduced from the es reaction rate or the nc reaction rate .
this is summarized in figure [ fig : smoking_gun ] .
recent results @xcite from the first measurements of the solar @xmath2b neutrino flux by the sno detector using the cc and es reactions are presented in this paper .
the measured @xmath10(@xmath7 ) is consistent with the high precision es measurement by the super - kamiokande collaboration @xcite .
the measured @xmath9(@xmath1 ) at sno , however , is significantly smaller and is therefore inconsistent with the null hypothesis of a pure @xmath1 constituent in the solar neutrino flux .
.summary of solar neutrino observations at different solar neutrino detectors . [ cols="<,<,<,<,<",options="header " , ] [ tbl : sysuncert ]
the @xmath2b neutrino flux can be determined from normalizing the observed integrated event rate above the energy threshold . assuming the @xmath2b spectrum from ref .
@xcite , the flux deduced from the cc and the es reactions are : @xmath11 where the theoretical uncertainty is the cc cross section uncertainty @xcite .
radiative corrections to the cc cross section have not been applied to the cc cross section , but they are expected to decrease the measured @xmath12(@xmath1 ) by up to a few percent @xcite .
the difference between @xmath12 and @xmath13 is [email protected]@xmath1510@xmath16 @xmath17s@xmath18 , or 1.6@xmath5 .
the ratio of @xmath12 to the predicted @xmath2b solar neutrino flux from bpb01 solar model @xcite is [email protected] where all the uncertainties are added in quadrature .
independent analyses using n@xmath19 as energy estimator , or in various fiducial volumes up to 620 cm with the inclusion of background pdfs in the signal extraction give consistent results .
the super - kamiokande experiment has made a high precision measurement of the @xmath2b solar neutrino flux by the es reaction : @xmath20 @xmath13(@xmath7 ) and @xmath21(@xmath7 ) are consistent with each other . assuming that the systematic uncertainties are normally distributed , the difference is @xmath22 or 3.3@xmath5 .
the probability that the observed @xmath12(@xmath1 ) is a @xmath233.3@xmath5 downward fluctuation of @xmath21(@xmath7 ) is 0.04% .
if @xmath1 s from @xmath2b decays in the sun oscillate exclusively to sterile neutrinos , the sno cc - derived @xmath2b flux with @[email protected] mev would be consistent with the integrated super - kamiokande es - derived flux above a threshold of 8.5 mev @xcite .
the difference between these derived fluxes after adjusting for the es threshold @xcite is [email protected]@xmath25 @xmath17s@xmath18 , or 3.1@xmath5 .
the probability of a @xmath233.1@xmath5 downward fluctuation is 0.13% .
therefore , the results presented here are evidence for a non - electron type active neutrino component in the solar neutrino flux . the cc energy spectrum can be extracted from the data by repeating the signal extraction with the cc energy spectral constraint removed .
this is shown in figure [ fig : sno_egy_spec ] .
there is no evidence for spectral distortion under the no - oscillation hypothesis .
[ fig : sno_egy_spec ]
recall from figure [ fig : smoking_gun ] that the es reaction is sensitive to all active neutrino flavors , but with reduced sensitivity for @xmath3 and @xmath4 . using the high precision es measurement @xmath21(@xmath7 ) and the pure @xmath1 flux from @xmath12(@xmath1 )
, one can infer the flux of non - electron flavor active neutrino @xmath26 : @xmath27 this is shown in figure [ fig : fig_3 ] , in which @xmath26 is shown against @xmath28 .
the two data bands are 1@xmath5 measurements of @xmath12 and @xmath21 , and the error ellipses are 68% , 95% and 99% joint probability contours for @xmath28 and @xmath26 . the best fit to @xmath26is @xmath29 the total @xmath2b flux derived from the sno and the super - kamiokande experiments is shown as the diagonal band ( @xmath30 ) in figure [ fig : fig_3 ] .
the agreement with the standard solar model prediction ( @xmath31 ) is good .
the total flux of of active @xmath2b neutrinos is found to be @xmath32 this is the first determination of the total flux of @xmath2b neutrinos generated by the sun .
= 3.5 in [ fig : fig_3 ]
two significant results are reported in this paper .
the data from sno represent the first direct evidence that there is an active non - electron flavor neutrino component in the solar neutrino flux , and an exclusive oscillation from @xmath1 s to sterile neutrinos is disfavored at the 3.1@xmath5 level .
this is also the first experimental determination of the total flux of active @xmath2b solar neutrinos , which is in good agreement with the solar model predictions .
the sno collaboration is now analyzing the data from the pure d@xmath0o phase with a lowered energy threshold .
efforts are devoted to understanding the low energy @xmath33 and the photodisintegration contribution to the nc measurement .
results from this analysis will be reported in the near future .
the sno experiment has just finished the first phase of the experiment .
the deployment of nacl to enhance the nc capability began on may 28 , 2001 .
figure [ fig : salt_time ] shows the detector background level seen in the cherenkov data before , during and after the nacl injection .
the increase in the event rate during the injection is attributed to @xmath34na , which were activated by neutrons from the cavity wall when the nacl brine was stored in the underground laboratory prior to the injection . after the injection has ended
, one sees the decay of @xmath34na with a characteristic @xmath35=15 hours . the background level in the detector returned to the pre - injection level after several days .
after 8 months of running in this configuration , sno will be able to make a definitive , solar model - independent statement ( better than 6@xmath5 ) on the solar neutrino oscillation hypothesis .
[ fig : salt_time ]
this research was supported by the natural sciences and engineering research council of canada , industry canada , national research council of canada , northern ontario heritage fund corporation and the province of ontario , the united states department of energy and in the united kingdom by the science and engineering research council and the particle physics and astronomy research council .
further support was provided by inco , ltd . , atomic energy of canada limited ( aecl ) , agra - monenco , canatom , canadian microelectronics corporation , at&t microelectronics , northern telecom and british nuclear fuels , ltd .
the heavy water was loaned by aecl with the cooperation of ontario power generation .
et al . _ ,
lett . * 65 * , 1297 ( 1990 ) ; k.s .
et al . _ ,
phys . rev .
* d44 * , 2241 ( 1991 ) ; k.s .
et al . _ ,
phys . rev . * d45 * , 2170e ( 1992 ) ; y. fukuda _ et al .
lett . * 77 * , 1683 ( 1996 ) .
this is the combined results from kamiokande ii and iii .
et al . _ ,
phys . rev .
* c60 * , 055801 ( 1999 ) .
this is the most recent published result .
the most recent unpublished result from the sage website ( http://ewiserver.npl.washington.edu/sage/sage.html ) is 77@xmath146@xmath143 snu , which include extractions up to december 2000 .
s. nakamura , t. sato , v. gudkov and k. kubodera , phys .
* c63 * , 034617 ( 2001 ) ; m. butler , j .- w .
chen , and x. kong , phys .
* c63 * , 035501 ( 2001 ) ; g. t hooft , phys . lett . *
37b * 195 ( 1971 ) .
the butler _
cross section with @xmath36=5.6 @xmath37 is used .
towner , j. beacom , and s. parke , private communication ; i.s .
towner , phys . rev . *
c58 * 1288 ( 1998 ) ; j. beacom and s. parke , hep - ph/0106128 ; j.n .
bahcall , m. kamionkowski and a. sirlin , phys .
* d51 * 6146 ( 1995 ) .
fogli , e. lisi , a. palazzo , and f.l .
villante , phys . rev . *
d63 * , 113016 ( 2001 ) ; f.l .
villante , g. fiorentini and e. lisi , phys .
* d59 * , 013006 ( 1999 ) ; g.l .
fogli , e. lisi , d. montanino , and a. palazzo , hep - ph/0106247 ( 2001 ) .
the most recent paper in this series indicates that the super - kamiokande threshold should be 8.6 mev , instead of 8.5 mev that is used here . | the sudbury neutrino observatory ( sno ) is a water imaging cherenkov detector .
its usage of 1000 metric tons of d@xmath0o as target allows the sno detector to make a solar - model independent test of the neutrino oscillation hypothesis by simultaneously measuring the solar @xmath1flux and the total flux of all active neutrino species .
solar neutrinos from the decay of @xmath2b have been detected at sno by the charged - current ( cc ) interaction on the deuteron and by the elastic scattering ( es ) of electrons . while the cc reaction is sensitive exclusively to @xmath1 , the es reaction also has a small sensitivity to @xmath3 and @xmath4 . in this paper ,
recent solar neutrino results from the sno experiment are presented .
it is demonstrated that the solar flux from @xmath2b decay as measured from the es reaction rate under the no - oscillation assumption is consistent with the high precision es measurement by the super - kamiokande experiment .
the @xmath1 flux deduced from the cc reaction rate in sno differs from the super - kamiokande es results by 3.3@xmath5 .
this is evidence for an active neutrino component , in additional to @xmath1 , in the solar neutrino flux .
these results also allow the first experimental determination of the total active @xmath2b neutrino flux from the sun , and is found to be in good agreement with solar model predictions .
address = institute for nuclear and particle astrophysics , lawrence berkeley national laboratory , berkeley , ca , usa , [email protected] |
initial results from superkamiokande @xcite appear to confirm indications from imb , @xcite kamiokande @xcite and soudan @xcite of an excess of @xmath0 relative to @xmath1 in the atmospheric neutrinos .
one possible interpretation is that neutrino flavor oscillations play a role . in a two flavor mixing scheme , for example
, the probability that a neutrino of flavor @xmath2 and energy @xmath3 retains its identity after propagating a distance @xmath4 in vacuum is @xcite @xmath5,\ ] ] where @xmath6 is the difference in mass squared of the two neutrino mass eigenstates and @xmath7 is the mixing angle .
therefore , to evaluate the manifestation of the mixing in a detector that measures to some degree the direction and energy of neutrino - induced events , one needs to know the distributions of production heights of the neutrinos as a function of energy and zenith angle .
more complicated mixing schemes @xcite and effects of propagation in matter @xcite still require this basic information about the points of origin of the neutrinos .
information about origin of the neutrinos is implicit in any calculation of neutrino fluxes . here
we extract the relevant information from the simulation of ref .
@xcite , which has been compared to several other calculations in ref .
@xcite .
the paper is organized in three sections .
first we review the simulation we are using to calculate production of neutrinos in the atmosphere .
next we present the basic results of the calculation .
we discuss simple analytic approximations which offer insight into the systematics of the results and compare them to simulation results for zenith angles from the vertical to horizontal .
finally , we provide some parametrizations , based on the analytic approximations , that may be useful for practical application of the results .
the simulation was performed in the spirit of earlier calculations of the atmospheric neutrino flux @xcite .
the simulation code is one dimensional . in this approximation ,
all secondaries are assumed to move in the direction of the primary particles ( except for a small fraction of low energy secondaries with angles larger than @xmath8 to the beam , which are discarded ) .
the validity of this approximation has been checked in refs .
@xcite . the primary cosmic ray flux and its composition is the parametrization used previously in the calculation of agrawal _ et al . _
@xcite which in the multi - gev range falls in between the measurements of refs .
incident cosmic - ray nuclei are treated in the superposition approximation @xcite , with cascades generated separately for protons and neutrons in order to insure the correct ratios of neutrinos and antineutrinos .
the fraction of neutrons is derived from the fractions of nuclei heavier than hydrogen in the primary flux .
we consider three ranges of neutrino energies that correspond approximately to the three major types of experimental events in a detector the size of superkamiokande : contained events ; partially contained neutrino interactions and stopping neutrino induced muons ; and througoing muons .
the energy ranges are presented in two different ways : * @xmath9 gev ; @xmath10 gev ; @xmath11 gev and * @xmath12 and @xmath13 gev .
the integral form is more closely related to simple analytic approximations that we use as the basis of parametrizations of the results of monte carlo simulations .
our results are obtained with the geomagnetic cutoffs for kamioka and for the epoch of solar minimum , which is applicable to measurements performed currently ( @xmath14 ) . because of the high geomagnetic cutoffs at kamioka , it is not necessary to account precisely for the phase of the solar cycle . to illustrate the potential influence of geomagnetic effects at other locations we also tabulate some results for the much higher geomagnetic latitude of the sno experiment .
we have not included prompt neutrino production through charm decay because it is totally negligible in the considered energy ranges @xcite .
all neutrinos are generated either in pion and kaon decays or in muon decays .
the production heights are stored separately for neutrinos from @xmath15 and from muon decays .
the muon decay procedure accounts for the muon enegy loss during propagation in the atmosphere .
technically the muon lifetime is sampled in the muon rest frame and then the muon is propagated in the atmosphere with time dilation proportional to its decreasing energy . thus mouns decay on the average
sooner that they would have if one ( incorrectly ) sampled from a decay distribution using their lorentz factor at production .
before presenting the results for neutrinos , we show a comparison between measurements of gev muons at different altitudes in the atmosphere and our calculation made with the same monte carlo code @xcite .
this type of balloon measurement provides the most direct test of the validity of the cascade model and of the treatment of the muon propagation in the atmosphere because the muons and neutrinos have a common origin . the comparison shown in fig .
1 is with data of the mass experiment @xcite as discussed in ref .
@xcite .
2 shows the height of production of neutrinos of energy above 1 gev for cos@xmath16 = 0.75 .
the graph gives @xmath17 ( @xmath18s@xmath19sr@xmath19km@xmath19 ) , where @xmath20 is the slant distance from the neutrino production point to sea level .
contributions from muon decay and from @xmath15 decay are shown separately for @xmath21 and for @xmath22 .
the overall flux of @xmath23 from @xmath15 decay is much lower because it reflects primarily the contribution of @xmath24 decays , which is very low in this energy range .
the curves for electron and for muon neutrinos from muon decay are nearly equal .
they extend to lower altitudes with a slope that depends on the average energy of the parent muons . for higher energy neutrinos
this slope is signicantly flatter as a consequence of the higher parent muon energy and correspondingly longer muon decay length . for @xmath25 20 gev ,
most parent muons reach the ground ( except in nearly horizontal direction ) and stop before decaying . as a consequence ,
the height distributions for neutrinos from muon decay deep in the atmosphere are nearly flat .
it is instructive to look at a simple approximation for the height of production of neutrinos from decay of pions . in the approximation of an exponential atmosphere with scale height @xmath26 and the approximation of feynman scaling for the production cross sections of pions in interactions of hadrons with nuclei of the atmosphere
, a straightforward solution of the equations for propagation of hadrons through the atmosphere @xcite gives @xcite @xmath27 for the integral flux of neutrinos in the energy range @xmath28 where reinteraction of pions in the atmosphere can be neglected .
there is a similar expression for neutrinos from decay of kaons proportional to @xmath29 .
the meaning and approximate values of the quantities in these equations are given in table [ tab1 ] .
cccccccccc @xmath30 & @xmath31 & @xmath32 & @xmath33 & @xmath34 & @xmath35 & @xmath36 & @xmath37 & @xmath38 & @xmath26 + & @xmath18s@xmath19sr@xmath19(gev)@xmath39 & g/@xmath40 & g/@xmath40 & & & & & g/@xmath40 & km + & & & & & & & & & + 1.70 & 1.8 & 86 & 120 & 0.08 & 0.0075 & 0.5731 & 0.0458 & 1030 & 6.4 + & & & & & & & & & + [ tab1 ] in eq .
[ production ] @xmath41 is the slant depth in the atmosphere at which the pion is produced and decays .
we now convert this into distance @xmath42 from the detector in the approximation of an exponential atmosphere in which @xmath43 . \label{atmos}\ ] ] for @xmath44 , curvature of the earth can be neglected and @xmath45 in eq . [ atmos ] is cosine of the zenith angle to a good approximation .
the effective values of @xmath45 for larger angles are given below .
the corresponding approximate expression for the distribution of production distances is @xmath46 \times\exp\left[-{\ell\cos\theta\over h_0}\right ] , \label{nupi}\ ] ] where @xmath41 is to be evaluated as a function of @xmath42 from eq .
[ atmos ] . assuming a primary cosmic ray nucleon flux with the normalization given in table [ tab1 ] ( and including the small contribution from decay of kaons ) the normalization factor is @xmath47 0.020 .
taking parameters from table [ tab1 ] gives the most probable distance of production as @xmath48 which is @xmath49 km for vertical neutrinos from decay of pions .
3 shows the distance distribution for neutrinos from @xmath15 decay ( @xmath50 1 gev ) for cos@xmath16 = 1.00 , 0.75 , 0.50 , 0.25 , 0.15 and 0.05 . here
@xmath7 is the zenith angle of the neutrino trajectory at the surface of the earth . for large zenith angles ,
the curvature of the earth is significant , and it is necessary to use effective values of @xmath51 that represent the convolution of the locations of neutrino production with the local zenith angle as it decreases moving upward along the trajectory . we treat @xmath51 as a free parameter in fitting eqs .
[ nupi ] and [ numu ] to the monte carlo results .
the values are included in table [ tab2 ] .
lccllc @xmath52 & @xmath53 & @xmath54(eq .
[ mode ] ) & @xmath55 & @xmath56 & @xmath57 + & ( km ) & ( km ) & & & + & & & & & + 1.00 & 13.8 & 14.0 & 1.00 & 1.00 & 0.69 + 0.75 & 21.6 & 21.2 & 0.75 & 0.75 & 0.71 + 0.50 & 38.4 & 37.0 & 0.50 & 0.50 & 0.77 + 0.25 & 88.4 & 87.5 & 0.26 & 0.26 & 0.83 + 0.15 & 155 . &
& 0.164 & 0.168 & 1.00 + 0.05 & 382 .
& 0.084 & 0.087 & 1.86 + & & & & & + [ tab2 ] up to @xmath45 = 0.25 the agreement between the monte carlo calculation and the analytic estimate is quite good .
note that for nearly horizontal neutrinos the height distribution from the monte carlo calculation is artificially narrow and irregular .
the atmospheric model used does not treat exactly the atmospheric densities at vertical depths of less than few g/@xmath40 .
this intruduces a sharp cutoff in the height distribution for strongly inclined showers and also decreases the width of the height distribution .
the height distribution of @xmath0 from @xmath15 decay has a similar shape with much lower normalization , because only @xmath24 have decay mode with @xmath0 s ( @xmath58 . to estimate the height of production for neutrinos from decay of muons is more complicated because of the competition between decay and energy loss for muons in the multi - gev energy range .
one starts from the distribution of production points for muons , which is similar to eq .
[ production ] with different coefficients .
the resulting approximate expression @xcite for the distribution of production distances ( differential in the energy of the parent muons as well as the slant height of production ) is @xmath59^{-p}\over[e_\mu+\alpha(x - y)]^{\gamma+1}},\ ] ] where @xmath60 is the muon lifetime , @xmath61 and @xmath62.\ ] ] at high altitude muon energy loss ( @xmath63 ) can be neglected and the expression [ numu ] is proportional to slant depth @xmath41 given by eq .
[ atmos ] .
this expression gives a good account of the high - altitude exponential falloff of the neutrinos from muon decay .
an approximation that is adequate for fitting the distribution for all distances ( integrated over neutrino energy ) is @xmath64^{-(p+\gamma+1 ) } , \ ] ] where @xmath57 is an overall normalization factor used to fit the monte carlo results ( see table [ tab2 ] ) .
4 shows the height of production distributions for muon neutrinos of energy above 1 gev from muon decay .
the lines are calculated according to eq .
[ numuapprox ] with values of @xmath65 as given in table [ tab2 ] . to obtain the fits shown in fig . 4 the approximations of eq . [ numuapprox ]
have also been renormalized as indicated in table [ tab2 ] . at high altitude
the height of production distributions have the same shape as the ones from neutrinos from @xmath15 decay , shifted to lower altitudes by one muon decay length ( 6.24 km for 1 gev muons ) . at lower altitude
the shapes are quite different .
the production height for neutrinos from muon decay extend to much lower altitude because of the slow attenuation of the parent muon flux , an effect which becomes more pronounced as the energy increases .
it is interesting to observe that at high zenith angles the yield of neutrinos from ( daughter ) muon decay exceeds the yield of neutrinos from the decay of the parent pions and kaons .
the reason is that muon neutrinos from muon decay in flight have a spectrum extending almost to @xmath66 ( where @xmath67 ) , while the neutrinos from @xmath68 decay can only reach @xmath69 = @xmath70 .
the corresponding @xmath71factors @xmath72 and @xmath73 are 0.133 and 0.087 respectively , including the effect of muon polarization in pion decay .
the result is that for large zenith angles , when almost all muons decay the @xmath1 yield from muon decay becomes slightly larger than that from @xmath15 decay ( @xmath74 for @xmath45 = 0.05 ) .
fig . 5 compares the distributions of distance to production for @xmath1 from muon decay with @xmath75 above 1 , 10 , and 100 gev .
at high neutrino energy the muon decay length becomes comperable or larger than the total dimension of the atmosphere .
the neutrino height of production then becomes constant deep in the atmosphere .
the height distribution for @xmath0 from muon decay is analogous to that of @xmath1 .
the only difference is the slightly lower normalization , which reflects the ratio @xmath76 = 0.133/0.129 = 1.03 ( including muon polarization ) . in table
[ tab3 ] we show the average height of production and the contributions of @xmath15 and muon decays for neutrinos in the three energy bins ( @xmath77 gev ; @xmath78 gev ; @xmath79 gev ) which roughly correspond to contained neutrino events , semicontained events and stopping neutrino induced muons and throughgoing neutrino induced muons .
for each angle and neutrino flavor table [ tab3 ] first gives the average height of production ( slant depth ) in km and the width of the height of production distribution .
then it gives the contribution ( in % ) of @xmath15 decay @xmath80 and the corresponding @xmath81 and @xmath82 , then the same quantities ( @xmath83 for neutrinos from muon decay .
the calculation was done with the geomagnetic cutoffs of kamioka , except for the three lines ( @xmath52 = 1.00 , 0.75 and 0.50 , for the lowest energy bin ) that are also calculated for the high geomagnetic latitude of sno @xcite .
the numbers for high geomagnetic latitude are slightly higher ( 2 10 % ) for both neutrino sources because of the contribution of low energy protons .
this difference becomes negligible at higher angles . .
production height ( slant distance , km ) of neutrinos for six values of @xmath84 and three neutrino energy ranges .
the calculation is for the geomagnetic location of kamioka with three lines for the lowest energy range calculated for sudbury , canada .
[ cols= " > , > , > , > , > , > , > , > , > , > , > , > , > , > , > , > , > " , ] there are two obvious trends in the numbers in table [ tab3 ] .
the height of production for neutrinos from @xmath15 decay grows slightly with the neutrino energy because higher energy mesons preferentially decay ( rather than interact ) in the tenuous atmosphere at high altitude .
neutrinos from muon decay , on the other hand , are generated at lower altitude at high energy because of the increasing muon decay length .
this second feature is much stronger because of the proportionality of decay length and muon energy .
the average heights of production also reflect the relative yields of the two neutrino sources . for low energy @xmath85 , for example , the contribution of @xmath86 is small , so the average height of production is dominated by muon decay . at higher energy
the relative contribution of @xmath86 grows , especially at directions close to the vertical , and @xmath87 becomes intermediate between those of the two processes with correspondingly larger width .
generally the contribution from muon decay increases significantly with the zenith angle since even 20 gev muons easily decay in cascades developing in nearly horizontal direction .
we have calculated the distribution of pathlengths of atmospheric neutrinos for a range of angles and energies relevant for current searches for neutrino oscillations with atmospheric neutrinos .
accounting correctly for the pathlenght will be particularly important for neutrinos near the horizontal direction where the pathlength through the atmosphere of neutrinos from above the horizon is of the same order of magnitude as the pathlength through the earth of neutrinos from below the horizon .
we have also given simple approximations that may be useful in interpolating the tables and adapting the results for different energy ranges and directions .
the influence of the geomagnetic effects on the calculated height of neutrino production distributions is not very strong .
the difference in the average production heights for neutrinos detected at kamioka and sno is of order several per cent in directions relatively close to the vertical .
this difference diminishes with angle and becomes totally negligible for upward going neutrinos , where the geomagnetic cutoffs becomes approximately equal , being averaged over the geomagnetic fields of the opposite hemisphere .
the authors express their gratitude to w. gajewski , j.g . learned , h. sobel , and y. suzuki for their interest in the height of production problem that inspired us the complete this research .
this work is supported in part by the u.s .
department of energy under de - fg02 - 91er40626.a007 .
z. conner , hightlight talk at the 25th international cosmic ray conf .
( durban , south africa , 1997 ) to be published ; j.g . learned ( superkamiokande collaboration ) , _ proc .
25th international cosmic ray conference _ , eds .
potgieter , b.c .
raubenheimer and d.j .
van der walt , * 7 * , 73 ( 1997 ) .
r. becker - szendy _ et al .
_ , ( imb collaboration ) , _ phys . rev . _ * d46 * ( 1992 ) 3720 .
see also d. casper _ et al .
. rev . letters _ ,
* 66 * ( 1991 ) 2561 .
et al . _ , ( kam - ii collaboration ) , _ phys .
letters _ , * b280 * ( 1992 ) 146 and y. fukuda _ et al .
_ , _ phys .
letters_,*b335 * ( 1994 ) 237 .
et al . _ , ( soudan 2 collaboration ) , _ proc .
25th international cosmic ray conference _ , eds .
potgieter , b.c .
raubenheimer and d.j .
van der walt , * 7 * , 77 ( 1997 ) . , by felix boehm & peter vogel ( cambridge university press , 1992 ) .
fogli and e. lisi , _ phys .
_ , d*52 * ( 1995 ) 2775 .
bernstein & s. parke , _ phys .
d _ , * 44 * , 2059 ( 1991 )
. v. agrawal _ et al .
_ , _ phys .
_ d*53 * ( 1996 ) , 1314 .
_ , _ phys .
_ d*54 * ( 1996 ) 5578 .
giles barr , t.k .
gaisser & todor stanev , phys . rev .
d39 ( 1989 ) 3532 .
h. lee & s.a .
bludman , _ phys .
_ d*37 * ( 1988 ) 122 .
gaisser & todor stanev , proc .
24th icrc ( rome ) vol . 1 ,
webber , r.l .
golden and s.a .
stephens , in _ proc .
_ ( moscow ) * 1 * , 325 ( 1987 ) .
ap . j. _ * 378 * , 763 ( 1991 ) .
j. engel _ et al .
rev . _ * d46 * , 5013 ( 1992 ) . m. thunmann , g. ingelman & p. gondolo , _ astropart . physics _ * 5 * ( 1996 ) 309 .
volkova , w. fulgione , p. galeotti & o. saavedra , _
nuovo cimento _ c*10 * ( 1987 ) 465 .
r. bellotti _ et al .
_ d*53 * , 35 ( 1996 ) .
marco circella , carlo de marzo , t.k .
gaisser & todor stanev , proc .
. cosmic ray conf .
( durban ) * 7 * , 117 ( 1997 ) .
, by thomas k. gaisser , ( cambridge university press , 1990 ) .
paolo lipari , _ astroparticle physics _ * 1 * , 195 ( 1993 ) .
moorhead ( for the sno collaroration ) , _ nucl .
( proc . suppl . ) * 48 * 378 ( 1995 ) . | we present the distribution of the production heights of atmsopheric neutrinos as a function of zenith angle and neutrino energy .
the distributions can be used as the input for evaluation of neutrino propagation under various hypotheses for neutrino flavor oscillations . |
it is estimated that 30% or more of the light emitted from stars in the universe is absorbed and re - emitted by dust in the infrared ( ir ) . the cosmic ir background , therefore , records the cumulative ir emission from galaxies at all redshifts and provides an important constraint on the global history of star formation .
this emission is affected by dust evolution in galaxies , in particular at high redshift ( @xmath12 ) , where the presence of dust can influence the galaxy number counts and the evolution of quasars ( e.g. obscuring star - forming galaxies or active galactic nuclei ; smail , ivison & blain 1997 ; hughes et al .
1998 , maiolino et al .
2003a ) .
evidence for the presence of dust at high redshifts comes from observations of damped ly@xmath13 systems ( pettini et al . 1994 ; prochaska & wolfe 2002 ; ledoux , bergeron & petitjean 2002 ) and from the detection of dust thermal emission from high redshift quasars ( qsos ) selected from the sloan digital sky survey ( sdss ) re - observed at millimetre wavelengths ( omont et al . 2001 ; carilli et al .
2001 ; bertoldi & cox 2002 ) . the inferred far - ir ( fir ) luminosities of samples of @xmath14 quasars are consistent with thermal emission from warm dust ( @xmath15 k ) , with dust masses in excess of @xmath16 ( bertoldi et al . 2003 ; robson et al . 2004 ; beelen et al .
2006 ) . despite the fact that these high redshift quasars are rare objects , hardly representative of the dominant star forming galaxy population ,
the fir and sub - mm observations indicate that early star formation leads to rapid enrichment of their host galaxies interstellar medium ( ism ) with metals and dust .
this is consistent with the super - solar metallicities inferred from the optical emission - line ratios for many of these systems ( pentericci et al .
2002 ; freudling et al .
2003 ; maiolino et al .
2003b ) . in present - day galaxies , the major sources of interstellar dust
are believed to be low and intermediate - mass evolved stars during the asymptotic giant branch ( agb ) phase .
since these stars take a long time to evolve to their dust producing stages , it is commonly believed that they can not be responsible for all of the dust seen at high redshifts .
in fact , agb stellar lifetimes ( @xmath17 to @xmath18 yr ) are comparable to the age of universe at redshift @xmath19 ( morgan & edmunds 2003 ; marchenko 2006 ) . as a result ,
if the observed dust is produced by stellar processes , supernova ( sn ) ejecta appear to be the only viable sites of grain condensation fast enough to explain these large dust masses .
this scenario has been tested through observations of the reddened quasar sdssj1048 + 46 at @xmath20 ( maiolino et al .
2004 ) and of the spectral energy distribution ( sed ) of the grb 050904 afterglow ( stratta et al . 2007 ) . in both sources ,
the inferred dust extinction curve is different with respect to any of the extinction curves observed at low z ( hopkins et al .
2004 ) , and it shows a very good agreement with the extinction curve predicted for dust formed in sn ejecta .
this is an indication that the properties of dust evolve beyond @xmath21 .
theoretical models , based on classical nucleation theory , predict that a few hundred days after the explosions silicate and carbon grains can form in expanding sn ejecta , with condensation efficiencies in the range @xmath22 .
this implies @xmath23 of dust per sn for stellar progenitor masses in the range of core - collapse sn ( @xmath24 ) ( kozasa et al .
1991 ; todini & ferrara 2001 ; clayton , deneault & meyer 2001 ; nozawa et al .
2003 ) .
there is now clear observational evidence for dust formation in core - collapse supernovae , but the quantity of dust formed within the ejecta is still a subject of debate . the observed ir emission for a limited number of sne , such as 1987a ( wooden et al .
1993 ) , 1999em ( elmhamdi et al . 2003 ) and 2003gd ( sugerman et al . 2006 ; meikle et al .
2007 ) implies dust masses which are generally smaller than @xmath25 , corresponding to condensation efficiencies which are at least two orders of magnitude smaller than what theory predicts . observations of young galactic sn remnants by the _ spitzer _
telescope provide unambiguous evidence of dust formation ( hines et al .
the inferred total mass of freshly formed dust is @xmath26 for cas a ( rho et al . 2008 ) , and @xmath27 for 1e 0102.2 - 7219 ( rho et al .
while masses up to @xmath28 have been estimated from fir and submm observations ( dunne et al . 2003 ) , the interpretation of these data is complicated by strong foreground contamination ( krause et al . 2004 ; wilson & batrla 2005 )
. the apparent discrepancies between observational estimates and theoretical models can be accomodated taking into account the partial destruction of newly synthesized dust in the reverse shock of the sn ( bianchi & schneider 2007 ) .
the evolution of dust in the early universe is still poorly understood .
models developed so far show that the current observational constraints on dust production in sne are still consistent with the scenario where most of the dust in qso at @xmath29 is produced by sne .
however these estimates either neglect dust destruction by interstellar shocks ( maiolino et al .
2006 ) or make extreme assumptions on sn dust condensation factors ( dwek , galliano & jones 2007 ) .
it is clear that our ability to interpret the observed properties of high redshift qsos depends on a detailed understanding of the star formation history and the history of metal and dust pollution in their host galaxies .
this , in turn , requires an adequate description of metal and dust stellar yields . in this work
we study the evolution of dust produced by low , intermediate and high - mass stars as a function of time and metallicity , taking into account the stellar evolutionary timescales , i.e. without adopting an instantaneous recycling approximation .
the aim of this study is to explore the relative importance of the two main stellar sources , agb stars and sne , in the evolution of dust at high redshift for galaxies with different star formation histories ( sfh ) and stellar initial mass function ( imf ) .
the paper is organized as follows . in section 2 , we describe the adopted grid of dust yields for agb stars and sne with different initial metallicities . in section 3
we discuss the time evolution of the predicted dust mass for two different sfh . in this section
we also explore the dependence of the results on the adopted imf and the stellar mass - lifetime relation . in section 4
we present a chemical evolution model which includes dust destruction by interstellar sn shocks and , in section 5 , we apply it to the qso sdss j1148 + 5251 at @xmath7 , using the star formation history of the host galaxy predicted by the numerical simulation of li et al .
finally , in section 6 we discuss and summarize the results of our work . in
what follows , we assume a @xmath30cdm cosmology with @xmath31 , @xmath32 , @xmath33 , and @xmath34 km / s / mpc .
the age of the universe at a redshift @xmath7 is about 840 myr .
dust production in the ism is extremely inefficient as the timescales for dust formation are longer than the destruction timescales ( tielens 1998 ) .
thus , formation sites during stellar evolution are required to explain the vast amounts of dust observed in the universe .
most of the theoretical models for dust formation are based on classical nucleation theory ( feder et al .
1966 ) : the formation of solid materials from the gas phase can occur only from a vapour in a supersaturated state .
condensation of a given species occurs when its partial pressure in the gas exceeds its vapour pressure in condensed phase ( whittet 1992 ) , with subsequent particle growth by random encounters leading to cluster formation .
chemical reactions will further assist cluster production .
the rate of grain growth therefore depends on the temperature , t@xmath35 , and pressure , p@xmath35 , of the gas and the condensation temperature of a species , t@xmath36 .
the optimum conditions for nucleation are thought to occur in the range @xmath37p@xmath35(pa)@xmath38 and t@xmath39 k ( salpeter 1974 ) .
this gives very stringent restrictions on where nucleation can occur .
stars seen to produce dust , only satisfy the conditions needed for the nucleation processes at the very latest stages of their evolution .
the main stellar sites of dust grain condensation are the cool , dense atmospheres of agb stars and the expanding sn ejecta .
the type of dust formed depends on the relative abundance of carbon and oxygen at the onset of nucleation . in this work
, we use the results obtained by bianchi & schneider ( 2007 ) to estimate the mass and properties of dust formed in type ii sn ejecta and those presented by zhukovska , gail & trieloff ( 2008 ) to take into account grain condensation in the atmospheres of agb stars . all stars with initial masses @xmath40 on the main sequence
evolve through the agb phase of stellar evolution shortly before they become white dwarfs . in these stars ,
the sites of dust production are their cool , dense atmospheres during the thermally pulsating ( tp ) agb phase . in this evolutionary stage , stars loose mass at an extremely high rate by a stellar wind and develop optically thick circumstellar dust shells by condensation of solid particulates in the outflowing gas .
depending on the relative abundances of carbon and oxygen ( c / o ) of the ejected matter in which grains are formed , different dust mixtures are produced .
ferrarotti & gail ( 2006 ) provide models which consistently describe the dependence of dust production by agb stars on stellar initial mass and metallicity .
they calculate the dust production rate of the dominant dust species , namely silicates , carbon , silicon carbide ( sic ) and iron , which are formed from the most abundant elements .
they use agb stellar yields from van den hoek & groenewegen ( 1997 ) and karakas et al .
( 2003 ) . on the basis of this model calculations , zhukovska et al .
( 2008 ) propose refined grids of the ejected mass of different dust species , for different star masses in the range @xmath41 and at various ( sub - solar and solar ) metallicities .
these models predict that silicate dust is produced by stars with initial masses @xmath42 and @xmath43 , which during the agb phase are oxigen - rich .
these stars are spectroscopically identified as m ( characterized by c / o@xmath44 ) and s type ( c / o@xmath45 ) . on the other hand , carbon and sic dust production is dominated by stars with initial mass @xmath46 , during the carbon - rich phase of evolution on the agb ( c stars , c / o @xmath47 ) .
iron dust should be formed in all m , s and c type stars .
however , only some hints of its existence have been found up to now ( e.g. kemper et al . 2002 ) . to ease the comparison with the dust mass produced by sne , in fig .
1 we plot the mass of carbon , silicate and _ other _ dust species produced by agb stars as a function of the initial star mass and for three different metallicities , @xmath48 . here we indicate as _ other _ dust the sum of iron and sic grains .
it is clear from the figure that agb stars can produce @xmath49 at all metallicities while comparable amounts of silicate can only be produced when @xmath50 .
in fact , silicate grain condensation requires the presence of a sufficient amount of si , o , mg and fe in the stellar outflow .
these metal species can not be produced by the star and have to be collected by the pre - enriched ism prior to star formation .
stars with masses @xmath51 end their life exploding as core - collapse sne . main sites of dust formation are their expanding ejecta , a few hundred days after the explosions .
dust formed in stellar winds during the pre - supernova phase ( i.e. by red supergiant stars or wolf - rayet stars ) is almost entirely destroyed by the subsequent shock wave ( see discussion in zhukovska et al . 2008 and references therein ) .
most of the available models proposed to investigate the process of dust formation in expanding sn ejecta are based on classical nucleation theory and grain growth .
the onset of grain formation depends on the temperature structure in the sn ejecta , whereas the grain composition mainly reflects its chemical composition , which depends on the nucleosynthesis occurring during stellar lifetime ( i.e. on the progenitor mass ) and explosion mechanism .
thus , models of dust formation in sn ejecta are based on specific prescriptions for the chemical composition and thermodynamics of the expanding gas .
for the present analysis , we use the results of the model initially developed by todini & ferrara ( 2001 ) but recently revisited by bianchi & schneider ( 2007 ) in order to follow the evolution of newly condensed grains from the time of formation to their survival through the passage of the reverse shock in the sn remnant . in these works ,
the ejecta are taken to have a uniform composition and density , with initial temperature and density chosen to match the observations of sn1987a .
the initial composition depends on the metallicity and mass of the progenitor star while the dynamic is given by the mass of the ejecta and the kinetic energy of the explosion : the initial sn models were taken from woosley & weaver ( 1995 ) .
therefore , we consider dust formed in the ejected matter by sne with initial progenitor star masses in the range @xmath52 and metallicities @xmath53 .
bianchi & schneider ( 2007 ) derive the fraction of dust which survives the passage of the reverse shock , as a function of progenitor star mass and metallicity
. dust destruction due to the passage of the reverse shock depends on the density of the surrounding ism .
the denser is the gas in the ism , the higher is the fraction of the destroyed dust mass .
we refer the reader to the original paper for more details on the model . in what follows
, we adopt the results of the reference model by bianchi & schneider ( 2007 ) , i.e. the model which starts grain condensations from seed clusters made of @xmath54 monomers , assumes a sticking coefficient @xmath55 ( all the gas particles colliding on a grain will stick to it ) and an ism density of @xmath56 g @xmath57 . in this model ,
about @xmath58 of dust are formed in the ejecta but only 7% of the initial dust mass survives the passage of the reverse shock .
note , however , that these numbers are sensitive to variations on the assumed values of @xmath59 , @xmath13 and @xmath60 .
for a larger @xmath59 and/or a smaller @xmath13 , dust grains are characterized by smaller radii , which are more easily destroyed .
if the sn explodes in a denser ( @xmath61 g @xmath57 ) ism , only about 2% of the dust mass survives .
conversely , for a lower density ism ( @xmath62 g @xmath57 ) a larger fraction , about 20% , is left . in fig .
2 we present the final mass of carbon , silicate and other dust species ( essentially magnetite , fe@xmath63o@xmath64 , and corundum , al@xmath65o@xmath63 ) for the fiducial model , as a function of the initial stellar mass and metallicity , as given by bianchi & schneider ( 2007 ) .
carbon dust condensation efficiency does not vary much with metallicity , with @xmath66 , while silicates production appears to be inefficient at low metallicity .
it is worth noting that the model by bianchi & schneider ( 2007 ) shows that sne are expected to produce mainly carbon , magnetite and corundum grains .
this is at odd with the models developed by nozawa et al .
( 2003 , 2007 ) , in which the composition is dominated by silicate and iron grains .
the differences are likely to be due to different initial sn models and their assumption of complete formation of sio and co molecules .
while the first molecule is necessary to study the formation of si - bearing grains , the second may be a sink for carbon atoms that otherwise would accrete on grains ( bianchi & schneider 2007 ) . in both models , though , a similar amount of dust is formed and survives the reverse shock .
a thorough comparison between different theoretical models is deferred to future studies .
finally , it should be mentioned that in the present model we do not consider the possible contribution to dust production by pair - instability and type ia sne .
the former class is predicted to be the end - product of the evolution of very massive population iii stars with masses in the range @xmath67 and might provide an important source of dust at high redshift ( schneider , ferrara & salvaterra 2004 ; nozawa et al .
2003 ) . however , our aim here is to compare the relative contribution of the two main stellar sources of dust assuming a conventional stellar initial mass function with an upper mass cut - off of @xmath68 . for the latter class , there are no available theoretical models for dust production and destruction , although clayton et al .
( 1997 ) have presented arguments in favor of dust production in type ia sne , with important consequences for the composition of iron dust in the interstellar medium ( dwek 1998 ) .
type ia sne might be relevant for dust evolution at high redshift given the recent observational evidence for a `` prompt '' component exploding in a time scale of the order of 100 myr after the stellar birth ( mannucci , della valle & panagia 2006 ) and that a sizeable fraction of stars with masses in the range @xmath69 is predicted to explode as type ia sne ( maoz 2008 ) .
however , adding the contribution of type ia sne to ( iron ) dust production is beyond the scope of the present analysis .
as a first step of the analysis , we compute the time evolution of the total dust mass produced , @xmath70 , considering the contribution of agb stars and type ii sne discussed above .
the aim of this study is to investigate the relative importance of these two formation channels and the characteristic timescales at which one dominates over the other .
we take into account that stars of different masses evolve on different timescales . therefore
, the time evolution of the total dust mass can be expressed as : @xmath71 where @xmath72 is the progenitor stellar mass , @xmath73 is the dust mass produced by a star of mass @xmath72 , @xmath74 is the stellar initial mass function ( imf ) , @xmath75 is the star formation rate , and @xmath76 is the stellar lifetime of a star of mass @xmath72 .
the lower limit @xmath77 is the minimum progenitor mass contributing to the dust production at a time @xmath78 , i.e. the mass corresponding to a stellar lifetime @xmath79 .
note that we have no prescriptions for dust production by stars in the mass range @xmath80 m@xmath81 , which evolve as agb and finally explode as core - collapse sne ( ii - p sne , smartt et al .
dust yields for stars in this mass range ( and for different metallicity ) are not yet available in the literature . therefore we extrapolate from the dust mass produced by the largest available agb model ( 7 @xmath82 ) and by the smallest sn progenitor ( 12 @xmath82 ) to account for metal and dust production in this mass range .
however , the contribution to the total mass of dust from stars in this mass range is less than 5% ) . to investigate the dependence on the star formation history , we make two extreme hypotheses : ( i ) a burst occurring at @xmath83 , where all stars are born at the same epoch , in an instantaneous episode of star formation ; ( ii ) a constant @xmath75 , where stars form continuously and at a constant rate for 10 gyr .
it is clear that the relative contribution of agb stars and sne depends on the shape of the stellar imf and on the adopted stellar lifetimes . in our reference model
, we assume that stars form with masses in the range @xmath84 m@xmath81 according to a larson imf ( larson 1998 ) , which follows a salpeter - like power law at the upper end but flattens below a characteristic stellar mass : @xmath85 where @xmath86 , @xmath87 and we normalize the integral of @xmath88 in the mass range @xmath89 m@xmath81 to unity .
the lifetimes of stars of different mass and metallicity are computed according to the following simple parametric form proposed by raiteri et al .
( 1996 ) , @xmath90 where @xmath91 is expressed in yr , @xmath72 in solar units and the coefficients are : @xmath92 although this relation has been tested to reproduce the stellar evolutionary data in the metallicity range @xmath93 , we apply it also to stars with lower metallicities . in fig .
[ fig : mdust_ref ] we show the time evolution of the dust mass normalized to the final mass of stars for a burst - like ( right panels ) and a constant star formation history ( left panels ) and assuming that all stars form at the same initial metallicity of @xmath94 ( upper panels ) , we have assumed that the agb dust mass for lower metallicities is the same as at @xmath95 . for both agb and sn dust , dust mass values at metallicities not available in the grids
are obtained through interpolation . ]
( upper panels ) and @xmath96 ( lower panels ) .
the curves represent the mass of dust , summed over all the components ( carbon , silicates and other dust species ) produced by sne ( dashed lines ) and agb stars ( dot - dashed lines ) . at the beginning of the evolution ,
dust production is dominated by sne ; however , for @xmath97 myr the contribution of agb stars rapidly increases and becomes dominant in a characteristic time which depends both on the assumed star formation history and stellar metallicity , ranging between 150 myr ( for a burst at @xmath98 ) and 500 myr ( for a constant star formation rate at @xmath99 ) .
these results suggest that agb stars might give a non - negligible contribution to dust production at early cosmic epochs and that their role as dust factories at @xmath100 might have been over - looked .
gyr in our reference model ( see text ) .
all the stars have an initial metallicity of @xmath94 ( upper panels ) and @xmath96 ( lower panels ) .
the rigth panels show the results obtained assuming that star formation occurs in a burst at @xmath83 whereas the left panels are computed for a constant star formation rate lasting for 10 gyr.,width=340 ] we have also investigated how these results depend on different model assumptions , such as the stellar lifetime relation and imf .
it can be shown that the timescales at which agb stars start to be the dominant dust sources change by less than 20% if we adopt the stellar lifetimes proposed by padovani & matteucci ( 1993 ) , which predicts slightly longer lifetimes for stars of intermediate mass and high metallicity compared to our reference model .
the largest uncertainty is introduced by the shape of the stellar imf , which can be quantified through variations of the stellar characteristic mass @xmath101 in eq .
[ imf ] . in fig .
[ fig : mdust_mch ] we show the time evolution of dust mass produced by sne and agb stars assuming a constant stellar metallicity of @xmath96 and three different values for the characteristic mass @xmath102 and @xmath103 ( from bottom to top ) . as in fig . 3 ,
stars are formed at a constant rate for @xmath104 gyr ( left panels ) and in a burst at @xmath83 ( right panels ) and the resulting dust mass is normalized to the total stellar mass formed . for comparison , in fig .
[ fig : mdust_mch ] the corresponding reference models are also shown ( light curves ) .
as it can be inferred from the figure , the evolution is almost identical to the reference model for @xmath105 ; for @xmath106 , the mass of dust contributed by sne is increased ( as a consequence of the larger sne rate ) and agb stars produce a comparable amount of dust on timescales @xmath107 gyr . finally , when @xmath108 the agb dust component is always sub - dominant , as a result of the larger sn rate and of the smaller number of stars formed with intermediate masses .
thus , it is clear that variations of the stellar characteristic mass over cosmic history due to the smaller gas metallicity and/or to the larger cosmic microwave background temperature ( see , e.g. , omukai et al . 2005 ) might significantly affect the origin and properties of dust in the high redshift universe . but
assuming that all stars form with a metallicity @xmath96 and a larson imf with characteristic masses @xmath109 , @xmath110 , and @xmath103 ( from bottom to top ) ; left panels are computed assuming a constant star formation rate for 10 gyr ; in right panels all stars are formed in a burst at @xmath83 .
the mass of dust is normalized to the final stellar mass formed . for comparison , in each panel
we also show the results of the fiducial model of fig .
[ fig : mdust_ref ] with thin lines.,width=321 ]
in this section , we describe a simple chemical evolution model with dust which allows to follow the time evolution of the mass of gas , stars , metals in specific environments , such as high redshift galaxies and quasars , given the appropriate star formation history .
we then apply this model to a `` template '' qso host galaxy at high redshift . in this model
, we make the simplifying assumptions that ( i ) there are no mass exchanges with the surrounding intergalactic medium ( igm ) ( `` close - box '' model ) , and that ( ii ) dust produced by agb stars and sne does not grow in molecular clouds ( mc ) by accretion of metals .
indeed , dust grains formed in stellar outflows can serve as growth centers for accretion of icy mantels in dense mc ( e.g. draine 1990 ) .
however , grain mantles are amorphous and heterogeneous , thus mc - grown dust has different properties from dust produced by stars and must be treated as a separate component ( see e.g. zhukovska et al .
2008 and references therein ) .
since we are primarily interested in the study of the relative contribution of the two main stellar sources , we do not follow the evolution of this component .
the model consists of a set of differential equations regulating the time evolution of the different components : @xmath111 where @xmath112 and @xmath113 are the total masses of stars and dust , @xmath75 is the star formation rate . since a fraction of dust
can be destroyed by interstellar sn shocks and returns metals back into the gas phase , we avoid complications indicating with @xmath114 the total mass in the interstellar medium ( gas and dust ) and with @xmath115 the total mass of metals ( diffused in the gas phase and condensed into dust grains ) . at any given time , the mass of _ gas _ can be computed as the difference between @xmath114 and @xmath113 , and the mass of _ gas phase _ metals as the difference between @xmath115 and @xmath113 ( with proper stoichiometric coefficients when we are interested in a particular gas phase element ) .
the quantities @xmath116 and @xmath117 are the total metal and dust abundances in the ism .
the terms @xmath118 , @xmath119 and @xmath120 represent the rates at which mass ( gas and dust ) , heavy elements ( gas - phase and condensed ) and dust are returned to the ism .
these terms are given by : @xmath121 where @xmath72 is the initial mass of the progenitor star , and the lower limit @xmath77 is the mass with a lifetime @xmath122 .
the total mass of metals ( pre - existing and newly synthesized ) produced by a star of initial mass @xmath72 and metallicity @xmath123 , @xmath124 , and the mass of the stellar remnant @xmath125 , are taken from van den hoeck & groenewegen ( 1997 ) for agb stars with initial metallicities @xmath126 and masses @xmath127 and from woosley & weaver ( 1995 ) for sne with initial masses @xmath24 metallicities @xmath128 z@xmath81 .
the grid of dust masses , @xmath129 , is the same as described in section 2 and shown in figs . 1 and 2 .
note that integrating eq .
[ eq : ejerate ] over time we obtain the mass of dust as given in eq .
the last term of eq .
[ eq : dust ] is the dust destruction rate in the ism by thermal sputtering in high - velocity ( @xmath130 km s@xmath131 ) sn shocks .
the destructive effect of the sn shock waves on dust grains depends on the dust species and on the destruction timescale , @xmath132 , i.e. the lifetime of the dust grains against destruction by sn remnant ( jones , tielens & hollenbach 1996 ) .
following dwek , galliano & jones ( 2007 ) we assume that @xmath133 where @xmath134 is the effective ism mass that is completely cleared of dust by a single sn remnant ( i.e. a measure of the grain destruction efficiency ) , which depends on the dust species , and @xmath135 is the sn rate , @xmath136 grain destruction efficiencies by radiative shocks have been modeled by jones et al .
( 1994 ) , and typical grain lifetimes of 0.6 and 0.4 gyr have been estimated , respectively , for carbonaceous and silicate grains in the milky way ( jones et al .
dwek , galliano & jones ( 2007 ) investigate the dependence of grain destruction efficiency , @xmath134 , on the ism density , both for silicate and carbon dust .
following their study , for a uniform ism with density @xmath56 g @xmath57 , we assume @xmath137 for silicates and @xmath138 for carbon and other dust species .
finally , in what follows we always consider our fiducial model , i.e. a larson imf with a characteristic mass of 0.35 @xmath82 and stellar lifetimes from raiteri et al .
( 1996 ) as described in section 3 .
we apply the model described above to the most distant quasar , sdss j1148 + 5251 , discovered at redshift @xmath139 ( fan et al . 2003 ) .
it is a very luminous quasar powered by a super massive black hole ( smbh ) with mass @xmath140 accreting close to its eddington limit ( willott et al . 2003 ) .
the fir luminosity is @xmath141 , suggesting a mass of dust in excess of @xmath142 ( bertoldi et al .
2003 ; robson et al . 2004
; carilli et al . 2004 ; beelen et al .
if all the fir emission is assumed to come from young star , the derived @xmath75 is about @xmath143 yr@xmath131 ( bertoldi et al . 2003 ) .
such a high @xmath75 is also supported by the first detection of the carbon [ cii ] line at 158 @xmath144 m ( maiolino et al .
2005 ) . alternatively ,
if the active galactic nucleus ( agn ) contributes to dust heating , the inferred @xmath75 would be much lower , in agreement with what implied by a simple application of the schmidt - kennicutt law adopting a total gas mass of @xmath145 ( walter et al .
2004 ) , which yields a value of @xmath146 yr@xmath131 ( dwek , galliano & jones 2007 ; li et al .
2008 ) .
the interpretation of the observed properties of j1148 + 5251 and of its host galaxy is therefore contradictory .
our aim here is to explore the relative importance of sne and agb in the build - up of the dust mass inferred for this system .
we adopt the star formation history of the host galaxy of j1148 + 5251 predicted by a recent numerical simulation by li et al .
we then apply our chemical evolution model with dust and discuss the dependence of the results on the adopted star formation history .
note that this approach is not entirely self - consistent given that , as we will see below , the simulated system is not a closed - box .
a more accurate study of the chemical properties of the host galaxy of j1148 + 5251 implementing dust evolution in the semi - analytical merger - tree code gamete ( salvadori , ferrara & schneider 2008 ) is currently underway .
the numerical simulation by li et al . ( 2007 ) follows the hierarchical assembly of the sdss j1148 + 5251 halo in a @xmath147cdm cosmology .
their model accounts for the quasar activity , the host galaxy properties and includes a self - regulated growth of the smbh .
the quasar host galaxy is found to form through seven major mergers of gas rich starburst progenitors between @xmath148 and @xmath149 .
intense starbursts are triggered by the gravitational interactions between the merging galaxies and rapid black hole accretion is fuelled by the highly concentrated gas ( see li et al .
2007 for details ) .
this model is able to reproduce most of the observed properties of sdss j1148 + 5251 , such as the black hole mass , the quasar luminosity and metallicity . for convenience , fig .
[ fig : lisfr ] shows the time evolution of the star formation rate found by li et al .
the star formation rate ranges between @xmath150 yr@xmath131 , both at the onset of the interactions between the progenitor galaxies ( at about 0.29 gyr , corresponding to a redshift @xmath151 ) and at the time of the final coalescence ( 0.8 gyr , corresponding to @xmath152 ) , to values as large as @xmath153 yr@xmath131 at the final major merger , represented by the peak at 0.6 gyr ( @xmath154 ) .
the horizontal line shows the average value of the star formation rate over the lifetime of the simulated qso host galaxy , found to be @xmath155 yr@xmath131 .
note that the simulation predicts a rate of about 100 m@xmath81 yr@xmath131 at 840 myr ( corresponding to redshift @xmath156 ) , an order of magnitude smaller than that inferred by fir observations .
therefore , li et al .
( 2008 ) suggests that the fir luminosity of the qso is not dominated by young stars but instead by the agn ( for over 80% ) .
finally , the stellar mass of the quasar host at @xmath157 is predicted to be @xmath158 , in agreement with what would be required by the observed smbh mass if the local black - hole - bulge relation were to hold at these high - redshifts .
however , this result can not be reconciled with observations of molecular gas : while the dynamical mass estimated from co observations within 2.5 kpc from the qso center is @xmath159 , a conservative estimate of the contribution of the predicted stellar bulge within this region is a few @xmath160 , about an order of magnitude larger ( walter et al . 2004 ) . ) .
the @xmath75 is expressed in m@xmath81 yr@xmath131 and the time in gyr .
the horizontal line shows the average value of the star formation rate over the lifetime of the simulated qso host galaxy , found to be @xmath155 yr@xmath131 , while the vertical line indicates the age of universe ( @xmath161 myr ) corresponding to a redshift @xmath156 , at which the quasar is observed .
, width=302 ] using the simulated star formation rate evolution as an input to the chemical evolution model presented above , we can compute the corresponding evolution of the mass of stars and dust . following li et al .
( 2007 ) , we assume that the qso is hosted by a dark matter halo of mass @xmath162 and that the initial mass in baryons is equal to the universal value , @xmath163 , i.e. @xmath164 . in the left panel of fig .
[ fig : liqso ] we show the evolution of the mass of stars ( dotted line ) , and dust ( solid line ) of the host galaxy .
the vertical solid lines indicate the age of universe at @xmath156 when the quasar is observed ( @xmath161 myr ) and the data point represents the dust mass inferred from mm / submm observations by beelen et al .
( 2006 ) , @xmath165 , with error - bars which account for the range of values quoted in the literature , @xmath166 .
the lower dashed , dotted and dashed - dotted lines indicate the time evolution of the three main dust components , carbon , silicate and other dust grains , respectively . for these components ,
thin lines represent the contribution of agb stars only .
the right panel shows the same quantities computed assuming a constant star formation rate equal to @xmath167yr@xmath131 ( the average value shown with the horizontal line in fig .
[ fig : lisfr ] ) . in this model
the total mass of stars formed at @xmath7 is @xmath168 , marginally consistent with the observed dynamical mass after correcting for the extension of the stellar bulge .
it is clear from the figure that our chemical evolution model with the simulated @xmath75 predicts a total mass of dust of @xmath169 at @xmath170 in agreement with observations .
we stress that this result is obtained following dust production by stellar sources on their characteristic stellar lifetimes and dust destruction by astration and sn shocks .
the latter mechanisms are responsible for the down - turns in the evolution of the dust masses shown in the figure .
the amount of dust destroyed by interstellar sn shocks depends on the age of the system : at the time corresponding to @xmath7 , we find that this effect reduces the mass of dust by 30% in both star formation models . on longer timescales ,
dust destruction is more pronounced in the case of a constant star formation rate ( sn progenitors continue to form ) resulting in more than 50% dust mass reduction at @xmath171 .
conversely , assuming the simulated star formation history , the fraction of dust destroyed at @xmath171 is only 20% .
this is due to sfr ( and therefore sn rate ) quenching by agn - feedback and by the larger relative contribution of agb stars in this model ( see below ) .
as expected on the basis of the adopted stellar dust yields , the dominant dust component is found to be carbon dust , with silicates and other dust species being a factor 10 less abundant .
moreover , agb stars are found to give a significant contribution to dust production at high redshift , accounting for @xmath17280 % of the total mass of dust at @xmath7 .
in particular , agb stars start to contribute to carbon dust production already 200 myr after the onset of star formation , but their contribution remains sub - dominant until the last major burst of star formation , which occurs at @xmath173 ( 0.6 gyr ) . following this burst , the mass of dust produced by agb stars steeply rises for all the dust species .
in fact , stars formed in this last major burst have nearly solar metallicities and their contribution to silicates and other dust components is no longer negligible .
these results depends on the adopted star formation history . when a constant star formation rate is assumed , as shown in the right panel of fig .
[ fig : liqso ] , the total mass of dust at @xmath174 is a factor 2 smaller and the contribution of agb stars is @xmath17250 % . since the stellar metallicities in this case are smaller , the contribution of agb stars to silicate and other dust components are negligible .
we stress that , although the star formation history of the qso host galaxies is still largely unknown , our analysis indicate that the contribution of agb stars to dust production can not be neglected , even at the most extreme redshifts currently accessible to observations . additional constraints on the chemical evolution model can be obtained from observations of the metallicity and elemental abundances in the broad line region ( blr ) of the qso .
in fact , the detection of metal lines such as fe ( barth et al .
2003 ) , cii ( maiolino et al . 2005 ) , and oi ( in absorption , becker et al .
2006 ) indicates that the ism of the qso host was significantly enriched by heavy elements . in a recent analysis , juarez et al .
( 2009 ) investigated the metallicity of the blr of a sample of 30 qsos in the range @xmath175 , including j1148 + 5251 , using the ( siiv+oiv)/civ ratio , which is a good metallicity tracer ( nagao et al . 2006 ) .
according to this analysis , the observed ratios do not show any correlation with redshift and correspond to metallicities which are several times solar . in fig .
[ fig : liqsomet ] we show the predicted evolution of the total metallicity in solar units ( solid line ) for the two star formation models shown in fig . [
fig : liqso ] .
the shaded area indicates the metallicity level inferred from observations of the narrow line regions ( nlr ) of a sample of high-@xmath176 radio - galaxies , which shows no evolution over the redshift range @xmath177 ( matsuoka et al .
juarez et al . ( 2008 ) determined the metallicity of the blr of j1148 + 5251 .
since the blr is a small nuclear region , of less than a few pc , contaning a total mass of @xmath178 , we consider this measurement to be an upper limit to the metallicity of the host galaxy .
conversely , the nlr has a size which is roughly comparable to the size of the host galaxy and represents a good tracer of the chemical properties on galactic scales ( matsuoka et al .
2009 and references therein ) . adopting the star formation history predicted by the simulation of li et al .
( 2007 , left panel ) , the ism reaches the solar metallicity already at @xmath179 , and remains constant thereafter .
a delayed metal - enrichment is found if a constant star formation rate is assumed ( right panel ) and the predicted metallicity at @xmath7 is @xmath180 .
both models are consistent with the upper limit to the total metallicity derived from observational data for j1148 + 5251 , and are in good agreement with the data inferred from nlrs observations at @xmath181 .
finally , we would like to comment on the so - called `` carbon - problem '' pointed out by juarez et al .
these authors argue that the lack of evolution of the observed ( siiv+oiv)/civ ratio between @xmath182 suggests that the abundance of carbon relative to silicon and oxygen also does not evolve significantly .
if carbon is mostly produced by agb stars , the bulk of its production must be delayed with respect to that of @xmath13-elements and we should expect to see an evolution at @xmath183 . however , the dotted lines in fig.[fig : liqsomet ] show that both star formation models predict a ( si+o)/c ratio of @xmath184 at @xmath156 , with only a mild evolution at lower redshifts .
the evolution is more pronounced in the left panel ( a factor 2 over the redshift range @xmath185 ) because at @xmath186 we expect to see the contribution to carbon ( and carbon dust ) enrichment of intermediate mass stars formed in the last major burst of star formation . as already discussed above ,
the contribution of intermediate mass stars is much less important when a constant star formation rate is assumed ( right panel ) .
the resulting mild evolution indicates that carbon production by massive short - lived stars , progenitors of sne , is not negligible relative to that of longer - lived intermediate mass stars , at least according to the stellar metal yields adopted in the present paper .
in this work we show that the role of agb stars in cosmic dust evolution at high redshift might have been over - looked .
we develop a chemical evolution model which follows the evolution of metals and dust on the stellar characteristic lifetimes , taking into account dust destruction mechanisms .
we use a grid of stellar dust yields as a function of the initial mass and metallicity over the range @xmath2 and @xmath3 .
we adopt the stellar dust yields by zhukovska et al .
( 2008 ) for agb stars with stellar masses in the range @xmath187 and the model of bianchi & schneider ( 2007 ) to account for dust production and destruction in the reverse shock of sne , with progenitor stellar masses in the range @xmath24 .
our results can be summarized as follows : * a comparison of dust yields expected from agb stars and sne shows that carbon grains are the dominant dust component .
silicates and other dust components can only be produced by agb stars with solar metallicities and by sne when @xmath188 but are always sub - dominant with respect to carbon grains .
this is at odd with what has been generally adopted in chemical evolution models which , assuming a dust condensation efficiency of 1 for all refractory elements present in the ejecta , predict that sne dominate silicate dust production and that agb stars lead to a delayed injection of carbon dust ( dwek 1998 ; dwek 2005 ; galliano , dwek & chanial 2008 ) . * at early times ,
dust production is dominated by sne as a consequence of the shorter lifetimes of their progenitor stars .
agb stars start to produce dust after about 30 myr ( i.e. when a @xmath189 star evolves off the main sequence to the agb phase ) . for our reference model , where we adopt a larson imf with @xmath4 and stellar lifetimes from raiteri et al .
( 1996 ) , we find that the characteristic timescale at which agb stars dominate dust production ranges between 150 and 500 myr , depending both on the assumed star formation history and on the initial stellar metallicity .
hence , we conclude that these stellar dust sources must be taken into account at early cosmic epochs . *
this conclusion is only moderately dependent on the adopted stellar lifetimes , but it is significantly affected by variations of the imf : for a @xmath5 , dust from agb starts to dominate only on timescales larger than 1 gyr , and sne are found to dominate dust evolution when @xmath6 .
thus , variations of the stellar characteristic mass over cosmic history due to the smaller gas metallicity and/or to the larger cosmic microwave background temperature ( see , e.g. , omukai et al .
2005 ) might significantly affect the origin and properties of dust in the high redshift universe .
* we apply the chemical evolution model with dust to the host galaxy of the most distant qso at @xmath7 , sdss j1148 + 5251 . given the current uncertainties on the star formation history of the host galaxy , we have considered two models : ( i ) the star formation history obtained in a numerical simulation by li et al .
( 2007 ) which predicts that a large stellar bulge is already formed at @xmath7 , in agreement with the local smbh - stellar mass relation , through a series of strong bursts triggered by galaxy mergers in the hierarchical evolution of the system and ( ii ) a constant star formation rate of @xmath8/yr , as suggested by the observations if most of the fir luminosity is due to young stars . in the latter case ,
the stellar mass at @xmath170 is compatible with estimates of the dynamical mass derived by observations .
the total mass of dust predicted at @xmath7 by the first model is @xmath9 , within the range of values inferred by observations , with a substantial contribution ( @xmath10% ) of agb - dust .
when a constant star formation rate is adopted , the contribution of agb - dust decreases to @xmath11% but the total mass of dust formed is a factor 2 smaller .
both models predict a rapid enrichment of the ism with metals and a relatively mild evolution of the carbon abundance , in agreement with observational constraints .
the evolution of dust in the high - redshift universe is still very uncertain . some of the chemical evolution models with dust developed so far have been aimed at reproducing the elemental abundances of the gas and dust phases of the ism of our galaxy ( dwek 1998 ; falceta - goncalves 2008 ; zhukovska et al .
2008 ) and of local ellipticals and dwarf irregulars ( calura , pipino & matteucci 2008 ) . in early galaxies ,
a crucial unknown is the relative contribution of sne and agb stars .
it is generally found that there is no difficulty in producing highly dusty galaxies at @xmath190 if sne are an important contributor to interstellar dust .
however , the models developed so far assume that all refractory elements present in sn ejecta form dust grains with a condensation efficiency of 1 ( dwek et al . 2007 ) or neglect dust destruction mechanisms ( morgan & edmunds 2003 ) .
the evolution of metals and dust in the host galaxy of the most distant quasar at @xmath170 sdss j1148 + 5251 has been also investigated by dwek et al .
( 2007 ) which show that an average sn must condense a dust mass of @xmath191 to reproduce the dust - to - gas ratio inferred from the observations ( 0.0067 ) , when a gas mass of @xmath192 and a dust mass of @xmath193 are assumed . since the observed dust yield is much lower ( the largest observed dust yield is @xmath194 , rho et al . 2008 ) , the authors suggest that mc - grown dust or dust formed around the agn ( i.e. in outflows from the broad line regions , elvis et al .
2002 ) must be taken into account .
our results suggest the alternative possibility that agb stars might significantly contribute to dust production even at redshifts @xmath100 . for a comparison
, we can infer from figure 8 in dwek et al .
( 2007 ) that , in order to reproduce the dust - to - gas ratios predicted by our sfr models for sdss j1148 + 5251 ( @xmath195 at @xmath7 ) , a dust yield per sn of about @xmath196 would be required .
this is larger than our adopted sn yields , the difference being compensated by the additional contribution to the dust budget provided by agb stars .
more recently , li et al .
( 2008 ) have applied a three - dimensional monte carlo radiative transfer code to the hydrodynamic simulations of li et al .
( 2007 ) to model the dust properties of j1148 + 5251 and find that a supernova - origin dust model may be able to explain the observed dust properties . yet
, dust destruction in the ism is not explicitely taken into account .
moreover , these models make the a - priori assumption that at redshift above 6 , when the age of the universe is less than 1 gyr , the major alternative source of interstellar dust , namely agb stars , have not yet evolved to their dust - production stage .
our results show that stellar sources can account for the huge dust masses inferred from observations of distant qsos .
although the star formation history of the qso host galaxies is still largely unknown , our analysis indicate that the contribution of agb stars to dust production can not be neglected , even at the most extreme redshifts currently accessible to observations .
yet , since agb winds are not efficient to disperse dust through the galaxy , sne are still required to efficiently re - distribute dust on galactic scales ( jones 2005 ) .
the nature and properties of the grains responsible for dust extinction in high - redshift qsos and grbs host galaxies ( maiolino et al .
2004 ; stratta et al . 2007 ; gallerani et al . in prep
) are likely to depend more on the specific host galaxy star formation history than on redshift .
future observations aimed at constraining the properties of dust in high redshift qsos might be a powerful probe of star formation and qso feedback regulating the bh - galaxy coevolution .
we thank the anonymous referee for the careful reading and the insightful comments which helped us to improve the clarity of the paper .
we are grateful to a. ferrara , s. gallerani , h. gomez , a. jones , r. maiolino , d. maoz , a. marconi and s. salvadori for useful suggestions .
r. schneider acknowledges the support and hospitality of the dark cosmology center .
the dark cosmology centre is funded by the danish national research foundation .
99 barth a.j . ,
martini p. , nelson c.h .
, ho l.c . , 2003 , apj , 594 , l95 beelen a. , cox p. , benford d.j . , dowell c.d . , kovacs a. , bertoldi f. , omont a. , carilli c.l .
, 2006,apj,642 , 694 becker g.d . ,
sargent w.l.w . ,
rauch m. , simoncoe r.a . , 2006 , apj , 640 , 69 bertoldi f. & cox p. 2002 , a&a 884,l11 bertoldi f. , carilli c.l . , cox p. , fan x. , strauss m.a .
, beelen a. , omont a. , zylka r. , 2003 , a&a , 406 , l55 bianchi s. , schneider r. , 2007 , mnras , 378 , 973 calura f. , pipino a. , matteucci f. , 2008 , a&a , 479 , 669 carilli c.l .
, bertoldi f. , rupen m.p . , fan x. , strauss m.a . ,
menten k.m , kreysa e. , schneider d.p .
, bertarini a. , yun m.s . , zylka r. , 2001 , apj 555 , 625 carilli c.l . , walter f. , bertoldi f. , menten k.m , fan x. , lewis g.f . , strauss m.a . , cox p.,beelen a. , omont a. , mohan n. , 2004 , aj , 128 , 997 clayton d. d. , arnett d. , kane j. & meyer b. s. , 1997 , apj , 486 , 824 clayton d. d. , deneault e. a .-
n . , meyer b. s. , 2001 , apj , 562 , 480 draine b.t .
, 1990 , in the evolution of the interstellar medium , ed
. l. blitz .
( san francisco asp ) , asp conf .
, 12 , 193 dunne l. , eales s. , ivison r. , morgan h. , edmunds m. , 2003 , nature , 424 , 285 dwek e. , 1998 , apj , 302 , 363 dwek e. , galliano f. , jones a.p .
, 2007 , apj , 662 , 927 dwek e. , 2005 , in the spectral energy distributions of gas - rich galaxies : confronting models with data , ed.c . c. popescu & r.j
( heidelberg , germany ) aip conf .
proc . , 761 , 103 elmhamdi a. , danziger i. j. , chugai n. , pastorello a. , turatto m. , cappellaro e. et al . , 2003 , mnras , 338 , 939 elvis m. , marengo m. , karovska m. , 2002 , apj , 567 , l107 falceta - goncalves d. , 2008 , a&a , 478 , 151 fan x. , strauss m.a . ,
schneider d.p . ,
becker r.h .
, white r.l .
et al . , 2003 ,
aj , 125 , 1649 feder j. , russell k. c. , lothe j. , pound g. m. , 1966 , advances in physics , 15 , 111 ferrarotti a.s .
, gail h.p . , 2006 , a&a , 447 , 553 freudling w. , corbin m.r .
, korista k.t . , 2003 , apj , 587 , l67 galliano f. , dwek e. , chanial p. , 2008 ,
apj , 672 , 214 hines d. c. , rieke g. h. , gordon k. d. , rho j. , misselt k. a. , woodward c. e. et al . , 2004 , apjs , 154 , 290 hopkins p. f. , strauss m. a. , hall p. b. , richards g. t. , cooper a s. , schneider d. p. et al .
2004 , aj , 128 , 1112 hughes d.h . , serjeant s. , dunlop j. , rowan - robinson m. , blain a. , mann r.g .
, ivison r. , et al . 1998 , nature , 394 , 241 jones a.p .
, tielens a.g.g.m . ,
hollenbach d.j .
, mckeec.f . , 1994 ,
apj , 433 , 797 jones a.p .
, tielens a.g.g.m . , hollenbach d.j . , 1996 ,
apj , 469 , 740 jones a.p . , 2005 , in the dusty and molecular universe : a prelude to herschel and alma , ed .
by a. wilson ( noordwijk esa ) , esa sp 577 , 239 juarez y. , maiolino r. , mjica r. , 2008 , rmxac , 32 , 105 juarez y. , maiolino r. , mujica r. , pedani m. , marinoni s. , nagao t. , marconi a. , olive e. , 2009 , a&a , 494 , 25 karakas a.i . , lattanzio j.c . , 2003 , pubblications of the astronomical society of australia , 20 , 279 kemper f. , de koter a. , waters l.f.b.m . , bouwman j. , tielens a.g.g.m . , 2002 , a&a , 384 , 585 kozasa t. , hasegawa h. , nomoto k. , 1991 , a&a , 249 , 474 krause o. , birkmann s. m. , rieke g. h. , lemke d. , klaas u. , hines d. c. , 2004 , nature , 432 , 596 larson r.b . , 1998 , mnras , 301 , 569 ledoux c. , bergeron j. , petitjean p. , 2002 ,
a&a , 385 , 802 li y. , hernquist l. , robertson b. , cox t.j .
, hopkins p.f . ,
springel v. , gao l. , di matteo t. et al .
, 2007 , apj , 665 , 187 li y. , hopkins p.f . , hernquist l. , finkbeiner d.p . , cox t.j . ,
springel v. , jiang , l. , fan x. , yoshida n. , 2008 , apj , 678 , 41 maiolino r. , comastri a. , gilli r. , nagar n. m. , bianchi s. , bker t. , colbert e. , krabbe a. , marconi , a. , matt , g. , salvati , m. , 2003a mnras , 344 , l59 maiolino r. cox , p. , caselli p. , beelen a. , bertoldi f. , carilli c. l. , et al .
2005,a&a , 440 , l51 maiolino r. , juarez y. , mujica r. , nagar n.m . , olive e. , 2003b , apj , 596 , l155 maiolino r. , nagao t. , marconi a. , schneider r. , bianchi s. , pedani m. , pipino a. , matteucci f. , cox p. , caselli p. , 2006 ,
mmsai , 77 , 643 maiolino r. , schneider r. , oliva e. , bianchi s. , ferrara a. , mannucci f. , pedani m. , roca sogorb m. , 2004 , nature , 431 , 533 mannucci f. , della valle m. , panagia n. , 2006 , mnras , 370 , 773 maoz d. 2008 , mnras , 384 , 267 marchenko s. v. , 2006 , in lamers h. j. g. l. m. , langer n. , nugis t. , annuk k. , eds , asp conf . ser .
353 , 299 matsuoka k. , nagao t. , maiolino r. , marconi a. , taniguchi y. , 2009 , submitted to a&a meikle w. , mattila s. , pastorello a. , gerardy c. , kotak r. , sollerman j. et al .
2007 , apj , 665 , 608 morgan h.l . , edmunds m.g . , 2003 , mnras , 343 , 427 nagao t. , marconi a. , maiolino r. , 2006 a&a , 447 , 157 nozawa t. , kozasa t. , umeda h. , maeda k. , nomoto k. , 2003 , apj , 598 , 785 nozawa t. , kozasa t. , habe a. , dwek e. , umeda h. , tominaga n. , maeda k. , nomoto k. 2007 , apj , 666 , 955 omont a. , cox p. , bertoldi f. , mcmahon r.g .
et al . 2001 , a&a 374 , 371 omukai k. , tsuribe t. , schneider r. , ferrara a. , 2005 , apj , 626 , 627 padovani p. , matteucci f. , 1993 , apj , 416 , 26 pentericci l. , fan x. , rix h.w . , strauss m.a .
, narayanan v.k . , richards g.t . ,
schneider d.p . , krolik j. , heckman t. , et al .
2002 , aj , 123 , 2151 pettini m. , smith l.j . , hunstead r.w . , king d.l . , 1994 ,
apj , 426 , 79 prochaska j.x . , wolfe a.m. , 2002 , apj , 566 , 68 raiteri c.m . , villata m. , navarro j.f . , 1996 ,
a&a , 315 , 105 rho j. , kozasa t. , reach w. t. , smith j. d. , rudnick l. , delaney t. , ennis j. a. , & gomez h. , 2008 , apj , 673 , 271 rho j. et al .
2009 , conference proceeding for `` cosmic dust - near and far '' ( heidelberg , germany ) , eprint arxiv:0901.1699 robson i. , priddey r.s .
, isaak k.g . ,
mcmahon r.g . , 2004 , mnras , 351 , l29 salpeter e.e . , 1974 , apj , 193 , 579 salvadori s. , ferrara a. , schneider r. 2008 , mnras , 386 , 348 schneider r. , ferrara a. , salvaterra r. 2004 , mnras , 351 , 1379 smail i. , ivison r.j .
, blain a.w . , 1997 , apj , 490 , l5 smartt s. j. ,
eldridge j. j. , crockett r. m. , maund j. r. , 2009 , mnras , submitted , eprint arxiv:0809.0403 stratta g. , maiolino r. , fiore f. , delia v. , 2007 , apj , 661 , l9 sugerman b. e. k. , ercolano b. , barlow m. j. , tielens a. g. , clayton g. c. , zijlstra a. a. et al .
2006 , science , 313 , 196 tielens a.g.g.m . , 1998 ,
apj , 499 , 267 todini p. , ferrara a. , 2001 , mnras , 325 , 726 van den hoek l.b .
, groenewegen m.a.t . , 1997 , a&as , 123 , 305 walter f. , carilli , c. , bertoldi , f. , menten , k. , cox , p. , lo , k.y . , fan , x. , strauss , m.a . , 2004 ,
apj , 615 , l17 wilson t. , batrla w. 2005 , a&a , 430 , 561 willott c.j . , mclure r.j .
, jarvis m.j . , 2003 ,
apj , 587 , l15 whittet d.c.b 1992 , in dust in the galactic environment , iop publishing bristol woosley s.e . ,
weaver t.a . , 1995 ,
apj , 101 , 181 wooden d. h. , rank d. m. , bregman j. d. , witteborn f. c. , tielens a. g. , cohen m. et al 1993 , apjs , 88 , 477 zhukovska s. , gail h.p . , trieloff m. , 2008 , a&a , 479 , 453 | with the aim of investigating whether stellar sources can account for the @xmath0 dust masses inferred from mm / sub - mm observations of samples of @xmath1 quasars , we develop a chemical evolution model which follows the evolution of metals and dust on the stellar characteristic lifetimes , taking into account dust destruction mechanisms . using a grid of stellar dust yields as a function of the initial mass and metallicity over the range @xmath2 and @xmath3 ,
we show that the role of agb stars in cosmic dust evolution at high redshift might have been over - looked .
in particular , we find that ( i ) for a stellar population forming according to a present - day larson initial mass function ( imf ) with @xmath4 , the characteristic timescale at which agb stars dominate dust production ranges between 150 and 500 myr , depending both on the assumed star formation history and on the initial stellar metallicity ; ( ii ) this result is only moderately dependent on the adopted stellar lifetimes , but it is significantly affected by variations of the imf : for a @xmath5 , dust from agb starts to dominate only on timescales larger than 1 gyr and sne are found to dominate dust evolution when @xmath6 .
we apply the chemical evolution model with dust to the host galaxy of the most distant quasar at @xmath7 , sdss j1148 + 5251 . given the current uncertainties on the star formation history of the host galaxy , we have considered two models : ( i ) the star formation history obtained in a numerical simulation by li et al .
( 2007 ) which predicts that a large stellar bulge is already formed at @xmath7 , and ( ii ) a constant star formation rate of @xmath8/yr , as suggested by the observations if most of the fir luminosity is due to young stars .
the total mass of dust predicted at @xmath7 by the first model is @xmath9 , within the range of values inferred by observations , with a substantial contribution ( @xmath10% ) of agb - dust .
when a constant star formation rate is adopted , the contribution of agb - dust decreases to @xmath11% but the total mass of dust formed is a factor 2 smaller .
both models predict a rapid enrichment of the ism with metals and a relatively mild evolution of the carbon abundance , in agreement with observational constraints .
this supports the idea that stellar sources can account for the dust observed but show that the contribution of agb stars to dust production can not be neglected , even at the most extreme redshifts currently accessible to observations .
[ firstpage ] galaxies : evolution , high - redshift , ism ; quasars : general ; stars : agb and post - agb , supernovae : general , ism : dust , extinction |
in recent years approaches based on the bootstrap technique @xcite-@xcite , within the conformal field theory ( cft ) framework , allowed to improve the precision on universal numbers needed to characterize the n - point correlation functions , such as scaling dimensions and operator product algebra ( ope ) coefficients . in particular new results have been obtained for the 3d ising model universality class @xcite@xcite , whose scaling dimensions were already known with very high precision from monte carlo ( mc ) simulations @xcite . however
a similar comparison between different methods is still difficult for the ope coefficients due to the lack of mc results . in the reference
@xcite a method based on the conformal perturbation theory was implemented for calculating the ope coefficients from off - critical correlators .
the method exploits the short distance expansion of the perturbed two - point functions , written as a power series of the wilson coefficients derivatives and the one - point functions . this procedure was already known @xcite-@xcite , but it was never applied to the three dimensional case . in @xcite the practical feasibility of the method was shown and in particular it was used to extract the ope coefficients of the 3d ising model , setting the magnetic field as perturbation , and the following results were found : @xmath5 and @xmath6 . in order to improve the precision of these results , and also to extend this procedure to another kind of perturbation , we consider now the ising model perturbed from the critical point with a non uniform confining potential coupled to the spin , obtaining for example this hamiltonian : @xmath7 where @xmath8 is the position vector of the site @xmath9 with respect to the central site of the lattice , @xmath10 is the power exponent of the trap and @xmath11 is the typical length .
therefore we have a central symmetry potential growing towards the edges .
the aim of this paper is to show that the conformal perturbation theory can be applied also in presence of a confining potential and that the ope coefficients of a model , in our case the 3d ising model , can be obtained by studying the correlation functions modified by the trap .
there are many reasons to motivate this study : the behavior of the critical system in presence of a trap is well described by the trap size scaling ( tss ) , a framework based on renormalization group arguments that allows to obtain the scaling laws of the observables at the critical point as a function of the trap parameters .
the tss has been successfully tested in many works both for quantum and classical models @xcite-@xcite . besides the presence of a trap is a common feature of an experimental setup in which the system has to be confined into a limited region .
there are experimental studies reproducing this situation for systems in the u(1 ) universality class @xcite-@xcite .
finally from a purely theoretical point of view it is interesting to combine the trap size scaling ansatz with the perturbations in cft , aiming to find the corrections to the two - point functions , because it shows the general applicability of the method , even in this case in which the translational symmetry is explicitly broken .
moreover there could be some advantages , as we will see the tss ansatz introduces a further external parameter , that in principle could be used to improve the signal of the perturbation terms that have to be calculated in the mc simulations .
in order to write down the correlation functions of the 3d ising model with a trapping potential , it is useful to summarize the general framework of the tss as established in the original reference @xcite . from the renormalization group framework ( rg ) we can write the scaling form of the singular part of the free energy density , near the critical point , with the usual notation but keeping into account a further scaling variable @xmath12 given by the trap : @xmath13 where we inserted also the space position @xmath14 to stress that now the translational invariance is lost and the observables are point dependent .
iterating the rg transformation as in the standard approach ( near the critical point we can assume @xmath15 ) , we obtain : @xmath16 where we defined @xmath17 , that is the characteristic trap exponent and can be deduced by means of scaling arguments from the corresponding continuum theory .
in fact if we consider the potential ( [ trapot ] ) as a perturbation coupled to the spin relevant field of a generic conformal action , we can write : @xmath18 if we perform a scale change with parameter @xmath19 , we have @xmath20 , @xmath21 , and for the field @xmath22 , where @xmath23 is the spin scaling dimension . since they must compensate within the ( [ actrap ] ) , we deduce : @xmath24 this expression depends only on the dimensionality of the system , the geometry of the confining trap and the universality class of the model . from the free energy density at the critical point ( @xmath25 , @xmath26 ) , we can deduce the scaling behavior of the one - point functions of the spin and the energy in the central point of the trap , respectively : @xmath27 where @xmath28 , @xmath29 are non universal constants and the label zero emphasizes that we are considering only local observables , typically defined in the central point , since in general the mean values are point dependent .
moreover the correlation length @xmath30 exhibits a scaling law @xmath31 with the trap parameters .
therefore the singularities of the observables at the critical point are smoothed , being the correlation length limited by the trap .
however the tss has a precise and universal scaling behavior determined by the typical length and the exponent @xmath32 .
for this reason we expect the tss can be applied also in our case for the perturbed correlation functions .
we shortly summarize the method already applied in @xcite .
the starting point is the ope @xcite written for perturbed systems : @xmath33 where the @xmath34 are a complete set of operators in cft , the @xmath35 is the expectation value over the action perturbed with a small relevant parameter @xmath36 .
the wilson coefficients @xmath37 can be expanded as a taylor series with respect to the perturbation parameter in a regular way without infrared divergences , so that the coefficients of the expansion are the derivatives calculated at @xmath26 @xcite .
in particular we note that the ope coefficients can be extracted from the first correction term of the critical correlator .
hence by calculating via mc simulations the perturbed two - point functions of relevant operators and the one - point functions , whose behavior is well known from the rg framework , the ope coefficients can be deduced . for further details see the references @xcite-@xcite .
for example for the spin - spin correlator with the magnetic field perturbation we have : @xmath38 in our case the trap perturbation breaks the translational symmetry , but it still leaves the rotational symmetry due to the shape of the confining potential ( [ trapot ] ) .
hence if we fix one operator in the center of the trap and the other one at radial distance @xmath39 , the ( [ example ] ) is still valid and the fusion rules between fields are the same of the uniform case . then substituting the one - point functions ( [ mtrap ] ) and ( [ etrap ] ) , we see that the expansion can be written as power series of the scaling variable @xmath40 , that can be deduced also from the ( [ fren ] ) .
so there is perfect analogy with the uniform magnetic field case , provided that the scaling exponent of the trap and the scaling variable of the tss ansatz are substituted , and we obtain straightforwardly : @xmath41 as in the general case , the expansion series for the correlators , taken from the center up to the distance @xmath39 , converges for distances less than about one correlation length .
the desired ope coefficients are included in the leading perturbation term , so that we can extract them knowing the tss behavior of the one - point functions .
moreover there are two free parameters , @xmath42 and @xmath32 , that can be adjusted to maximize the relevance of the first correction term of the expansion .
our aim is to check the validity of this scenario via mc simulations and to extract the 3d ising model ope coefficients .
we perform the simulations on a cubic lattice of side @xmath43 with fixed boundary condition .
the trap is centered in the middle point of the cube . on the lattice
we denote with @xmath44 the spin located at distance @xmath39 from the center upon the axis @xmath9 .
hence we define the following observables : the spin one - point function on the central site @xmath45 , and the energy one - point function in the middle of the lattice , defined as @xmath46 , where @xmath47 is the energy bulk contribution at the critical point , that must be subtracted .
the correlation functions are calculated from the central site of the lattice up to the distance @xmath39 on the central axis , averaging between the six possible directions . also for correlators involving the energy
the constant bulk contribution has to be subtracted .
therefore we have these definitions : @xmath48 usually the site energy is obtained by averaging the products with the neighbor spins in all the directions .
however now there is no more translational invariance , therefore , except for the central site , the links in different directions have not the same energy . however in our case
the trap acts as a perturbation , so that the energy difference between two neighbor sites due to the broken translational symmetry is negligible .
we numerically find that this assumption is correct within the precision of our simulations .
the same attention is needed for the spin - energy correlation function , since in principle with the exchanging of the two operators actually we obtain different correlators .
however we anticipate that , for the distances and the trap lengths involved in our simulations , we do not observe differences .
this confirms the validity of the selected window of perturbation parameters , since in the theory the operators can be exchanged without differences . the constraints to select the most appropriate trap parameter @xmath42 and the exponent @xmath10 are the same as in the uniform magnetic field case , i.e. having a sufficient large correlation length to sample the correlator and avoiding the finite size effects .
after some tests we find that a potential with power @xmath49 is the best choice , in fact with a smaller exponent the correlation length of the system is short while with larger @xmath10 we have to increase too much the size of the lattice to avoid finite size effects . we have checked that a cubic lattice of side @xmath50 is large enough to prevent them within our current precision .
this is a very large lattice in three dimensions , however since all the observables are closely sampled around the central site of the lattice , we can increase the speed of the simulation by using the hierarchical upgrade .
this method can be used when the observable is local : instead of doing the monte carlo sweep on the whole lattice before sampling the observable , the sweep is performed only on a box around the point of interest .
actually the algorithm works by defining many boxes of increasing side around the central point , and then performing nested cycles of sweeps on every box .
the detailed description can be found for example in the reference @xcite . in this way we can save time and increase the maximum lattice size .
moreover we sample correlators at different distances in different simulations , so that our data are uncorrelated .
we fix the following constants to their known value : the energy bulk contribution @xmath51 and the critical temperature @xmath52 from reference @xcite , the scaling dimension @xmath53 and @xmath54 from @xcite , with whom we have @xmath55 .
the uncertainty on these constants is negligible in our data set . with these settings
we find that the optimal range for the trap parameter is @xmath56 for which the correlation length is around 18 lattice spacings . in these simulations
we choose a thermalization time of @xmath57 sweeps , with starting configuration of all spins up aligned , and we sample about @xmath58 configurations for each correlation function .
the two - point functions at the critical point have to be normalized to one as in the continuum , since , for example , the spin - spin correlator on lattice is : @xmath59 therefore the spin normalization is @xmath60 , and similarly for the energy @xmath61 .
from these we can convert all the quantities from lattice to continuum units .
for further details see @xcite and @xcite .
the constants @xmath62 , @xmath63 are also important because they fix the leading order term of the spin - spin and the energy - energy correlators ; as we see in the reference @xcite they are the main source of systematic errors on the final estimate of the ope coefficients .
we fix @xmath64 as in @xcite , while for @xmath62 we have refined the result by finite size scaling study of the critical correlator , obtaining @xmath65 . finally we have to fix the non universal constants @xmath28 , @xmath29 of the formulas ( [ tcor1])-([tcor3 ] ) by studying the one - point functions .
we sample these observables in the optimal @xmath42 range selected previously , so that we find the expected power law behavior of formulas ( [ mtrap])-([etrap ] ) without scaling corrections .
we obtain @xmath66 and @xmath67 . as we learn from the reference @xcite
, the spin - energy correlator is the most suitable observable to determine the coefficient @xmath68 since it is zero at the critical point and there is not systematic error .
hence we can fit only the leading order term of ( [ tcor3 ] ) .
the table [ tab_tem ] shows the results of the fit for the values of @xmath42 in the optimal range .
the data sets are consistent with the functional form predicted by the tss ansatz . for comparison
we remind the result of the bootstrap approach , for example from @xcite , @xmath69 .
our results are in good agreement with this estimate , and thanks to the higher statistics we improve also the previous mc results . combining all the data series
we obtain @xmath70 . in the upper panel of figure [ fig_trapem ]
the data set for @xmath71 is shown , while in the lower panel all the rescaled data series are plotted in order to show that they are distributed around the expected scaling function @xmath72 .
the energy - energy correlator instead is the most interesting to determine the less known coefficient @xmath73 .
however this observable has a low signal due to the rapidly decreasing power law , and it is affected by a systematic error due to the critical point term .
we adopt the same convention of @xcite denoting the fit statistical error in round brackets and the systematic error in square brackets .
the table [ tab_te ] shows our results , that are fully consistent with the estimate of @xcite , and again we find a good agreement with the expected power law behavior of ( [ tcor2 ] ) .
we quote the final estimate @xmath74 .
the figure [ fig_trapee ] shows the plot for the data set @xmath75 and the scaling function @xmath76 .
finally we consider as consistency check the spin - spin correlator , since it is the observable with largest systematic error .
moreover it is affected by short distance corrections , which instead were negligible for the other correlators @xcite .
the table [ tab_tm ] shows the results , that are fully consistent with the previous estimates .
.results of the one parameter least square fit for the spin - energy correlator for some values of @xmath42 .
the columns @xmath77 , @xmath78 report the range of sampled distances .
the error quoted on @xmath79 takes into account also the standard error propagation from the normalization constants . [ cols="<,^,^,^,^",options="header " , ]
we have verified that the method of extracting the ope coefficients from perturbed correlators works also in presence of a trap : when the confining potential has a very large typical length compared to the distances involved for the correlators , it can be considered as a perturbation and the ope expansion for the perturbed two - point functions can be wrote following the general prescription .
the trap size scaling ansatz provides the tools to understand the behavior of the one - point functions and the exponents of the power law terms , characterized by the trap exponent @xmath32 .
the possibility to combine these tools to describe the perturbed correlators in principle was not ensured , due to the explicit translational symmetry breaking of the potential coupled to the spin operator .
therefore finding the expected ope coefficients of the 3d ising model universality class is itself a non trivial result .
moreover we can extract with a good precision the coefficients @xmath80 and @xmath81 ; unfortunately the latter one , which is very interesting being the most difficult to obtain also with the bootstrap approaches , has higher uncertainty than the former one .
however we are confident that this result could be improved with further mc simulations .
another possibility could be coupling the confining potential to the energy operator , whose trap size scaling has been equally investigated in @xcite . given our results
, there is no reason for which the same kind of study could not be repeated in this other case .
the main limitations to the trap approach are the large size of the lattice generally required ( a problem partially recovered by the hierarchical upgrades algorithm ) , and the fine tuning of the free parameters , the power exponent @xmath10 and the trap length @xmath82 , which requires some preliminary studies to fix the non universal constants and to find the most suitable sampling window .
however this drawback is acceptable if it leads to isolate the terms containing the ope coefficients in the perturbed correlators expansion . for @xmath83 .
b ) plot of the scaling function @xmath84 for all the data series .
though this observable has larger error bars than the spin - energy correlator , we can observe that the data are distributed around the expected behavior.,title="fig : " ] for @xmath83 .
b ) plot of the scaling function @xmath84 for all the data series .
though this observable has larger error bars than the spin - energy correlator , we can observe that the data are distributed around the expected behavior.,title="fig : " ] we fit the spin - spin correlator with @xmath85 , with @xmath86 and @xmath87 as free parameters .
the last term takes into account the short distance corrections induced by the next to the leading order term of the correlator at the critical point on the lattice .
this anstaz is sufficient to fit the correlator as reported in table [ tab_tm ] . | recently the operator product expansion coefficients of the 3d ising model universality class have been calculated by studying via monte carlo simulation the two - point functions perturbed from the critical point with a relevant field .
we show that this method can be applied also when the perturbation is performed with a relevant field coupled to a non uniform potential acting as a trap .
this setting is described by the trap size scaling ansatz , that can be combined with the general framework of the conformal perturbation in order to write down the correlators @xmath0 , @xmath1 and @xmath2 , from which the operator product expansion coefficients can be estimated .
we find @xmath3 , in agreement with the results already known in the literature , and @xmath4 , confirming and improving the previous estimate obtained in the uniform perturbation case . |
metastable structures are found in a variety of physical systems , _ e.g. _ glasses , amorphous solids of colloids @xcite , stalk intermediate structures in biological membrane fusion process @xcite , and so on .
metastable states correspond to regions of the local free energy minima in phase space . in conventional simulation of the canonical ensemble ,
once the system is captured in these regions , the system often remain in these non - equilibrium states for enormously long time .
such metastable states are frequently found in macromolecular and colloidal systems , and make the simulation studies on equilibrium states difficult . even if the system can escape from the non - equilibrium states to the ordered equilibrium state , the constant number of particles and the constant system size cause defects in the ordered structure .
this situation breaks , in a global scale , the anisotropy of the ordered structure and forces the periodicity of the ordered structure to be commensurate to the system size . in order to find defect - free
equilibrated ordered structures , we should finely tune the system box size @xmath0 as well as the number of particles @xmath1 so that both the anisotropy and the periodicity of the ordered structure , which are not known _ a priori _ , are not violated by the periodic boundary conditions .
this fine tuning for the ordered structure is in general a tedious task .
advanced simulation techniques , _
e.g. _ multicanonical ensemble method @xcite and umbrella sampling @xcite , allow us to almost homogeneously sample the whole phase space at constant @xmath1 and @xmath0 , with the help of artificial weights that reduce the occurrence probability of non - equilibrium states .
however , microscopic states in equilibrium , sampled by these advanced techniques , are restricted to microstates with the given set of constant @xmath1 and constant @xmath0 .
free energy landscapes of the system with different sets of @xmath1 and @xmath0 at the same particle density are not searched .
these extensive variables simultaneously need manual fine tuning for the purpose of finding the most stable state in these free energy landscapes .
for example , both @xmath1 and @xmath0 of perfect crystals should be integer multiples of the unit structure .
furthermore these advanced methods also require an advanced programming and large amounts of complicated preparation , _
e.g. _ accurate calculation of free energy @xcite and the precise adjustments of the artificial weights , prior to the production simulation runs .
in addition , such unphysical sampling processes with the artificial weights make it difficult to trace physical trajectories in the phase space . here
we devise molecular monte carlo simulation method of systems connected to three reservoirs ( hereafter we call it `` three - reservoirs method '' ) @xcite , chemical potential @xmath2 , pressure @xmath3 , and temperature @xmath4 , for seeking the most stable states of the target systems , _
i.e. _ the equilibrium structures . due to gibbs - duhem equation , @xmath5 where @xmath6 is the entropy and
@xmath7 is the volume , the number of reservoirs is thermodynamically limited to no more than 2 for single component systems .
however , we connect the three reservoirs in order to overcome the above difficulties of the other conventional and advanced simulation techniques . in order to perform this ,
we introduce a method for adjusting these three reservoirs and obtaining thermodynamically stable states . the total number of particles @xmath1 and the system box size @xmath0 are additional degrees of freedom of the system connected to these three reservoirs .
these additional degrees of freedom correspond to additional dimensions of the phase space , which provide shortcuts from the non - equilibrium state to the equilibrium state .
in addition , unlike the other simulation techniques utilizing 2 or fewer reservoirs , these degrees of freedom allow the system itself to simultaneously tune @xmath1 and @xmath0 , so that the system reaches the true equilibrium ordered structure . furthermore , our method requires fewer efforts for the preliminary simulation prior to the production simulation runs .
guggenheim formally introduced boltzmann factor ( statistical weight ) of the ensemble with the three reservoirs @xcite .
prigogine and hill also studied the same ensemble later @xcite .
these early works , however , focused on mathematical aspects of the partition function , _
i.e. _ mathematical formalism of the ensemble , since their goal was to discover a universal and generalized expression for a partition function applicable to any thermodynamically acceptable ensembles @xcite .
by contrast , physical aspects of the ensemble were wholly left for the future . in the present work ,
we study the physical aspects of this ensemble intuitively and thought - experimentally .
in addition , we also analytically solve maximization problems of entropy densities , which corroborates our intuitive and thought - experimental study . finally , we design three - reservoirs method based on these physical aspects of the ensemble and show simulation results on non - trivial globally - anisotropic defect - free ordered structures of colloidal systems . in spite of gibbs - duhem equation
, a system connected to the three reservoirs can be realized in experiments .
for example , we can imagine a system that obeys the grand canonical ensemble ( _ i.e. _ constant @xmath2 , @xmath7 , and @xmath4 ) , and replace one of its walls with a free piston facing to a reservoir of pressure @xmath3 , _ i.e. _ the 3rd reservoir . in the present article ,
we theoretically construct thermodynamics and statistics of the system connected to the three reservoirs .
based on euler equation of thermodynamics , we also propose a method for measuring entropy and free energy directly from the simulations of the systems connected to the three reservoirs .
this measurement needs fewer computational efforts than the other free energy calculation methods using molecular simulation .
we design the algorithms of the three - reservoirs method based on conventional monte carlo ( mc ) simulation methods of the grand canonical ensemble ( @xmath8-ensemble ) and the isothermal - isobaric ensemble ( @xmath9-ensemble ) .
we give a brief description of these conventional molecular mc techniques in appendix [ sec : molecularmctechniqueinmuvtandnpt - ensembles ] .
thermodynamics , statistical mechanics , and simulation methods of the system with the three reservoirs are studied in section [ sec : muptensemble ] .
finally , we summarize the present work in section [ sec : conclusions ] .
here we discuss the basic formalism of the three - reservoirs method , where the method for adjusting the three reservoirs is also developed .
thermodynamic properties of this system are discussed in section [ subsec : thermodynamicsinmuptensemble ] .
a microscopic formulation of three - reservoirs method based on statistical mechanics will be given in section [ subsec : statisticalmechanicalpropertiesofparticlesinmuptensemble ] , where we solve the maximization problem of the statistical entropy per volume .
algorithms for the simulation based on this statistical formulation are constructed in section [ subsec : mcsimulationmethodinmuptensemble ] .
simulation results to demonstrate efficiency and stability of three - reservoirs method are given in section [ subsec : examinationof3-reservoirsmethod ] .
finally , a new and simple technique to measure the entropy and the free energy of the system is proposed in section [ subsec : entropyandfreeenergycalculationinmuptensemble ] . as an example of thermodynamic systems , we consider a gas contained in a diathermal box with a free piston .
this box is placed in an environment at constant @xmath4 and constant @xmath3 .
these two intensive variables @xmath11 determine the other intensive variables of this system , _
e.g. _ the chemical potential @xmath2 , the number density of particles @xmath12 , and the free energy per particle @xmath13 .
this means that thermodynamic degrees of freedom of this system are equal to 2 , which results from gibbs - duhem equation , eq . .
in conventional simulation methods , @xmath1 is also fixed at some value ( @xmath9-ensemble ) , whereas states and phases of the system are independent of @xmath1 , a change of @xmath1 only scales the extensive variables of the system . in other words ,
phase diagrams constructed in @xmath14-plane are independent of the extensive variables . instead of fixing this insignificant @xmath1 , we connect this system to a reservoir of @xmath2 , whose value is determined by @xmath11 through gibbs - duhem equation , eq . .
since gibbs - duhem equation is satisfied , this third reservoir does not affect the equilibrium state of the system .
this condition corresponds to a thermodynamically stable point based on gibbs - duhem equation , which results in an equation of state that links @xmath15 , and @xmath2 . at a thermodynamically stable point of this ensemble , the extensive variables of the system , _
e.g. _ @xmath7 and @xmath1 , are freely scaled , _
i.e. _ indeterminate and fluctuating , while the system keeps all the intensive variables fixed . in simulating systems connected to the three reservoirs at the thermodynamically stable points , we can choose simulation runs at small @xmath1 , which are computationally advantageous .
gibbs free energy per particle , which must be minimized in @xmath9-ensemble before adding the 3rd reservoir of constant @xmath2 , is unchanged even after this 3rd reservoir is connected to the system . in the above example
, @xmath11 is given from the outside of the system ; @xmath2 is adjusted according to @xmath11 and connected to the system as an additional reservoir .
two other combinations @xmath16 and @xmath3 , and @xmath17 and @xmath4 also work in a similar manner .
therefore , in addition to gibbs free energy per particle , both grand potential per volume and the thermodynamic potential of @xmath18-ensemble per @xmath6 , _ i.e. _ @xmath19 , are simultaneously minimized in the system connected to the three reservoirs , where @xmath6 and @xmath20 are entropy and internal energy respectively .
gibbs - duhem equation explains this simultaneous minimization of the 3 free energy densities .
when we add the 3rd reservoir to the system and set the system at the thermodynamically stable point , the intensive variable of the 3rd reservoir needs to be adjusted , in advance of the connection of this additional reservoir . through this adjustment of the 3rd intensive variable ,
the corresponding free energy density is minimized .
the same system at the same thermodynamically stable point is also constructed by the two other combinations of the 3 intensive variables . in other words , @xmath9 , @xmath8 , and @xmath18-ensembles simultaneously underlie the ensemble with the three reservoirs .
this results in the simultaneous minimization of the 3 free energy densities . in the other ensembles ,
however , any sets of corresponding 3 external parameters , _
e.g. _ @xmath21 in the canonical ensemble ( @xmath22-ensemble ) , can be selected arbitrarily and the adjustment of the external parameters is not demanded .
therefore , the corresponding free energy , _ e.g. _ helmholtz free energy in @xmath22-ensemble , is minimized in a conventional ensemble , whereas other free energies are not minimized due to the absence of underlying ensembles .
a system connected to the three reservoirs is sketched in fig .
[ fig : sketchmuptreservoirs ] .
the reservoirs 1 and 2 define the values of @xmath2 and @xmath3 of the system respectively .
the particle density of these 2 reservoirs , denoted by @xmath23 and @xmath24 respectively , determine the @xmath2 and @xmath3 .
the system and the reservoirs 1 and 2 are connected to a thermostat at @xmath4 , the reservoir 3 . in appendix
[ sec : molecularmctechniqueinmuvtandnpt - ensembles ] , we give a derivation of this @xmath2 , and its explicit expression is given in eq . . ,
where @xmath1 of the system is a dynamic variable . between the reservoir 2 and the system ,
a free piston is placed . this piston moves and changes the volume @xmath7 to fix the pressure of the system at @xmath3 .
the system and these reservoirs are connected to a thermostat at @xmath4 , the reservoir 3 . ] as another simple example , we thermodynamically consider a system composed of a single - component ideal gas . in this example , we assume that the reservoirs are also composed of the same ideal gas . at the thermodynamically stable point ,
a relation @xmath25 holds and the particle number density of the system , @xmath26 , also equals these particle number densities of the reservoirs ; _ i.e. _ @xmath27 . however , when @xmath28 , both @xmath1 and @xmath7 diverge , since the reservoir 1 continues to increase @xmath1 of the system aiming for a large @xmath26 value and the reservoir 2 increases @xmath7 aiming for a small value of @xmath26 .
when @xmath29 , both @xmath1 and @xmath7 vanish .
therefore , the system reaches equilibrium only at the thermodynamically stable point .
in other words , outside the thermodynamically stable point , the system is always in non - equilibrium and both the intensive and the extensive variables are indeterminate .
these results also apply to systems composed of interacting particles ( _ e.g. _ non - ideal gases ) .
we utilize these divergence and vanishment of the target system as a criterion for the equilibration , which can be used for the automatic adjustment of the three intensive variables , _
e.g. _ by a bisection method , to determine the thermodynamically stable point .
the system quickly diverges or vanishes outside the vicinity of the thermodynamically stable points @xcite .
the speed of the divergence and the vanishment increases with the difference in the intensive parameter sets from the thermodynamically stable point .
on the other hand , when we need a long simulation run in the vicinity of the thermodynamically stable point , indeterminate @xmath1 could cause a computational problem , since the extensive variables could become extremely large or vanish .
however , an appropriate choice of @xmath1 and @xmath7 makes the system last for a long time , within which good statistics of simulation results , _ e.g. _ particle density and lattice constants of crystals , are obtained @xcite .
we can determine such simulation results with accuracy enough to obtain the equilibrated structure of the system with the three reservoirs .
if we need a far longer simulation run , the ensemble can be switched to one of the conventional methods , _
e.g. _ @xmath9-ensemble or @xmath22-ensemble . as these switched ensembles
are free from the problem of the indeterminate @xmath1 , they allow us to perform a longer simulation run of the equilibrated structure obtained via three - reservoirs method . in the present article
, we tentatively call the ensemble of the systems connected to the three reservoirs @xmath10-ensemble , since this is obtained as an equilibrium condition between the three intensive variables and as a combination of @xmath8 , @xmath9 , and @xmath18-ensembles . here
the statistical mechanical properties of particles in @xmath10-ensemble are discussed . according to gibbs - duhem equation
, the thermodynamic potential of @xmath10-ensemble is identically equivalent to zero in the thermodynamic limit , whereas systems obeying this ensemble have certain degrees of freedom in the phase space .
this means that boltzmann factor fluctuates in statistical mechanics and that the partition function of this ensemble , _ i.e. _ summation of statistical weights over the whole phase space , is defined .
this boltzmann factor is calculated in the present section as a natural extension of those for @xmath8 and @xmath9-ensembles , and determines detailed balance conditions necessary for designing three - reservoirs simulation method . for determining this boltzmann factor
, we solve a maximization problem of the statistical entropy per volume in @xmath10-ensemble .
the statistical entropy is defined as @xcite , @xmath30 where the suffix , @xmath31 , represents microstates of the system , @xmath32 denotes corresponding occurrence probability , and @xmath33 is boltzmann constant . because the extensive variables of @xmath10-ensemble is indeterminate ,
we choose the entropy density to be maximized . with the use of eq .
, the statistical entropy per volume is given by , @xmath34 the equilibrium probability distribution , @xmath35 , which maximizes eq .
under the constraints in @xmath10-ensemble , is determined by : @xmath36 where @xmath37 is the thermal average specified according to the reservoirs .
the constraint of eq . represents the normalization condition .
the two constraints given by eq . instead of three constraints
are due to the thermodynamic degrees of freedom , which equal 2 . with the use of lagrange multipliers , @xmath38 , @xmath39 , and @xmath40 ,
this maximization problem is reduced to @xcite , @xmath41.\end{aligned}\ ] ] this reduced problem , eq .
, is solved by a partial derivative with respect to @xmath32 .
the solution is : @xmath42,\ ] ] where @xmath43 , @xmath44 , and @xmath45 .
equations - , and give the statistical entropy per volume determined as thermal average as : @xmath46 equating eq . with the thermodynamic relation for the thermodynamic potential of @xmath10-ensemble , denoted by @xmath47 , @xmath48
we obtain : @xmath49 in the thermodynamic limit , the right - hand side of eq .
is essentially zero compared with the other extensive variables in eq . .
this result coincides with the euler equation in thermodynamics .
furthermore , for finite @xmath7 , the last term of eq . is monotonically increasing with decreasing @xmath7 .
this indicates that due to the principle of maximizing entropy a finite size system as is used in the computer simulation has a tendency to shrink even at the thermodynamically stable point . finally , from eq .
, the probability distributions of the microstates are obtained as , @xmath50.\ ] ] we confirmed that these results are also obtained by maximizing the statistical entropy per particle or the statistical entropy per internal energy instead of the statistical entropy per volume as has been done in eq . .
the boltzmann factor of @xmath10-ensemble is , @xmath51,\ ] ] where @xmath52 denotes the total kinetic energy of the system , @xmath53 the potential energy of the system , and @xmath54 is the spatial coordinates of the particle @xmath55 . as we have seen above ,
this is a result of a natural extension of @xmath8 and @xmath9-ensembles .
this boltzmann factor determines the statistical properties of systems equilibrated with the three reservoirs and is equivalent to the boltzmann factor of @xmath9-ensemble at fixed @xmath1 , and is equivalent to the boltzmann factor of @xmath8-ensemble at fixed @xmath7 .
guggenheim formally derived the boltzmann factors ( statistical weight ) of various ensembles @xcite . assuming that the ensemble averages of extensive variables , e.g. @xmath7 and @xmath1 , were determined in @xmath10-ensemble
, guggenheim also introduced statistical weight of this ensemble , based on analogy between other conventional ensembles @xcite .
however , in the calculation of the maximization of entropy density discussed in section [ subsec : statisticalmechanicalpropertiesofparticlesinmuptensemble ] , guggenheim s assumption corresponds to keeping the averages @xmath56 and @xmath57 fixed , instead of the constraints eq . .
guggenheim s assumption contradicts the indetermination of the extensive variables , as was pointed out by prigogine and further discussed by sack @xcite .
prigogine showed that the summation of the resulting boltzmann factor over the phase space diverges and therefore concluded that the resulting partition function of @xmath10-ensemble does not have any physical meanings @xcite .
the thermodynamic potential of @xmath10-ensemble , @xmath47 , determined from such partition function could be indefinite while it should identically equal zero in the thermodynamic limit .
the true thermodynamic potential that dominates this ensemble , similar to the helmholtz free energy in @xmath22-ensemble , remains unknown since guggenheim s article . in the following
, we will answer prigogine s criticism and give an explicit expression of the thermodynamic potential for @xmath10-ensemble . in conventional ensembles ,
the statistical weight takes non - zero values only in the vicinity of the averages @xmath57 or @xmath58 in the phase space . outside this vicinity
, the statistical weight quickly decreases to zero .
this suppresses the divergence of the partition functions , _
i.e. _ the summation of the statistical weights , in the case of conventional ensembles . on the other hand , in @xmath10-ensemble , there is no such limitation because of the indeterminate extensive variables .
the statistical weight of @xmath10-ensemble at each microstate keeps non - zero values at any @xmath1 or @xmath7 .
this results in the divergence of the partition function of @xmath10-ensemble .
however , ratios of the statistical weights between any pair of microstates are still defined .
this feature guarantees the physical validity of @xmath10-ensemble . in this case , the trajectory of the system in the phase space is similar to a free random walk in infinitely large space without boundaries .
moreover , in the present study , equations and calculated with the constraints eqs . and indicate that the summation of the statistical weights equals @xmath59 for @xmath10-ensemble .
therefore , the corresponding thermodynamic potential @xmath47 , eq . , is negligibly small compared with the other extensive variables in eq . and vanishes in the thermodynamic limit .
furthermore , by the thermodynamic consideration given in section [ subsec : thermodynamicsinmuptensemble ] , we have shown that @xmath10-ensemble is obtained by combining 3 underlying ensembles each with 2 reservoirs , _
i.e. _ @xmath9 , @xmath8 , and @xmath18-ensembles .
this thermodynamic consideration means that free energy densities of these underlying ensembles , _
i.e. _ gibbs free energy per particle , grand potential per volume , and @xmath19 , are simultaneously minimized in @xmath10-ensemble , rather than @xmath47 .
this corresponds to the minimization of helmholtz free energy in @xmath22-ensemble .
the present simulation method is constructed based on conventional simulation methods in the grand canonical ensemble ( @xmath8-ensemble ) and @xmath9-ensemble .
the simulation algorithms of the particle insertion and deletion in @xmath8-ensemble ( see appendix [ subsec : mcsimulationmethodinmuvtensemble ] ) and the system size change in @xmath9-ensemble ( see appendix [ subsec : mcsimulationmethodinnptensemble ] ) are directly utilized in our method .
this compatibility between the present method and the conventional methods demonstrates that the algorithms of the present method satisfy detailed balance condition .
one simulation step of the present method is composed of the following 4 trial steps : a. with probability @xmath60 , trial particle insertion into the system , b. with probability @xmath60 , trial particle deletion from the system , c. with probability @xmath61 , trial system size change , d. with probability @xmath62 , trial displacement by metropolis algorithm _
i.e. _ perturbation of one particle , is chosen , where @xmath63 and @xmath61 are constants fixed in an interval @xmath64 . with the use of the simulation algorithms of trial particle insertion and deletion in @xmath8-ensemble , the insertion and deletion of the present method ,
steps i ) and ii ) , are performed . during this particle exchange between the system and the reservoir 1 ,
the system size @xmath65 is fixed .
this particle exchange in the present method satisfies the detailed balance condition because it is guaranteed in @xmath8-ensemble .
the trial system size change , step iii ) , is performed with use of the simulation algorithms in @xmath9-ensemble , during which @xmath1 is fixed .
this system size change in the present method also satisfies the detailed balance condition . unlike the conventional mc simulations in @xmath9-ensemble based on mcdonald s method @xcite , @xmath66 , and @xmath67
are independently changed in the present method . the trial move of one particle , step iv ) ,
is performed by metropolis algorithm in @xmath22-ensemble , which also satisfies the detailed balance condition .
therefore , the present mc simulation method for @xmath10-ensemble fulfills the principle of detailed balance .
see also section [ subsubsec : detailedbalanceconditionandergodicityof3-reservoirsmethod ] .
our algorithm indicates that a short - time average of an intensive physical quantity in @xmath10-ensemble is approximated by ensemble averages of @xmath8 and @xmath9-ensembles at corresponding @xmath1 and @xmath7 , which is discussed in appendix [ sec : ensembleaverageateachnandvinmuptensemble ] .
our simulation is performed in a rectangular system box with independently changing system size @xmath65 . since any crystals fit into rectangular boxes with periodic boundary conditions
, we do not have to introduce parrinello - rahman method @xcite , which allows the change in the shape of the simulation box .
our simulation algorithm is similar to gibbs ensemble technique , which is utilized for simulation of phase equilibria in @xmath22-ensemble @xcite , where the phases coexisting in the same system box in @xmath22-ensemble exchange both volume and particles .
a system connected to the three reservoirs at a thermodynamically stable point corresponds to this gibbs ensemble , when one of the coexisting phases in gibbs ensemble is assumed to be infinitely large that plays the role of the two reservoirs @xmath2 and @xmath3 . in this section
, the detailed balance condition and the ergodicity of three - reservoirs method are discussed .
steps i ) and ii ) change @xmath1 according to the detailed balance condition as in the same way that the @xmath8-ensemble does .
step iii ) changes @xmath65 according to the detailed balance condition as in the @xmath9-ensemble .
the particle coordinates are updated in step iv ) , by the standard metropolis algorithm of @xmath22-ensemble . as a result ,
three - reservoirs method satisfies the principle of detailed balance and ergodicity , based on the conventional ensembles which fulfill ergodicity .
this indicates that @xmath1 , @xmath0 , and the particle coordinates are simultaneously updated in the phase space , so that the system realizes the equilibrium state .
this also means that statistics of @xmath10-ensemble contradicts none of the three underlying ensembles with 2 reservoirs , _
i.e. _ @xmath8 , @xmath9 , and @xmath18-ensembles . in this section
, we demonstrate the efficiency and stability of the three - reservoirs method using several examples .
the first example is a model polymer - grafted colloidal system , which was observed to show various exotic metastable phases each of which has a long life time @xcite .
colloidal particles are made from metals , polymers , _ etc .
_ and are often modeled with hard spheres @xcite . on the other hand ,
owing to van der waals attraction acting on particle surfaces , colloids are aggregating and make precipitates after a long time .
for the purpose of stabilizing colloidal dispersions against the precipitation , linear polymer chains are often grafted onto the surfaces of the colloids .
these are called polymer - grafted colloids . depending on the physical and chemical properties of the grafted polymers , interaction between polymer - grafted colloids
changes significantly @xcite .
polymer - grafted colloids have several industrial applications due to such useful characteristics , for example filler particles immersed in a polymer matrix , and the particles in electro and magnet rheological fluids . in our previous work
@xcite , we studied the phase behavior of colloidal particles onto which diblock copolymers are grafted .
pair interaction potential between these polymer - grafted colloids was numerically determined via self - consistent field calculation @xcite as a function of the distance between centers of the particles , @xmath68 .
this potential has been approximated by spherically symmetrical repulsive square - step potential with a rigid core of diameter @xmath69 and a square - step repulsive potential of diameter @xmath70 and height @xmath71 as : @xmath72 where the step potential is originated from the grafted polymer brushes .
this interaction potential @xmath73 is purely repulsive . simulating particles interacting via @xmath73 in @xmath22-ensemble ,
we have studied phase behavior of these colloidal systems .
these mc simulation results @xcite show that , at low temperature , high pressure , and @xmath74 , our particles self - assemble into string - like assembly .
the positions and the mean - square displacement of the particles show that the string - like assembly is observed in disordered solid phases .
actually , such a string - like assembly has been observed recently in experiments @xcite .
in addition to such string - like assembly , various structures , _ e.g. _ dimers and lamellae @xcite , and also glass transition @xcite are observed in the same model system .
it was also shown that the particles interacting via continuous repulsive potential similar to the above @xmath73 show these string - like and other various assemblies @xcite . in these recent studies using @xmath22-ensemble at finite @xmath4 in both 2 and 3-dimensions
, it was shown that this string - like assembly with a local alignment in the same direction but with a global isotropy is a metastable structure .
although a variety of ground states of the same model system at zero temperature have been discovered via genetic algorithms @xcite , equilibrium states at finite temperature have not been understood yet .
we simulate the equilibrated states of our model system at finite @xmath4 via three - reservoirs method . in the present simulation work , @xmath69 and @xmath75
are taken as the unit length and the unit energy , respectively .
simulation is performed on 2 dimensional systems .
we define dimensionless chemical potential as @xcite , @xmath76 the thermal de broglie wave length @xmath77 , defined in eq . in appendix
[ sec : molecularmctechniqueinmuvtandnpt - ensembles ] , is removed from @xmath2 in eq .
, since simulation results are independent of @xmath77 , which is discussed in appendix [ sec : molecularmctechniqueinmuvtandnpt - ensembles ] . in the present article
, @xmath78 is fixed .
in the initial state , @xmath79 particles are arranged on a homogeneous triangular lattice in a square system box , @xmath80 , with the periodic boundary condition . for the trial move of the particles , _
i.e. _ metropolis algorithm , a particle is picked at random and given a uniform random and isotropic trial displacement within a square whose sides have length @xmath81 .
@xmath82 is fixed at @xmath83 .
the probability @xmath84 and @xmath85 . with this @xmath61 ,
the computational time for the simulation is about twice as long as the simulation in @xmath22-ensemble
. 1 monte carlo step ( mcs ) is defined as @xmath79 simulation steps .
the mersenne twister algorithm @xcite is adopted as a uniform random number generator for our simulation . in our previous @xmath22-ensemble simulation with the potential step width , @xmath78
, we found @xcite that the string length diverges at @xmath86 and @xmath87 . at this low @xmath88 , using three - reservoirs method , we simulate the system . in the present simulation , the density @xmath89 is initially set at 0.451 for @xmath90 system .
first , simulating the system at various values of @xmath91 , we search for the thermodynamically stable point at fixed @xmath92 .
the given @xmath91 is , 1 ) : @xmath93 , 2 ) : @xmath94 , 3 ) : @xmath95 , 4 ) : @xmath96 , and 5 ) : @xmath97 . as an example
, a snapshot of the system at 1 ) is presented in fig .
[ fig : frog2dmupts20snapshots](a ) . despite different @xmath91 ,
all the systems of 1 ) to 5 ) show similar well - aligned globally - anisotropic defect - free string - like assembly , though the system in @xmath22-ensemble at @xmath98 and @xmath99 presents the globally - isotropic string - like assembly as is shown in fig .
[ fig : frog2dmupts20snapshots](c ) @xcite . only small and short - lived defects caused by the thermal fluctuation
can be generated in the systems simulated with @xmath10-ensemble , whereas many long - lived defects are observed in @xmath22 simulation .
time evolutions of @xmath89 and @xmath1 at 1 ) to 5 ) are given in fig .
[ fig : frog2dmupts20t012forarticle - rhoandn ] . all the data for the time evolution of @xmath89 shown in fig .
[ fig : frog2dmupts20t012forarticle - rhoandn](a ) are fluctuating in the vicinity of @xmath87 , regardless of @xmath91 .
it would be worth noting that the simulations started from different initial conditions , _
e.g. _ different initial particle density @xmath89 and different system aspect ratio of the simulation box @xmath100 , also reach the same results as long as the intensive variables of the reservoirs , @xmath91 , @xmath101 , and @xmath88 are the same . furthermore , outside this range of @xmath91 from 1 ) to 5 ) , the system diverges ( @xmath102 ) or vanishes ( @xmath103 ) just after the simulation starts .
these results illustrate that this region of @xmath91 is located in the vicinity of the thermodynamically stable point at this pressure , @xmath92 . , ( a ) , and time evolution of @xmath1 , ( b ) , simulated at @xmath86 and various @xmath104 , @xmath105 .
lines 1 ) to 5 ) are simulation results at @xmath92 .
1 ) : @xmath93 .
2 ) : @xmath94 .
3 ) : @xmath95 .
4 ) : @xmath96 .
5 ) : @xmath97 . lines 6 ) to 8) are simulation results at @xmath94 .
6 ) : @xmath106 .
7 ) : @xmath107 .
8) : @xmath108 . ]
on the other hand , time evolution of @xmath1 , plotted in fig .
[ fig : frog2dmupts20t012forarticle - rhoandn](b ) , indicates that the total system size depends on @xmath91 in a systematic manner . at small mcs , line 1 )
indicates the tendency of the vanishment and lines 2 ) to 5 ) the tendency of the divergence .
this means that , in a short computational time , the thermodynamically stable point is expected to lie between 1 ) : @xmath93 and 2 ) : @xmath94 , which corresponds to a relative error of some percent .
the abrupt time evolution of @xmath1 stops at less than @xmath109 mcs , whereas the system slowly continues diverging or vanishing at larger mcs .
the mcs needed for the divergence or the vanishment becomes larger when we choose parameter sets close to the exact values at the thermodynamically stable point .
this is utilized as a criterion for measuring the convergence of the thermodynamic intensive variables in @xmath10-ensemble .
although we stop the simulation with this relative error of some percent , the accuracy of the thermodynamically stable point could be raised , _
e.g. _ by the bisection method improving the accuracy of @xmath91 .
lines 6 ) to 8) in fig .
[ fig : frog2dmupts20t012forarticle - rhoandn ] show the results of a similar series of simulations for a fixed value of @xmath94 ( the same value as that for line 2 ) ) and changing the value of @xmath101 .
this @xmath91 is consistent with the thermodynamically stable point at @xmath92 obtained above within the relative error of some percent .
the given @xmath101 of lines 6 ) to 8) ranges @xmath106 to 1.1 . a snapshot of the system at 8) is presented in fig . [ fig : frog2dmupts20snapshots](b ) .
despite different intensive parameter sets , all the systems of 6 ) to 8) show the globally - anisotropic defect - free string - like assembly similar to fig .
[ fig : frog2dmupts20snapshots](a ) .
long - lived defects are absent in these systems .
all the lines 6 ) to 8) in fig .
[ fig : frog2dmupts20t012forarticle - rhoandn](a ) are fluctuating around lines 1 ) to 5 ) , _ i.e. _ the vicinity of @xmath87 .
for the values of @xmath101 outside this range , @xmath1 quickly diverges or vanishes . from these data , we recognize that the present thermodynamically stable point lies between 7 ) : @xmath107 and 8) : @xmath108 .
this thermodynamically stable point is equal to the one we obtained above within a relative error of some percent , as is expected .
once the thermodynamically stable point is determined accurately in @xmath10-ensemble , we can exchange the ensemble to a conventional one , _ e.g. _ @xmath9-ensemble or @xmath22-ensemble , and can perform a longer simulation run , which is free from the divergence or the vanishment of @xmath1 .
as is expected , we can perform longer simulation runs even with the @xmath10-ensemble if the thermodynamically stable point is determined with higher accuracy .
next , we check the stability of our simulation method . starting from the instantaneous microstate shown in fig .
[ fig : frog2dmupts20snapshots](a ) , we resume simulation runs after disconnecting the reservoirs 1 or 2 , which corresponds to simulations with @xmath9 or @xmath8-ensemble . snapshots of the system and the time evolution of @xmath89 in these resumed simulation runs are given in figs .
[ fig : frog2dmupts20snapshotsinvariousensembles ] and [ fig : frog2dmupts20rd001e14p10t012innptmuvt - rho ] respectively .
both of these simulation runs preserve the same string - like assembly . in @xmath9-ensemble case
, a value of @xmath89 that is close to that for three - reservoirs method is also obtained .
therefore , the equilibrium state obtained via three - reservoirs method does not change even after the ensemble is switched .
these results justify our method for determining the thermodynamically stable point with @xmath10-ensemble . .
black dots represent the centers of the particles and grey lines denote networks of overlaps between the particles . from the instantaneous state of fig .
[ fig : frog2dmupts20snapshots](a ) , simulations are resumed after disconnecting the reservoir 1 or 2 , i.e. simulations are resumed with @xmath9 or @xmath8-ensemble .
( a ) : simulation result at @xmath110 mcs , after resuming the simulation with @xmath9-ensemble .
( b ) : simulation result at @xmath111 mcs , after resuming the simulation with @xmath8-ensemble . ] in simulations with @xmath9 and @xmath8-ensembles that are resumed from the instantaneous state shown in fig .
[ fig : frog2dmupts20snapshots](a ) . in these simulation runs ,
the number of the particles is @xmath112 with minor fluctuations . for reference ,
the result of the simulation with @xmath10-ensemble is also shown . ] here , one can recognize occurrence of a few defects and undulation of the string - like assemblies after switching the ensembles .
this should be attributed to the difference in the nature of the thermal fluctuations of a finite system between different ensembles .
for example , when a particle is removed from a perfect triangular crystal in @xmath10-ensemble and simultaneously the ensemble is exchanged to @xmath9-ensemble , long - lived defects are created in the crystal .
figure [ fig : frog2dmupts20snapshotsinvariousensembles](a ) shows this finite size effect . in the thermodynamic limit where the system size becomes infinitely large
, such a difference should vanish .
we also recognize a slight drop in @xmath89 in @xmath8-ensemble by approximately 2% .
this is also related to the finite size effect of the @xmath8-ensemble , which will be further discussed in section [ subsubsec : aldertransitionoftheoutercores ] .
the undulation of the string - like assembly shown in fig .
[ fig : frog2dmupts20snapshotsinvariousensembles](b ) corresponds to zigzag instability typically observed in convection rolls in a fluid slab when its natural periodicity is suddenly changed @xcite .
this zigzag instability is consistent with the slight drop in @xmath89 , _
i.e. _ a slight increase in the system size , in @xmath8-ensemble shown in fig .
[ fig : frog2dmupts20rd001e14p10t012innptmuvt - rho ] .
these simulation results show that the physical properties of the system , in conventional ensembles , sensitively depend on @xmath1 and the system box size while they do not strongly depend on @xmath1 in @xmath10-ensemble . at extremely low temperature ,
the repulsive square - step of @xmath73 becomes far higher than the thermal energy , @xmath113 . due to this extremely high potential energy barrier ,
the phase behavior of the system is almost identical to the behavior of hard particle systems with diameter @xmath70 if the system volume @xmath7 exceeds the close - packed volume ( area ) of the outer cores of the particles , denoted by @xmath114 ( @xmath115 in 2 dimensions ) @xcite .
therefore , crystalization of the hard particles at high density , called alder transition @xcite , occurs in our system at @xmath116 and low temperature @xcite . here
we simulate this triangular crystal of the outer cores of our colloids with the diamter @xmath70 at low temperature @xmath117 , where we fixed the parameters @xmath118 that are determined via the iterative refinement of @xmath3 and @xmath2 .
the stability of this thermodynamically stable point is discussed later in the present section .
-ensemble at @xmath119 and @xmath120 at @xmath121 mcs .
the initial system volume is set at @xmath122 .
black circles represent the inner cores of the particles and white ones denote the outer cores . ] , ( a ) , and @xmath1 , ( b ) and ( c ) , in the simulations with the @xmath10-ensemble at @xmath119 .
black lines denote the results of that are started from the initial condition with @xmath122 and @xmath120 and grey lines from that with @xmath123 and @xmath124 . ] simulation with the @xmath10-ensemble is started from an initial state with @xmath122 and @xmath120 .
we prepared this initial configuration by removing the particles from the equilibrium configuration with @xmath125 particles arranged on a homogeneous triangular lattice , which results in an inhomogeneous particle configuration .
such an inhomogeneous configuration is swiftly equilibrated in @xmath10-ensemble as is shown in fig .
[ fig : frog2dmupts20rd01e09p045t01inin1089iniv4778.38_009000000mcs ] . a snapshot of the system simulated with the @xmath10-ensemble , which is presented in fig .
[ fig : frog2dmupts20rd01e09p045t01inin1089iniv4778.38_009000000mcs ] , shows the defect - free triangular crystal .
temporary small defects due to thermal fluctuation are sometimes found in the system , whereas long - lived defects are absent .
time evolution of @xmath89 and @xmath1 is plotted in fig .
[ fig : frog2dmupts20rd01e09p045t01inin1089iniv4778.38inin1221iniv5647.18-rhoandn ] . small fluctuation in @xmath1 indicates the high stability of this defect - free crystalline state .
different initial particle configuration also results in a similar defect - free crystalline state with the same average @xmath89 .
time evolutions of @xmath89 and @xmath1 for simulation on the system with different @xmath79 and initial @xmath126 are also plotted in fig .
[ fig : frog2dmupts20rd01e09p045t01inin1089iniv4778.38inin1221iniv5647.18-rhoandn ] , which show the same defect - free structure .
this demonstrates the high stability of the defect - free crystalline state obtained via three - reservoirs method .
next , we compare physical characteristics of @xmath10-ensemble with these of the conventional ensembles . for this purpose , we perform simulations also with the conventional ensembles , _ i.e. _ @xmath9 and @xmath8-ensembles , with the same parameters @xmath119 . different from the @xmath10-ensemble case , in the present @xmath9-ensemble case , we have to manually tune the value of @xmath1 so that the perfect ordered equilibrium structure can be obtained .
for this reason , we perform a series of simulations with @xmath9-ensemble for all the values of @xmath1 within an interval @xmath127 , where @xmath128 and 1 mcs = @xmath1 simulation steps are fixed and the initial system volume is still @xmath122 .
the only parameter that is changed from the above @xmath10-ensemble simulation is @xmath1 .
typical examples of the time evolution of @xmath89 in these simulation runs are plotted in fig .
[ fig : frog2dnpts20p045t01n - rho ] . for reference ,
the result with the parameter optimized in @xmath10-ensemble , @xmath129 , is also plotted in this figure .
the system with this optimized parameter shows the defect - free triangular crystal and its value of @xmath89 is close to the results of the @xmath10-ensemble .
however , with any value of @xmath1 in the above interval , long - lived defects , mostly point defects , appear in the system , as is shown in fig .
[ fig : frog2dnpts20p045t01n1082n1089_020000000mcs ] .
moreover , the average value of @xmath89 changes with the change in @xmath1 slightly , nonmonotonically , and sensitively .
this behavior is different from the results of the @xmath10-ensemble .
this means that , although the external intensive variables @xmath11 are fixed , physical properties of the system in @xmath9-ensemble sensitively depend on the external extensive variable , @xmath1 . in @xmath9-ensemble at @xmath130 .
] in @xmath8-ensemble , the system size has to manually be tuned .
actually , we try such a manual tuning by performing the @xmath8-ensemble simulation with various system sizes in an interval @xmath131 .
aspect ratio of the system box is kept at a typical value for molecular simulation , @xmath80 .
simulation parameters that are changed from those of the corresponding @xmath10-ensemble simulation are @xmath7 and @xmath132 .
we assume that 1 mcs = @xmath79 simulation steps .
all the systems we have simulated show the defect - free triangular crystals , since point defects are directly removed by the particle insertion and deletion processes .
long - lived defects are not found in these simulation runs with @xmath8-ensemble . however , the average value of @xmath89 sensitively and nonmonotonically depends on the value of @xmath126 .
typical examples of time evolution of @xmath89 are given in fig .
[ fig : frog2dmuvts20rd01e09t01iniv - rho ] . although the intensive variables @xmath16 are specified by the reservoirs , physical properties of the system in @xmath8-ensemble significantly depend on the extensive variable @xmath7 . the data for the simulation with @xmath133 , which is the optimized value obtained in the three - reservoirs method , are also plotted by a black line .
we can confirm that the optimized value of @xmath89 is preserved during the simulation run .
this result justifies the simulation results obtained using three - reservoirs method .
obtained with @xmath8-ensemble at @xmath134 .
the black line denotes the result at @xmath133 , which is the optimized value obtained with three - reservoirs method . ] in the @xmath8-ensemble simulations , the particle density @xmath26 can be adjusted only discretely because of the discrete nature of @xmath1 , and therefore the behavior of @xmath26 in @xmath8-ensemble is rather abrupt and sensitive to the other parameters .
however , in this @xmath8-ensemble case , the point defects can rather easily be removed even in high - density states _
e.g. _ in a triangular crystal . on the other hand in @xmath9-ensemble , @xmath26 is changed by the change in the continuous dynamic variables , @xmath65 .
this results in a smaller change in the average @xmath26 compared with that in @xmath8-ensemble .
however , long - lived defects are frequently found in @xmath9-ensemble .
three - reservoirs method overcomes these disadvantages of @xmath8 and @xmath9-ensembles and , at the same time , it inherits the advantages of these ensemble methods . in @xmath10-ensemble ,
both point defects and line defects can be removed easily , and the extensive variables can finely and spontaneously be tuned .
the simulation results in section [ subsec : examinationof3-reservoirsmethod ] show that it is a tedious task to perform the manual tuning of @xmath1 and the system size in conventional ensembles , whereas these extensive variables are automatically and finely tuned in our @xmath10-ensemble . in the conventional ensembles where
at least one extensive variable is fixed , physical properties of the system are significantly dependent on the value of such a fixed extensive variable .
this illustrates that , for example , the equilibrium state of a system in @xmath9-ensemble depends on @xmath1 even though the intensive variables @xmath4 and @xmath3 are fixed . as another example , the equilibrium state of a system in @xmath22-ensemble changes with @xmath1 and @xmath65 even though the external intensive variables @xmath4 and @xmath12 are fixed .
the equilibrium state in conventional ensembles is specified by the parameter sets of the external extensive variables as well as the external intensive variables . in the present article
, we tentatively define the local equilibrium state as the equilibrium state at each parameter set of the external extensive variables with the fixed external intensive variables . in @xmath10-ensemble ,
however , the extensive variables are finely and spontaneously tuned and the most stable state over the local equilibrium states at the given @xmath135 is automatically obtained . in the present article , we tentatively call this equilibrium state in @xmath10-ensemble the global equilibrium state .
as these dependences of the physical properties on the external extensive variables in conventional ensembles are regarded as a finite size effect , the three - reservoirs method is a technique to remove the finite size effect of the conventional ensemble methods .
any local equilibrium states are , in the thermodynamic limit , identical to the global equilibrium state . in molecular simulations , physical quantities of the simulation system are defined through the ensemble average over the probability distribution of the microstates in the phase space .
the evaluation of the free energy of such a system , however , is equivalent to the evaluation of its partition function . as the partition function is not an averaged quantity over the ensemble , its evaluation demands a high dimensional integral over the whole phase space , resulting in unrealistically large amounts of computational cost . instead of evaluating the partition function directly , a derivative of the free energy with a control parameter , _
e.g. _ pressure in the canonical ensemble , is calculated for the sake of evaluating the free energy in standard molecular simulation @xcite .
this technique , which is usually called the thermodynamic integration , gives the free energy difference between two different thermodynamic states .
however , with this technique one can not go across a first - order phase transition .
for example , when the two thermodynamic states are located in a solid phase and in a fluid phase respectively , a first - order transition occurs in the middle of the integration path . due to
possible hysteresis at this transition point , forward and backward integration paths between the two states in general gives different values for the free energy difference .
this problem also affects the other free energy calculation techniques , _
e.g. _ histogram reweighting technique @xcite and the method of expanded ensembles @xcite . in order to overcome this difficulty
, we need to find appropriate reference states , whose free energy values are already known @xcite , _
einstein solid for the free energy calculation of crystals @xcite .
these reference states provide reversible integration paths .
if we can not find such reference states , we have to find integration paths that bypasses the first - order transition line , _
e.g. _ an integration path that is arranged with the help of artificially introduced external fields @xcite . after setting a reversible integration path using these techniques , we run simulations at a large number of state points along the integration path , and integrate the derivative of the free energy along this path .
in addition to the discretization error in the integration along the path , occurrence of defects in the system also affects the accuracy of the free energy evaluation of the ordered structures .
here we propose , based on euler equation in thermodynamics , a convenient method for the free energy calculation using the systems connected to the three reservoirs .
the entropy of the system per particle , denoted by @xmath136 , satisfies euler equation , @xmath137 where @xmath138 denotes the total kinetic energy of the system , @xmath53 denotes the total potential energy , and @xmath139 is the ensemble average at @xmath140 . in the right - hand side of this equation , the 3 intensive variables , @xmath4 , @xmath3 , and @xmath2 of the system , relax to the equilibrium values that are equal to those given by the reservoirs .
@xmath141 is determined via the equipartition theorem , _ e.g. _
@xmath142 for monatomic molecules .
the ensemble averages @xmath143 and @xmath144 can directly and readily be evaluated through the simulation runs of the three - reservoirs method .
therefore , according to this euler equation , @xmath145 is determined from our simulation at one state point @xmath140 , which means that this free energy evaluation method requires far smaller amounts of computation than the other standard methods do .
the free energy of the system , _
e.g. _ helmholtz free energy , gibbs free energy , and grand potential , is also obtained from this result in a similar manner . since
our entropy and free energy calculation method is free from any thermodynamic integration paths , the first - order phase transition does not affect our evaluation method .
in addition , with the three - reservoirs method we can easily eliminate the defects which is the main origin of the error in the free energy evaluation for the ordered phases .
this raises the accuracy of the evaluation . in our molecular simulation
, @xmath2 includes the thermal de broglie wave length @xmath77 , as is discussed in appendix [ sec : molecularmctechniqueinmuvtandnpt - ensembles ] , _ i.e. _ eq . . @xmath146 and @xmath147 in @xmath77
cancel when the free energy difference between two different state points is calculated .
therefore , we do not have to take care of these @xmath146 and @xmath147 .
although the above method based on euler equation is , in principle , applicable to simulations with the other ensembles , _
e.g. _ @xmath22 and @xmath9-ensembles , one has to measure intensive variables @xmath3 and/or @xmath2 as ensemble averaged values .
such a procedure requires a computationally expensive analysis .
for example , the measurement of @xmath2 , _
i.e. _ gibbs free energy per particle , demands a vast amount of simulation , especially in high particle density regions @xcite .
however , in the @xmath8-ensemble or @xmath10-ensemble , we do not have to evaluate @xmath2 because it is already given by the reservoir .
same is true for the evaluation of the pressure @xmath3 . as a result , with the use of the @xmath10-ensemble
, we can skip tedious evaluations of the intensive variables because all the essential intensive variables @xmath2 , @xmath3 , and @xmath4 are already specified by the reservoirs .
in addition , metastable structures and defects , which frequently appear in conventional ensembles , affect the results of this entropy and free energy evaluation .
when the system is in metastable states or outside the global equilibrium state , the intensive variables given from the reservoirs are inconsistent with the values of these variables in the simulation system .
this means that both @xmath2 and @xmath3 need to be analyzed in the simulation rather than to use the specified value by the reservoirs .
however , as the defects are quickly eliminated and the system reaches the global equilibrium in the simulations with the @xmath10-ensemble , the evaluation of the free energy is free from the above problem associated with the metastable states and the local equilibrium states . when one tries to construct the equilibrium phase diagrams using conventional simulation methods , candidates for the equilibrium structure have to be chosen prior to the simulation and the free energy of each candidate should be measured and compared with each other with high accuracy , for example , with a typical error level @xcite of @xmath148 to @xmath149 . a variety of phases , _
e.g. _ crystals , the string - like assemblies , and the other ordered and disordered structures should be considered the candidates for the equilibrium phase . in actual calculation
, we empirically select some of these potential candidates and discard the others .
however , we can not deny the possibility that we have discarded the true equilibrium structure in this selection process .
in addition , there is another possibility that the equilibrium structure is a totally new structure which has not been discovered yet . with the use of @xmath10-ensemble ,
however , one can obtain the global equilibrium structure directly as a result of the fine tuning of the extensive variables .
therefore , the above - mentioned problem encountered in the construction of the phase diagram using the standard ensembles can be avoided when we use @xmath10-ensemble .
we have studied thermodynamics , statistical mechanics , and molecular mc simulation algorithms of @xmath10-ensemble .
guggenheim formally introduced boltzmann factor ( statistical weight ) of this ensemble with an assumption that the averages of extensive variables can be determined @xcite . however
, this assumption contradicts the indetermination of extensive variables in @xmath10-ensemble .
in addition , other characteristics of this ensemble were totally absent in guggenheim s discussion and other early works @xcite .
these early researchers concentrated on the formalism , i.e. mathematical aspects of the partition function of this ensemble , only at the thermodynamically stable point @xcite .
physical aspects of the ensemble have not seriously been discussed . in the present work ,
we have shed light on these problems and have discovered thermodynamic and statistical mechanical characteristics of @xmath10-ensemble , _
e.g. _ the thermodynamically stable point , thermodynamic degrees of freedom , thermodynamics outside the thermodynamically stable point , maximization of entropy density , quick equilibration due to shortcuts in the phase space , _ etc .
_ we have also shown that @xmath10-ensemble is built as a combination of the 3 ensembles each of which is combined to 2 reservoirs , _
i.e. _ @xmath9 , @xmath8 , and @xmath18-ensembles .
the 3 corresponding free energy densities , _
i.e. _ gibbs free energy per particle , grand potential per volume , and @xmath19 , are simultaneously minimized in @xmath10-ensemble , rather than the thermodynamic potential of @xmath10-ensemble , denoted by @xmath47 .
we have proposed a molecular mc simulation method based on @xmath10-ensemble ( three - reservoirs method ) @xcite .
we can show that this three - reservoirs method gives a physically acceptable ensemble , which allows us to trace the physical trajectories in the phase space . since three - reservoirs method is built as a combination of conventional @xmath9 and @xmath8-ensembles , programming is lighter than other advanced techniques .
in addition , the thermodynamically stable point is determined according to gibbs - duhem equation in a short computational time .
these features mean that three - reservoirs method requires a small amount of preparation and that we can quickly start production simulation runs , although other advanced techniques demand a large quantity of complicated preparation , _
e.g. _ advanced programming , the precise adjustments of the artificial weights necessary for multicanonical technique , and accurate free energy measurement essential for the expanded ensemble technique @xcite .
these advantages over other simulation techniques facilitate and reduce the total work flow of our three - reservoirs method compared with the conventional methods such as @xmath9 , @xmath8 , or multicanonical method .
furthermore , only with the three - reservoirs method , we can _ simultaneously _ and _ automatically _ tune the number of particles @xmath1 and the system size to obtain the equilibrium ordered state . this unique advantage of the three - reservoirs method could enhance the understanding of those systems that were obtained via the other standard simulation techniques .
for example , for perfect crystals , both @xmath1 and @xmath0 must be integer multiples of the unit structure of the crystal structure , which is in general not known _ a priori_. in principle , by measuring the free energy density of various structures at each set of @xmath1 and @xmath0 ,
these extensive parameters can manually be tuned in simulation of @xmath9-ensemble , @xmath8-ensemble , multicanonical ensemble , _ etc .
_ in practice , however , this manual tuning requires much computational effort . on the other hand , with the use of our three - reservoirs method
, we can automatically achieve such optimization . for a solid at finite temperature @xmath4 ,
the system with the @xmath10-ensemble reaches a globally - anisotropic defect - free ordered state as the equilibrium state by crossing the metastable states through the shortcuts in the phase space due to the additional degrees of freedom . on the other hand , in conventional ensembles , physical properties of the system
sensitively and discretely depend on @xmath1 and/or @xmath0 even though external intensive variables are fixed .
this results in a requirement of the manual tuning of these extensive variables to obtain the global equilibrium state .
these results illustrate that our method can be applied to a variety of physical systems for the sake of studying ordered structures in equilibrium at finite @xmath4 , _
e.g. _ lamellae composed of diblock copolymers , smectic phase of liquid crystals , and fluid bilayer membranes .
three - reservoirs method can also be applied to numerical calculation of the equation of state that relates @xmath2 , @xmath3 , and @xmath4 .
we have also shown that the entropy and the free energy can quickly be evaluated in @xmath10-ensemble based on euler equation .
this feature is essentially important in the construction of the phase diagram of condensed materials .
the authors wish to thank professor komajiro niizeki and mr masatoshi toda for helpful suggestions and discussions .
we also thank the anonymous referees for their valuable suggestions and comments .
this work is partially supported by a grant - in - aid for science from the ministry of education , culture , sports , science , and technology , japan .
in this appendix , we show molecular mc simulation technique @xcite for single component systems based on the @xmath8 and @xmath9-ensembles .
we use the following notations : @xmath54 denotes the spatial coordinates of the particle @xmath55 , @xmath147 mass of a particle , @xmath113 the thermal energy , @xmath53 the potential energy of the system , and @xmath146 planck s constant .
the chemical potential of a system consisting of real particles in the canonical ensemble ( @xmath22-ensemble ) is given by , @xmath150 where @xmath151 and @xmath152 here , @xmath153 denotes the chemical potential of an ideal gas in @xmath22-ensemble and @xmath77 is called the thermal de broglie wave length .
the quantity @xmath154 denotes the excess chemical potential that is originated from the interaction between the real particles @xcite .
simulation results in @xmath8-ensemble are independent of @xmath77 .
this means that @xmath77 only appears in the chemical potential @xmath2 at the reference point in @xmath8-ensemble simulation , which is discussed in section [ subsec : mcsimulationmethodinmuvtensemble ] . in the present section ,
we discuss the mc method of a single component system in the grand canonical ensemble , which is also called @xmath8-ensemble because @xmath2 , @xmath7 , and @xmath4 are kept fixed . in simulation with the @xmath8-ensemble @xcite , particles are inserted into and deleted from the system in addition to metropolis trial displacement of particles . in one simulation for this ensemble
, these steps are included in : a. with probability @xmath60 , trial particle insertion into the system b. with probability @xmath60 , trial particle deletion from the system c. with probability @xmath155 , trial displacement based on metropolis algorithm , _
i.e. _ perturbation to one particle is chosen , where @xmath63 is a constant fixed in an interval @xmath156 .
algorithms of trial particle insertion and deletion are discussed in the following .
we assume that one particle is inserted into the system that is composed of @xmath1 particles .
the position of this inserted particle @xmath157 is chosen uniformly over the system box .
the coordinates of the @xmath1 particles , @xmath158 , are fixed during this particle insertion step .
the state after the insertion , _
i.e. _ the state of @xmath159 particles , is accepted as a new state with the probability , @xmath160
\right ) , \\ & u^{\text{ins}}_{\text{excess } } : = u \left ( \rvector_1 , \dots , \rvector_{n+1 } \right ) - u \left ( \rvector_1 , \dots , \rvector_n \right ) . \notag\end{aligned}\ ] ] a function @xmath161 returns the smaller of two arguments , @xmath162 and @xmath163 .
if the trial insertion is rejected , the state before the insertion is kept for the next simulation step .
we assume that one particle is randomly chosen and attempts to be removed from the present system composed of @xmath1 particles .
this chosen particle , denoted by index @xmath31 , is removed from the system with the probability , @xmath164 } \right ) , \\
u^{\text{del}}_{\text{excess } } : = u \left ( \rvector_1 , \dots , \rvector_n \right ) - u \left ( \rvector_1 , \dots , \rvector_{j-1 } , \rvector_{j+1 } , \dots , \rvector_n \right ) .
\notag\end{gathered}\ ] ] if the trial deletion is rejected , the state before the deletion is kept for the next simulation step .
when @xmath165 is substituted for @xmath2 in eqs . and , these acceptance criteria are independent of @xmath77 .
this illustrates that simulation results in @xmath8-ensemble are free from the actual value of @xmath77 .
therefore , @xmath77 appears only in the chemical potential @xmath2 at the reference point in the @xmath8-ensemble simulation @xcite .
mc simulation method in isobaric - isothermal ensemble @xcite , also called @xmath9-ensemble , is briefly summarized in this section .
in addition to @xmath1 and @xmath4 , the system pressure , denoted by @xmath3 , is given from the outside , in this ensemble .
in simulation of @xmath9-ensemble , the system size is changed , in addition to the trial move of particles . in one simulation step , we include the following steps : a. with probability @xmath61 , trial system size change b. with probability @xmath166 , trial displacement based on metropolis algorithm , _ i.e. _ perturbation of one particle is chosen , where @xmath61 is a constant fixed in an interval @xmath167 .
algorithms of trial system size change are discussed in the following .
to change the system box size , we use the following 4-step algorithm .
we assume that the system box size , denoted by @xmath0 , is changed to a new system size , @xmath168 .
unlike the conventional mc simulations in @xmath9-ensemble based on mcdonald s method @xcite , each element of the system size @xmath0 is independently changed in our algorithm . 1 .
the new system size is chosen , @xmath169 where @xmath82 is a small constant length , @xmath170 , and @xmath171 are random numbers uniformly distributed over an interval @xmath172 , and @xmath173 denotes the volume of the new system box .
coordinates of all the particles before the system size change , @xmath174 , are homogeneously scaled , @xmath175 this is new particle coordinates after the change .
3 . the potential energies of the systems before and after the system size change , @xmath176 and @xmath177 respectively , are calculated .
the new system size and particle coordinates are accepted with the probability , @xmath178
\right ) . \end{gathered}\ ] ] if this trial system size change is rejected , the state before the trial is kept for the next simulation step .
the ensemble average at each @xmath1 in @xmath10-ensemble is equivalent to the average in @xmath9-ensemble at the same @xmath1 .
we illustrate this in the present section .
we assume that a finitely long simulation run is performed in @xmath10-ensemble at a thermodynamically stable point .
after the equilibration of the simulation system , a finitely large number of microstates of the system are visited .
unnormalized boltzmann factor of the system is denoted by @xmath179 , where the suffix @xmath55 represents these microstates .
@xmath180 is kronecker delta . @xmath181 .
after the equilibration of the system , the ensemble average of an intensive physical quantity , @xmath182 , obtained in the present simulation run is : @xmath183 } { \sum_i \delta_{n_i , n } \frac{1 } { \varlambda^{3n } \ , n ! } \exp \left [ - \beta \left ( p v_i - \mu n + u_i \right ) \right ] } \notag \\ & \quad = \sum_{n=0}^{\infty } f(n ) \frac { \sum_i \delta_{n_i , n } a_i \exp \left [ - \beta \left ( p v_i + u_i \right ) \right ] } { \sum_i \delta_{n_i , n } \exp \left [ - \beta \left ( p v_i + u_i \right ) \right ] } \notag \\ & \quad \cong \sum_{n=0}^{\infty } f(n ) \ , \langle a \rangle _ { t , p , n},\end{aligned}\ ] ] where the summation @xmath184 runs over the microstates visited in the simulation , and @xmath185 is defined as , @xmath186 and @xmath187 denotes the ensemble average of @xmath182 in @xmath9-ensemble .
@xmath185 denotes the occurrence probability of @xmath1 in the simulation run .
an expression similar to eq . also holds for the ensemble average of @xmath182 in @xmath8-ensemble .
this result illustrates that the short - time average of @xmath182 in @xmath10-ensemble is approximated by the ensemble average in @xmath9-ensemble or @xmath8-ensemble .
when the finite size effect of the system , which has been discussed in sections [ subsec : examinationof3-reservoirsmethod ] and [ subsec : globalequilibrium ] , is small , @xmath187 is dependent on @xmath11 and independent of @xmath1 .
therefore , eq . yields @xmath188 .
a similar relation , @xmath189 , also holds , where @xmath190 denotes the ensemble average of @xmath182 in @xmath8-ensemble .
therefore , @xmath191 .
53ifxundefined [ 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 10.1038/nmat2404 [ * * , ( ) ] link:\doibase { 10.1209/epl / i2005 - 10288 - 6 } [ * * , ( ) ] link:\doibase 10.1126/science.1142614 [ * * , ( ) ] link:\doibase { 10.1038/nsmb.1455 } [ * * , ( ) ] link:\doibase { 10.1039/b901657k } [ * * , ( ) ] _ _ , http://webdoc.sub.gwdg.de/diss/2010/norizoe/norizoe.pdf[ph.d .
thesis ] , , ( ) , + http://webdoc.sub.gwdg.de/diss/2010/norizoe/ + \doibase doi : 10.1016/0370 - 2693(91)91256-u [ * * , ( ) ] @noop _ _ ( , , ) \doibase doi : 10.1016/0021 - 9991(77)90121 - 8 [ * * , ( ) ] @noop _ _ ( , , ) link:\doibase 10.1063/1.462133 [ * * , ( ) ] @noop ( ) , link:\doibase 10.1063/1.1750386 [ * * , ( ) ] \doibase doi : 10.1016/0031 - 8914(50)90070-x [ * * , ( ) ] @noop _ _ ( , , ) \doibase doi:10.1080/00268975900100021 [ * * , ( ) ] link:\doibase 10.1021/jp951819f [ * * , ( ) ] @noop _ _ ( , , ) link:\doibase 10.1103/physrevlett.45.1196 [ * * , ( ) ] link:\doibase 10.1063/1.328693 [ * * , ( ) ] link:\doibase 10.1016/0036 - 9748(83)90283 - 1 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physreve.80.061403 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.99.248301 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1063/1.2965880 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1063/1.3006065 [ * * , ( ) ] \doibase http://doi.acm.org/10.1145/272991.272995 [ * * , ( ) ] \doibase http://doi.acm.org/10.1145/146382.146383 [ * * , ( ) ] \doibase http://doi.acm.org/10.1145/189443.189445 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.65.851 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevlett.63.1195 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1063/1.448024 [ * * , ( ) ] link:\doibase 10.1103/physreve.51.r3795 [ * * , ( ) ] link:\doibase { 10.1039/b818111j } [ * * , ( ) ] link:\doibase 10.1063/1.1734110 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1063/1.1741967 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1063/1.462658 [ * * , ( ) ] and , eds .
, @noop _ _ ( , , ) @noop _ _ ( , , ) @noop * * , ( ) link:\doibase 10.1063/1.1667938 [ * * , ( ) ] link:\doibase 10.1063/1.1673047 [ * * , ( ) ] @noop * * , ( ) | in conventional molecular simulation , metastable structures often survive over considerable computational time , resulting in difficulties in simulating equilibrium states . in order to overcome this difficulty ,
here we propose a newly devised method , molecular monte carlo simulation of systems connected to three reservoirs : chemical potential , pressure , and temperature .
gibbs - duhem equation thermodynamically limits the number of reservoirs to 2 for single component systems .
however , in conventional simulations utilizing 2 or fewer reservoirs , the system tends to be trapped in metastable states . even if the system is allowed to escape from such metastable states in conventional simulations , the fixed system size and/or the fixed number of particles result in creation of defects in ordered structures
this situation breaks global anisotropy of ordered structures and forces the periodicity of the structure to be commensurate to the system size .
here we connect the such three reservoirs to overcome these difficulties .
a method of adjusting the three reservoirs and obtaining thermodynamically stable states is also designed , based on gibbs - duhem equation . unlike the other conventional simulation techniques utilizing no more than 2 reservoirs , our method allows the system itself to simultaneously tune the system size and the number of particles to periodicity and anisotropy of ordered structures .
our method requires fewer efforts for preliminary simulations prior to production runs , compared with the other advanced simulation techniques such as multicanonical method . a free energy measurement method , suitable for the system with the three reservoirs , is also discussed , based on euler equation of thermodynamics .
this measurement method needs fewer computational efforts than other free energy measurement methods do . |
the existence of black holes is accepted as a fact of the cosmos by many astrophysicists .
the evidence for the existence of massive objects that may be compact enough to be black holes is strong , however there is as yet no direct evidence of any mass that is contained within its schwarzschild radius and even that would not necessarily prove the object to be a black hole .
supermassive compact objects have been found in the nuclei of most galaxies and objects of stellar mass are abundant within our own and other galaxies . while these compact objects are commonly called black holes no compelling observational evidence of an event horizon , the quintessential feature of a black hole , has yet been found .
it has been pointed out previously ( abramowicz , kluzniak & lasota 2002 ) that it would be very difficult to prove the existence of event horizons if objects with large surface gravitational redshift exist .
extreme redshifts could make such objects nearly as dark as a black hole ( see appendix a ) . if not different from a black hole in some other way , it would not seem to matter much whether or not the black hole candidates are actual black holes . on the other hand ,
nearly every object of stellar mass or greater is known to be magnetic to some degree .
whatever their origins , magnetic fields and their synchrotron radiations are ubiquitous among the black hole candidates . instead of possessing event horizons ,
some of these compact objects might be found to be both intrinsically magnetic and highly redshifted . since magnetic moments
are forbidden attributes for black holes , there would be important consequences for astrophysics if some of the black hole candidates were found to possess them .
already some evidence for the existence of magnetic moments has been presented for both gbhc and agn ( robertson & leiter 2002 , 2004 , 2006 schild , leiter & robertson 2006 , 2008 ) .
others have reported evidence for very strong magnetic fields in gbhc , e.g. , a field in excess of @xmath0 g has been found at the base of the jets of grs 1915 + 105 ( gliozzi , bodo & ghisellini 1999 , vadawale , rao & chakrabarti 2001 ) . these findings provided motivation for the development of a fully general relativistic model of a gravitationally compact , intrinsically magnetic , eternally collapsing object that we call a meco ( leiter & robertson 2003 , mitra 2006a , b , c , robertson & leiter 2003 , 2004 , 2006 , hereafter rl03 , rl04 , rl06 ) .
a meco avoids rapid collapse to a black hole state by radiating away its mass - energy at an eddington limit rate .
it is characterized by both an extreme redshift and a strong intrinsic magnetic moment ( see appendix a - d ) .
the large redshift accounts for the low quiescent surface luminosities of gbhc and agn and their extremely long ( many hubble times ) radiative lifetimes .
the meco model has been shown ( mcclintock , narayan & rybicki 2004 ) to be consistent with the lower limit on the quiescent emission from the gbhc xte j1118 + 480 .
we show here that it is also consistent with the low luminosity of sgr a*. the meco model for disk accreting gbhc and agn accounts for the existence of their low / hard and high / soft x - ray spectral states .
the rotating magnetic moments provide a robust universal magnetic propeller mechanism for spectral state switches ( ilarianov & sunyaev 1975 , campana et al . , 1998 , 2002 , rl02 , rl03 ) .
the high / soft @xmath1 low / hard transition marks the start of a magnetic propeller regime , the end of accreting plasma being able to penetrate inside the corotation radius and the beginning of a low state jet outflow .
the meco model correctly predicts ( rl04 ) that this transition occurs at about @xmath2 of eddington limit luminosity for both gbhc and agn .
the luminosities at the transition and in quiescence have permitted the determinations of the magnetic moments and spin rates for meco - gbhc ( rl02 , rl06 , and eqs 6 & 10 , table 1 ) .
spectral state switches are common to dwarf novae , the neutron stars and gbhc of low mass x - ray binary systems and agn .
they are signatures of intrinsic magnetism , however they have not yet been accepted as such because it is generally believed that the black hole candidates are actual black holes and can not possess magnetic moments .
meco interacting with accretion disks , as shown in rl04 , can produce low / hard state jet outflows with correlated radio and x - ray emissions ( @xmath3 ) in accord with observations ( gallo , fender & pooley 2003 , markoff et al .
2003 , falcke , krding & markoff 2004 , maccarone , gallo & fender 2003 ) . to account for the similar correlations for agn and neutron star or dwarf nova systems
, it was necessary to explore the mass scaling relations for the meco model ( rl04 ) .
these meco mass scaling relationships , including those associated with the magnetic moments and spin rates that have been able to account for the radio / x - ray luminosity correlations of agn and gbhc ( rl04 ) , are listed in the right hand column of table 1 .
it is a remarkable fact that no further adjustments to these meco mass scaling parameters are needed for the successful application of the meco model to the case of sgr a*. they also account for the newly discovered quasar accretion structures ( schild , leiter & robertson 2006 , 2008 ) revealed by microlensing observations of the quasars q0957 + 561 and q2237 + 0305 .
the observed structures are consistent with the strongly magnetic meco model but do not accord with standard thin disk models of accretion flows into a black hole . since even the nearest gbhc are much too small to be resolved in the detail shown by the microlensing techniques which were used in the study of q0957 and q2237 , sgr a * is likely the only remaining black hole candidate for which resolved images might reveal whether or not it possesses a magnetic moment .
for this reason it is important that it be tested . because much of the spectrum of sgr a * appears to originate in synchrotron - cyclotron radiation
, a serious test of the hypothesis that sgr a * might possess an intrinsic magnetic moment will require global solutions for the magnetic field and electron density distributions for some kind of accretion flow . while this will require future detailed simulations that are beyond the scope of the present work , we show in this paper that analytic methods can be used to give here a good accounting of the physical properties of the radio / nir and x - ray spectral characteristics , luminosities and polarizations that have been observed for sgr a*. we find that we must consider bondi accretion for a quiescent meco and
then show that we can reconcile the observed low luminosity of sgr a * with the expected bondi accretion flow rate . for the discussion to follow and also for the convenience of readers who might wish to relate meco properties to observations of other objects , we have tabulated a number of useful relations in table 1 .
many of the parameters are given in terms of quiescent x - ray luminosity @xmath4 , or the luminosity , @xmath5 , at the transition high / soft @xmath1 low / hard state since these are often measurable quantities .
some new developments , minor corrections and general features for the meco model are presented in appendixes a - e. [ cols="<,<,<",options="header " , ] + as shown in appendix e , the luminosity produced in the axial inflows to distance @xmath6 from the meco is ( eq .
32 ) @xmath7 for @xmath8 , this yields @xmath9 . to @xmath10 and @xmath11 ,
the luminosity in the polar flow would be @xmath12 . for comparison , the spectrum reported by an et al .
( 2005 ) shows a luminosity of @xmath13 to a frequency of 560 ghz .
the discrepancy between these results occurs primarily in the lower frequencies .
the flux at @xmath14 consistent with eq .
2 is about one order of magnitude below the observed radio spectral trend .
this should be expected because most of the radio spectrum would be generated in the outflow , but eq.2 should be fairly accurate in the nir which is produced only in the inflow . at frequencies above about @xmath15 ,
the axial flow is optically thin and dominated by cyclotron fundamental frequencies that can be associated with position in the dipole magnetic field ; i.e. , @xmath16 .
thus inside @xmath17 , it is also shown in appendix e that luminosity varies with frequency along the axial flows such that ( appendix e eq .
33 ) @xmath18 thus the spectral index in the optically thin ir / nir would be @xmath19 for the conical inflow model .
the average of four measurements reported in bn06 in the nir wavelength range from @xmath20 is @xmath21 . using a distance to sgr a * of @xmath22 , the average flux calculated here from eq .
3 for the same four wavelengths is @xmath23 .
this result and the variability shown by measurements at the position of sgr a * strongly suggests that the nir spectrum originates in a variable accretion flow that is consistent with the meco model . while the flux calculated here
is somewhat larger than the lowest flux used in bn06 to constrain an accretion rate to an assumed hard surface , it should be remembered that variability in the accretion flow would produce opportunities to observe both higher and lower luminosities .
it should also be noted that the calculated luminosity is sensitive to the electron temperature , which is poorly constrained . on balance , the agreement between meco calculated and averaged observed luminosities seems reasonable .
in addition to sensitivity to the electron temperature , the luminosity generated in the inflow depends on the meco spin rate which then determines the size of the corotation radius ( @xmath24 ) and hence the fraction of the bondi flow that can reach the central meco .
we should expect the spin rate to be influenced by angular momentum transported to the meco by its accretion environment .
nevertheless , the similar corotation radii ( @xmath25 ) of disk accreting gbhc and agn are necessary for consistency with their mass scale independent radio cutoffs at @xmath2 of eddington limit luminosity . given this mass scale invariant property ,
it is not surprising that they seem to show little variability in spins that scale as @xmath26 .
while we must admit that it is fortuitous that a similarly scaled spin provides reasonable results for the meco model of sgr a * , it also suggests that it may have had an accretion history similar to that of other agn . for an optically thick slice in the inflow at frequencies below @xmath27
, the cyclotron spectrum would produce luminosity from an axial region of thickness @xmath28 that would be proportional to @xmath29 times lateral surface area @xmath30 .
the correlation of frequencies with position , @xmath6 , in the inflow is surely weaker in the broad radio spectral band , but if we still associate @xmath6 with frequency as @xmath31 , the luminosity from the band would then be proportional to @xmath32 , hence the spectral index for the optically thick band between @xmath33 and @xmath34 would be @xmath35 , compared to the reported @xmath36 ( an et al .
on the other hand , if viewed looking parallel to the equatorial plane , the luminosity in the outflow could be calculated in the same way in a series of slices of thickness @xmath28 with outflow radius proportional to @xmath37 ( @xmath38 ) to obtain the same spectral index .
both radio and nir spectral indexes are sensitive to the shape of the conical inflow pattern .
dipole magnetic field lines have some curvature that would slightly widen the cone at the top . if the cone
" radius were proportional to @xmath39 with @xmath40 , it would increase the overall luminosity and spectral index calculated for the nir and decrease the spectral index in the optically thick region below the spectral peak .
this suggests that some refinements of our model might be expected to produce closer agreement with observations but that is a task for another time .
electrons become mildly relativistic with lorentz factors of @xmath41 for small @xmath6 .
some x - ray emission would arise from compton scattering within @xmath42 in the polar inflow .
electron densities are also large enough in this region to produce some thermal brehmsstrahlung .
the calculation of synchrotron self - compton ( ssc ) contribution can be started by differentiation of eq . 2 to obtain the luminosity contribution from axial thickness @xmath28 as @xmath43 , where @xmath44 is the rate of production of synchrotron photons of average energy @xmath45 within @xmath28 . in passing through distance @xmath46 , on average , these photons will experience @xmath47 collisions , where @xmath48 is the electron density and @xmath49 the thompson cross - section
. the average energy gained per collision would be @xmath50 .
substituting for @xmath48 , @xmath51 and @xmath52 as a function of @xmath6 , and integrating over @xmath6 , gives the x - ray self - synchrotron luminosity as @xmath53 which provides about @xmath54 to @xmath55 .
for a gaunt factor of @xmath56 for small z , thermal brehmsstrahlung contributions can be calculated similarly .
the emissivity obtained is @xmath57 .
integration over the axial cone volume and all frequencies yields a thermal ( primarily x - ray ) luminosity of @xmath58 this provides another @xmath59 to @xmath55 .
though both calculated x - ray luminosities might be increased somewhat by considering a flared cone " , their combined contributions should still fall well below the @xmath60 observed in the 0.5 - 7 kev band ( baganoff et al .
nevertheless , both depend on the square of electron density which could enhance their contributions to the luminosity in flares relative to synchrotron luminosity . these x - ray luminosity variations in flares would be strongly correlated with the nir synchrotron luminosity variations and without time delays since both originate in the same population of electrons .
a substantial fraction of the quiescent x - ray luminosity appears to come from a spatially extended source ( baganoff et al . 2003 ) .
there is a considerable volume in the meco magnetosphere in which temperatures and densities would be high enough to produce thermal x - rays . within the rough bounds of the magnetosphere outside the corotation radius there is a volume of @xmath61 . using @xmath62 as an average radius in the bondi flow , we estimate the corresponding electron temperature to be @xmath63 and the electron density to be @xmath64 . for these parameters , a gaunt factor of 3 and the @xmath65 bandwidth from 0.5 - 7 kev
, we find an average thermal brehmsstrahlung emission rate of @xmath66 .
multiplying by the magnetosphere volume , we obtain @xmath67 , which is reasonably close to the observed quiescent x - ray luminosity of sgr a*. the axial inflows constitute such a small fraction of the bondi accretion rate that their contributions to the spectrum below @xmath68 are much smaller than the radiation from the much larger outflow
. we do nt know how much the outflows might contribute in harmonic frequencies above @xmath69 , but since the luminosity calculated for the axial flow did not account for the observed luminosity or flux to a frequency of @xmath34 , there is a significant contribution from the outflow that has not been considered .
accurately quantifying the radio luminosity will require knowledge of the plasma and magnetic field distributions , as previously mentioned .
all that we can do here is set an upper limit on what might be produced in the outflow . for an accretion rate of @xmath70
, electrons would flow outward at a rate of @xmath71 .
if they each had the energy they could extract from the protons while reaching the corotation radius , there would be about @xmath72 available for radiation .
the actual radiation rate would probably be much smaller as few of the electrons that would flow out would reach such small distances from the meco . in this regard ,
in reaching a steady state , the magnetic propeller eventually must drive the outflow beyond the bondi radius , cutting off a large fraction of the flow into the bondi sphere and eventually producing a large torus where the outflow stops beyond the bondi radius and temperature @xmath73 at the bondi radius . ] .
a luminosity of a few times @xmath74 would seem to be entirely reasonable .
the meco - bondi flow pattern has two major regions of strong , ordered magnetic fields , with the strongest fields along the polar axis . with the nir generated within the the polar inflow
, its linear polarization direction would be essentially perpendicular to the axis .
the equatorial outflow would produce magnetic field lines stretched generally radially along the outflow .
these field lines would be perpendicular to the magnetic axis and there are several possibilities for what might be observed , depending on the orientation of the magnetic axis and equatorial plane relative to an observer .
yusef - zadeh et al . ( 2006 ) showed some evidence of outflow from sgr a*. the apparent flow was in a generally wsw direction from sgr a*. proper motion studies have provided confirmation ( muzic et al . , 2007 ) of a general , uncollimated outflow .
gravity , a recently proposed high resolution interferometric imaging system ( eisenhauer 2008 ) , should be capable of revealing the nature of the plasma flows in the vicinity of sgr a * in unprecedented detail .
if what has been observed at radio frequencies would be a view into the meco - bondi equatorial outflow , we would see the limb brightened regions of the outflow extending wsw ( and ene ) against the plane of the sky , with magnetic field lines stretched out along the flow .
linear polarizations from the outflow would then be generally perpendicular to the radio luminosity and in the sse - nnw direction .
a sketch of the way this might appear is shown in figure 2 .
the radio polarization has been observed at @xmath75 ( bower et al .
2003 ) and clearly displayed , aligned perpendicular to the wsw - ene direction of the limb brightened part of the outflow .
( see fig .
2.1 , http://www.cfa.harvard.edu/sma/newsletter/smanews_21dec2006.pdf ) .
weaker polarization would be produced further out in the flow where the field is weaker and fundamental cyclotron frequencies would be lower .
in addition to the polarization direction at radio frequencies matching expectations for an outflow , strong polarization has been observed in the nir with a direction roughly 90 degrees ( ene ) different from that seen at radio frequencies ( eckart et al .
as noted , this is consistent with the polarization that should be seen perpendicular to the field lines in the inflow if viewed at high inclination to the inflow axis .
a view with this orientation would be a view fairly directly into the equatorial outflow and directly down the magnetic field lines of the outflow . because of the dipole nature of the meco magnetic field , a view directly into the equatorial plane would have outgoing magnetic field lines on one side of the the plane and ingoing field lines on the other .
outflowing electrons would produce left circular polarization from outflow on one side of the equatorial plane and right circular polarization from the other .
since it is unlikely that the thickness of the outflow would be resolved , there would be a net circular polarization only if viewed at some inclination to the outflow equatorial plane .
this could occur while still having a large inclination relative to the inflow axis .
electrons spiraling outward on field lines directed toward us would produce a ( negative ) left circular polarization .
the circular polarization ( @xmath76 ) of sgr a * ( bower 2000 ) is unusually strong for an agn and approximately double the degree of linear polarization at the same frequency .
it seems unlikely that it would originate in a depolarizing medium .
while it exhibits short - term variability , it has maintained a stable ( negative ) sense for over 25 years . in this context
it is important to note that the observation of stable directions , with some variability in flares , for both linear and circular polarizations are unique observable features of the meco model for sgr a*. all of these properties are determined by the strength of the intrinsic magnetic field of the meco within the center of sgr a*. the observed polarizations are consistent with a view close to the equatorial plane of the outflow and across the axial inflow as shown in figure 2 .
for this orientation we should see an increasingly wide outflow zone at longer radio wavelengths . inside any surface of constant magnetic field strength cutting through the outflow , we would expect optically thick cyclotron radiant flux of @xmath77 to be generated .
the apparent size of its limb brightened photosphere image in the equatorial outflow would be roughly throughout . ]
proportional to @xmath78 and ellipsoidal as has been observed ( bower et al .
2004 ) . the major axis of the ellipsoid is apparently aligned with the direction of outflow ( muzic et al .
2007 , yusef - zadeh et al . 2006 , bower et al . 2004 ) and perpendicular to the radio polarization direction ( bower et al . 2003 ) .
these observations are consistent with our prediction that the apparent size and ellipsoidal shape of the radio images of sgr a * are due to the equatorial outflow from a meco in the center of sgr a*. scattering theory also predicts image sizes proportional to @xmath79 . the image sizes ( bower et al .
2004 , shen et al .
2005 ) have been interpreted only as scattering features superimposed on a compact object even though the major axis of the observed ellipsoids is twice the length of the minor axis .
the galactic center scattering screen is two to three orders of magnitude greater than what is seen in ngc6334b , the next most scattered source ( bower et al . 2004 ) .
a heavy screen would be expected for the base of the meco outflow where its highest radio frequencies would be generated .
the claimed detection of an intrinsic size of @xmath80 for sgr a * rests on apparent deviations from the scattering theory with its wavelength exponent of 2 , but for the cyclotron - synchrotron radiations considered here the exponent and deviations at distances inside the corotation radius could arise in a different way .
only much less luminous radio emissions should be generated in the inflow at wavelengths below about 0.3 mm ( @xmath15 ) .
33 ( appendix e ) gives an expected flux of @xmath81 for this wavelength and a distance of @xmath82 to sgr a*. as shown above , the surface emissions from a meco model for sgr a * would peak well below detection limits at @xmath83 . at wavelengths differing by a factor of even two from the thermal peak a meco would be as dark as a black hole ( see appendix a - c ) .
since we identify the @xmath15 to nir spectrum as originating in concentrated axial flows into the magnetic poles , we would expect image sizes at optically thin wavelengths to no longer be proportional to @xmath79 and to be elongated along the inflow axis rather than the equatorial plane . in the optically thin flow , axial distance from the meco should be correlated with frequency as @xmath84 . hence on the basis of the meco model for sgr
a * we are led to the prediction that , if viewed at high inclination to the polar axis , there would be two axial lobes , beginning at about @xmath85 and extending ever closer to a central dark source at shorter wavelengths . depending on the inclination there might be significant doppler boosting in the lobe with flow components
directed more toward us , however at these radial distances there would not be a uniform background of sub - mm to nir radiation against which a very dark shadow of the meco could be viewed .
there would be such a background only for wavelengths that originated inside @xmath86 and were extremely refracted gravitationally .
these wavelengths would necessarily be in the nir .
the meco - bondi model presented here can be compared with the radiatively inefficient accretion flow ( riaf ) into a black hole ( yuan , quataert & narayan 2003 ) .
to produce the observed polarization of the optically thick radio emission , the internally generated magnetic field of the riaf would need to be oriented generally e - w .
this might occur in two different ways .
( a ) if the riaf magnetic field responsible for the polarization were toroidal within the thick riaf disk , the riaf disk would need to be seen nearly edge - on in order to have a consistent magnetic field direction as seen from our side .
there would only be one direction of flow observed , either e - w or w - e , in the radio .
accordingly , a suitable radio interferometric array analogous to the proposed nir gravity array ( eisenhauer et al .
2008 ) might be able to distinguish a riaf from a meco outflow .
but in the nir , the riaf flow might show no preferential direction as observed with the gravity array because both near and far sides could be seen at once if sufficiently optically thin .
the nir , which would be produced closer to the central black hole is optically thin and would be produced in part by non - thermal electrons .
it is the power - law distribution of non - thermal electron energies in the optically thin part of the flow that would produce the orthogonal polarization of the nir ( agol 2000 ) .
( b ) if the riaf magnetic field were poloidal ; i.e. , with field lines emerging perpendicular to the disk , then the disk would need to be oriented generally n - s , but again seen nearly edge on . again
the flow would appear to be unidirectional .
the orthogonal polarization of the nir would then arise in the same way as before .
thus with either orientation , both polarizations would originate in the geometrically thick disk of the riaf and it should not look the same as the two zone flow of the meco - bondi model shown in figure 2 .
it is clear that the accretion rate in a riaf must be of order @xmath87 , and far below the bondi rate of @xmath88 in order to produce the observed strong linear polarization within the flow .
exactly how the riaf might accomplish this is unknown .
it has been suggested that the disk evaporates in a wind or that a jet outflow occurs . for a jet outflow to be consistent with observations ( yusef - zadeh et al .
2006 , muzic et al .
2007 , bower 2003 ) the riaf disk would need to be aligned generally n - s with a generally e - w jet . though this would give more of the appearance of figure 2 , a jet would be an actual collimated outflow rather than just a limb brightened region of a toroidal outflow and the flow in the riaf disk would still be unidirectional rather than into the central region from both above and below as in the flow into the poles of a meco .
the gravity array should be able to distinguish these possibilities .
ironically , if inflows were observed to disappear into an unseen dark object from both above and below , it could be mistaken for the flow into a black hole event horizon , but some explanation would then be required for an equatorial outflow which ought to also be observed , by means not presently known
. there would be too much luminosity generated by a kerr - metric ergospheric jet .
4 ) if the flow if hot enough to produce the quiescent soft x - rays over an extended region , why does it not produce more and harder luminosity in the riaf part of the flow 5 ) while it is clear that a source of non - thermal electrons must exist within the black hole riaf model there still remains an unsolved problem about the source of their energy . clearly these non - thermal electrons can not have been accelerated by magnetic reconnection processes because if that were the case then the black hole riaf would be too bright in the nir to fit the sgr a * observations . in spite of these problems ,
a black hole riaf can be used to account for most of the spectral distribution that has been observed , provided that it would be allowed to produce some radio emission in the disk as well as the jet .
this is necessary in order to account for the observed timing of nir / x - ray flares with both preceding and trailing radio flares . ] .
we have shown that the meco model accounts for the lack of observed surface luminosity from sgr a*. although it is necessary to know the global distributions of electron density and magnetic fields before spatial and spectral energy distributions can be accurately calculated , we have shown that a spectrum of approximately the correct luminosity and spectral indexes would be produced in the inflow - outflow zones of a bondi accretion flow into the magnetic field of a meco .
the equatorial outflow would produce optically thick cyclotron radiation with positive spectral index at frequencies below @xmath89 .
the axial inflow would produce steeply declining ( negative index ) optically thin nir emissions as well as some correlated x - ray ssc and brehmsstrahlung emissions that could be observed in flares caused by high density clumps in the inflow .
we have shown that timing of flares in radio / nir / x - ray bands is consistent with the meco - bondi model , with some weak sub - mm flaring preceding the strongest nir and ssc x - ray flares which are then followed by stronger delayed sub - mm and radio flare emissions .
the part of the bondi flow that eventually departs in the equatorial outflow would produce radio emissions , possibly to frequencies as high as @xmath89 , and including nearly everything below @xmath90 .
the bulk of the quiescent x - ray luminosity would be thermally generated within the magnetosphere in a mixed inflow / outflow pattern of @xmath91 size .
we have shown that the low bolometric luminosity of sgr a * can be reconciled with an expected bondi accretion rate in a completely natural way .
the magnetic propeller mechanism is a robust , stable physical mechanism for reducing the bondi accretion rate to levels compatible with the low luminosity of sgr a*. the only parameters that have been necessary for these calculations are the ion density and sound speed at the bondi radius , mass , magnetic moment and spin .
the first three of these have been taken from work reported by others .
the intrinsic magnetic moment of a meco is an inherent , mass dependent feature which is generated by the effects of the quantum electrodynamic stablility conditions that are required by the einstein - maxwell equations that describe the highly red shifted eddington limited meco collapse process ( rl06 ) .
its magnitude sets a natural high frequency limit for the synchrotron emissions in the axial inflow .
our estimate of spin has been taken from our previous work that has accounted for the radio / x - ray luminosity correlations and spectral state switches for agn and gbhc ( rl04 ) and the microlensed image of the quasar q0957 + 561 ( schild , leiter & robertson 2006 ) .
in retrospect , it is fortuitous that this choice has worked out well , but it suggests that the accretion history of sgr a * may be similar to that of other agn .
since they do not possess event horizons , highly red shifted general relativistic meco contain intrinsic magnetic moments that can interact with their environments .
the intrinsic meco magnetic moment automatically produces the ordered magnetic fields necessary to account for the observed strong linear polarization . in the meco - bondi inflow - outflow model for sgr
a * described here , there are regions of differing magnetic field strength and orientation , strong density variations and both inflows and outflows .
we have shown that there are places of origin in this flow for all of the spectral , spatial , polarization and timing features that have so far been observed for sgr a*. the patterns of inflow and outflow differ from those expected of black hole models and should be observable by the proposed gravity array ( eisenhauer et al .
2008 ) . on the basis of the meco - bondi model for sgr
a * we predict that high resolution images in radio frequencies should be elongated in the limb brightened outflow zone , while high resolution images in shorter wavelengths should be elongated along an orthogonal polar axis .
since the emissions in these shorter wavelengths are confined to the narrow axial inflow region , there would be no uniform background to provide a silhouette image of a dark meco except for strong gravitational refraction effects on nir frequencies generated inside @xmath92 .
everything inside @xmath93 would just be dark in the radio frequencies .
the qualitative consistency of the meco - bondi model provides added incentive for doing additional simulations which , in order to succeed , will require substantial computational facilities and expertise .
calculations of the global magnetic field and density distributions will be necessary first steps before synchrotron emissions in the outflow can be calculated . in order to add spectral details , smaller calculation grids close to the central dipole and consideration of light paths in strong field gravity will be required and could be challenging , even for the original us - russia supercomputer collaboration that produced the work of rttl03 .
calculations extending into the region near the photon sphere ought to be able to provide spectral and image details that could be compared with the increasingly high resolution images of sgr a*. further corroboration of the meco - bondi model for sgr a * could be found if further observations eventually reveal nir lobes for accretion flow into the magnetic polar regions , though they would likely be smeared into ellipsoids by gravitational refraction
. a pattern of expected polarizations of radiation should also be computed .
the meco parameters for spin and magnetic moment used here provide a place to start , but for best comparisons these parameters , as well as the bondi flow parameters should be varied .
rather detailed calculations of what might be seen for different axis orientations will also be necessary .
it is our hope that the present paper will provide motivation for the additional work to be done .
abramowicz , m. , kluzniak , w. & lasota , j - p , 2002 a&a , 396 , l31 agol , e. , 2000 apj , 538 , l121 an , t. , et al .
, 2003 apj , 634,49 baganoff , f. , et al . , 2003 apj , 591 , 891 bower , g. , 2000 gcnews 11 , 4 bower , g. , et al . , 2004 science , 304 , 704 broderick , a. , & narayan , r. , 2006 apj , 638 , l21 bn06 campana , s. et al . ,
2002 apj 580 , 389 campana , s. et al .
, 1998 a&a rev .
8 , 279 cavaliere , a. & morrison , p. 1980 apj 238 , l63 corbel , s. , fender , r.p . ,
tzioumis , a.k . , nowak , m.a . ,
mcintyre , v. , durouchoux , p.,sood , r. , 2000 a&a 359 , 251 corbel , s. , nowak , m.a . , fender , r.p . , tzioumis , a.k .
markoff , s. , 2003 a&a 400 , 1007 eckart , a. et al .
2008 a&a , july 8 in press /astro - ph0712.3165 eisenhauer , f. , et al . , 2008 proc .
spie astron .
, 22 - 28 june , marseille , france , in press @xmath94 astro - ph/0808.0063 gallo , e. , fender , r. , pooley , g. , 2003 mnras , 344 , 60 falcke , h. , krding , e. , markoff , s. , 2004 a&a , 414 , 895 falcke , h. , & markoff , s. , 2000 a&a 362 , 113 gliozzi , m. , bodo , g. & ghisellini , g. 1999 mnras , 303 , l37 heinz , s. , sunyaev , r. , 2003 mnras 343 , l59 harding , a. , 2003 , invited talk at pulsars , axps and sgrs observed with bepposax and other observatories , marsala , sicily , sept .
2002 astro - ph/0304120 .
ilarianov , a. & sunyaev , r. 1975 a&a , 39 , 185 kippenhahn , r. & weigert , a. 1990 ` stellar structure and evolution ' springer - verlag , berlin heidelberg new york landau , l. & lifshitz , e. 1958 ` statistical physics ' , pergamon press ltd , london leiter , d. , & robertson , s. , 2003 found .
lett . , 16 , 143 maccarone , t. , gallo , e. , fender , r. 2003 mnras 345 , l19 markoff , s. , falcke h. , & fender , r. 2001 a&al , 372 , 25 markoff , s. , nowak , m. , corbel , s. , fender , r. , falcke , h. , 2003 new astron .
47 , 491 mcclintock , j. , narayan , r. , and rybicki , g. , 2004 , apj , 615 , 402 mitra , a. , 2006 new astron .
, 12 , 146 - 160 mitra , a. , 2006 mnras 369 , 492 mitra , a. , 2006 phys .
d , 74 , 034010 muzik , k. , et al .
, 2007 , proc .
intl astron .
union , 2 , 415 robertson , s. & leiter , d. 2002 apj , 565 , 447 rl02 robertson , s. , leiter , d. 2003 apj , 596 , l203
rl03 robertson , s. , leiter , d. 2004 mnras , 350 , 1391 rl04 robertson , s. , and leiter , d. 2006 ` the magnetospheric eternally collapsing object ( meco ) model of galactic black hole candidates and active galactic nuclei ' , pp 1 - 45 ( in new developments in black hole research , ed .
p.v.kreitler , nova science publishers , inc .
isbn 1 - 59454 - 460 - 3 , novapublishers.com ) rl06 astro - phys/0602543 romanova , m. , toropina , o. , toropin , y. , & lovelace , r. , 2003 apj , 588 , 400 romanova , m. , ustyugova , a.,koldoba , a. , & lovelace , r. , 2005 apj , 635 , l165 schild , r. , leiter , d. & robertson , s. , 2006 aj , 132 , 420 schild , r. , leiter , d. & robertson , s. , 2008 aj , 135 , 947 shen , z. , et al . , 2005 nature , 438 , 62 stirling a. m. , et al . , 2001 mnras 327 , 1273 toropina , o. , romanova , m. , toropin , y. , & lovelace , r. , 2003 apj 593 , 472 vadawale , s. , rao , a. & chakrabarti , s. 2001 a&a , 372 , 793v yuan , f. , markoff , s. & falcke , h. , 2002 a&a , 383 , 854 yuan , f. , quataert , e. , narayan , r. , 2003 apj 598 , 301 yusef - zadeh et al . , 2006 ,
in vi microquasar workshop : microquasars and beyond , sept .
18 - 22 , 2007 , como , italy , astro - ph/0612156 yusef - zadeh et al . ,
2006 apj 650 , 189 yusef - zadeh et al . , 2008 apj , 682 , 361 zaumen , w. t. , 1976 apj , 210 , 776
* a. eco models * + an eternally collapsing object " eco is a gravitationally compact mass supported against gravity by internal radiation pressure ( mitra 2006 ) . in its outer layers of mass , a plasma with some baryonic content is supported by a _
net _ outward flux of momentum via radiation at the local eddington limit @xmath95 given by @xmath96 here @xmath97 is the opacity of the plasma , subscript s refers to the baryonic surface layer and @xmath98 is the gravitational redshift at the surface . in general relativity
@xmath6 is given by @xmath99 for a hydrogen plasma , @xmath100 @xmath101 /g and @xmath102 where @xmath103 is the mass in solar units .
since the temperature at the baryon surface is beyond that of the pair production threshold , there is a pair atmosphere further out that remains opaque .
the net outward momentum flux continues onward , but diminished by two effects , time dilation of the rate of photon flow and gravitational redshift of the photons .
the escaping luminosity at a location where the redshift is @xmath6 is thus reduced by the ratio @xmath104 , and the net outflow of luminosity as radiation transits the pair atmosphere and beyond is @xmath105 finally , as distantly observed where @xmath106 , the luminosity is @xmath107 for hydrogen plasma opacity of 0.4 @xmath101 g@xmath108 , and a typical gbhc mass of @xmath109 this equation yields @xmath110 erg s@xmath108 .
but since the quiescent luminosity of a gbhc must be less than about @xmath111 erg s@xmath108 , we see that it is necessary to have @xmath112 .
even larger redshifts are needed to satisfy the quiescent luminosity constraints for agn .
this is extraordinary , to say the least , but perhaps no more incredible than the @xmath113 of a black hole . at the low luminosity of eq .
10 , the gravitational collapse is characterized by an extremely long radiative lifetime , @xmath114 ( robertson & leiter 2003 , mitra 2006 ) given by : @xmath115 with the large redshifts that would be necessary for consistency with quiescent luminosity levels of bhc , it is clear why such a slowly collapsing object would be called an eternally collapsing object " or eco . for @xmath116 ,
radiation is impeded by passage through a small escape cone such that the fraction of radiation that could escape if isotropically emitted at radius r would only be @xmath117 for very large @xmath6 , @xmath118 and @xmath119 . at the outskirts of the pair atmosphere of an eco
the photosphere is reached . here
the temperature and density of pairs has dropped to a level from which photons can depart without further scattering from positrons or electrons .
nevertheless , the redshift is still large enough that their escape cone is small and most photons will not travel far before falling back through the photosphere . if we let the photosphere temperature and redshift be @xmath120 and @xmath121 , respectively , the net escaping luminosity is @xmath122 but in the radiation dominated region beyond the photosphere , the temperature and redshift are related by @xmath123 where @xmath73 is the distantly observed radiation temperature . substituting into the previous equation ,
we obtain @xmath124 and , for hydrogen plasma opacity @xmath125^{1/4}}~~~~k\ ] ] the left equality of eq .
15 can be written in terms of the distantly observed spectral distribution , for which the radiant flux density at distance r ( @xmath126 ) and frequency @xmath127 would be @xmath128 as previously discussed ( rl06 ) , if one naively assumes that the photon support for an eco originates from purely thermal processes , one quickly finds that the temperature in the baryon surface layer would be orders of magnitude higher than the pair production threshold . the compactness guarantees ( cavaliere & morrison 1980 ) that photon - photon collisions would produce numerous electron - positron pairs .
drift currents proportional to @xmath129 reactively generate extreme magnetic fields .
we assumed that the baryonic surface field reaches an equipartion level with a surface magnetic field of @xmath130 g ( harding 2003 , zaumen , 1976 ) that is capable of creating bound electron - positron pairs on the baryon surface .
the interior magnetic field in a stellar mass meco - gbhc is about what would be expected from flux compression during stellar collapse . at the meco surface radius ( @xmath118 ) , the ratio of tangential field on the exterior surface to the tangential field just under the meco surface is given by ( rl06 ) @xmath131 @xmath132 we have previously taken @xmath133 gauss as typical of the interior field that can be produced by flux compression during stellar gravitational collapse . using this value
the preceding equation has the solution . @xmath134
the magnetic moment of a meco would be @xmath135 we have found that this magnetic moment , and the spin rates given by table 1 , eq . 6
give the meco model a good correspondence with observations of spectral state switches and the radio luminosities of jets for both gbhc and agn ( rl02 , rl03 , rl04 , rl06 ) .
the photosphere , as a last scattering surface , can be found from the condition that ( kippenhahn & wiggert 1990 ) @xmath136 where @xmath137 is an increment of proper length in the pair atmosphere and @xmath138 is the combined number density of electrons and positrons along the path .
landau & lifshitz ( 1958 ) show that @xmath139 where p is the momentum of a particle , @xmath140 , k is boltzmann s constant , @xmath141 is planck s constant and @xmath142 , the mass of an electron . for low temperatures such that
@xmath143 this becomes : @xmath144 @xmath145 where @xmath146k .
we can express @xmath137 in terms of changing redshift as @xmath147 for @xmath118 substituting into eq .
23 and using @xmath148 beyond the photosphere we obtain the relation @xmath149 using eqs . 16 and 19 , we have numerically integrated this equation to obtain the photosphere temperatures and redshifts for various masses .
the results are represented with errors below 1% for @xmath150 solar mass by the relations : @xmath151 and @xmath152 these relations , though little different from those of rl06 , correct an error in our previous development of the photosphere temperature . * d. accretion efficiency for meco surface luminosity * + we need to reconsider how accreting material interacts with a meco .
bn06 assumed that for large redshifts , accretion energy would be converted to luminosity immediately with 100% efficiency . in previous work ( rl03 , rl06 )
we mistakenly made the same assumption .
it is true that a meco will eventually achieve 100% efficiency , but the conversion takes place on the time scale of the meco radiative lifetime .
accreting particles that reach the photosphere do not produce a hard radiative impact .
they first encounter soft photons , then photons of @xmath153 mev and electron - positron pairs near the photosphere and they eventually penetrate the baryon surface where the net outflowing luminosity is already at the local eddington limit rate . this provides a very soft landing .
perhaps the easiest way to show the difference between landing on a cushion of photons and striking a hard surface is to consider the pressure that accreting matter could exert if stopped dead at the photosphere and compare that to the radiation pressure already there .
radiation pressure at the photosphere ( hence the luminosity also ) would need to change by only about the same amount as the accretion pressure for the meco to remain stable .
consider a particle of rest mass @xmath154 in radial free fall from infinity .
its speed as it reaches the photosphere would be essentially light speed , @xmath155 , according to an observer at the photosphere , and it would have a momentum of @xmath156 , where @xmath157 is the lorentz factor , and proper time , @xmath114 , moving with the particle are related by @xmath158 .
an interval @xmath159 for an observer at rest at the photosphere is related by @xmath160 , from which it follows that @xmath161 . ] .
if particles arrive at the locally observed rate of @xmath162 , then the quantity of momentum deposited according to the observer at the photosphere would be @xmath163 . with the substitution of @xmath164 for @xmath159 ,
this becomes @xmath165 .
thus the rate of momentum transport , as observed at the photosphere is @xmath166 , where @xmath167 .
if momentum were deposited uniformly at the photosphere , the accretion pressure would be @xmath168 erg @xmath169 , using eq .
27 for @xmath121 and @xmath170 . for the mass of sgr
a * and @xmath171 g / s , this would yield @xmath172 erg @xmath169 .
the radiation pressure at the photosphere would be @xmath173 erg @xmath169 , using eq .
this pressure ratio would be @xmath174 , however , the actual ratio of accretion and radiation pressures would be smaller yet because the entire momentum is not deposited in a hard surface impact at the photosphere . only about @xmath175 of the incoming momentum
is transferred to the entire pair atmosphere .
this is just too little to affect the radiation leaving the photosphere .
finally , we note that in an equipartition magnetic field , the temperature of the pair plasma layer is buffered near the pair threshold temperature .
adding energy just produces more pairs rather than raising the temperature .
of course , for stability , the meco must adjust its radiation rate to accomodate a growing mass , but as distantly observed , the luminosity produced by the incremental increase of mass is negligible . for mass accretion rate @xmath176
the radiated luminosity increases at a rate of about @xmath177 .
we conclude that the efficiency and thermal equilibrium constraints of bn06 simply do not apply to the meco surface .
eventually the radial infall of accreting particles ends , on average somewhere below the meco baryon surface . with the temperature at the base of the pair atmosphere at @xmath178 k , eq .
23 shows the local pair density to be near @xmath179 .
the proper length thickness of the pair atmosphere is @xmath180 and the optical depth is @xmath181 . assuming that protons in an accreting plasma interact with the electron - positron pairs via coulomb scattering , the fraction of the accretion energy that they can transfer directly to the pair atmosphere is only @xmath182 .
photons in the pair atmosphere are somewhat more effective in absorbing accretion energy . for conditions in the pair atmosphere , about @xmath175 of the accretion energy can be removed by photon collisions with incoming electrons , however , the escape cone allows only about one in @xmath183 of the compton enhanced photons to escape .
most of the brehmsstrahlung produced is therefore buried below the meco baryon surface which is itself covered by an extremely optically thick layer of pairs for which the escape cone is negligibly small . with due consideration of the small escape cone , it can be easily shown that if accreting protons are stopped in a layer below the photosphere for which the redshift is @xmath184 , then the ratio of accretion pressure to radiation pressure already in the layer is @xmath185 , for @xmath186 .
accretion pressure is completely insignificant even for accretion rates exceeding the classical edddington limit . in effect
, a meco just swallows accreting particles about as effectively as a black hole . * d.1 high / soft spectral states * + reconsideration of the meco surface accretion efficiency necessitates a new interpretation of the high / soft spectral states of disk accreting systems . at accretion rates above the transition to the soft state ,
most of the luminosity of the high / soft state must arise from the accretion disk rather than the meco .
this can occur with high efficiency corresponding to an accretion disk that can penetrate well inside what would otherwise correspond to the marginally stable orbit of a non - magnetic black hole .
the inner disk of a meco is supported in part by magnetic pressure , as previously described in rl06 and disk accretion efficiencies can reach 42% at the photon sphere at @xmath55 . inside the photon sphere ,
gravitational redshift offsets the effect on luminosity of additional radiant energy release by accreting matter .
the energy equation for the bondi flow is @xmath187 where @xmath188 is the bondi flow speed , @xmath189 the ratio of specific heats for the gas , @xmath190 the mass of the central object and @xmath52 the radial coordinate distance from its center . the continuity equation for a rate of mass flow @xmath191 is @xmath192 we adopt the values @xmath193 , @xmath194 , @xmath195 ( baganoff 2003 ) and a specific heat ratio of @xmath196 for atomic hydrogen .
having adopted these values , they set the bondi radius as @xmath197 and they must be entered consistently in the energy equation .
we use @xmath198 , where @xmath199 for an adiabatic flow .
using @xmath200 and substituting into the energy equation produces an equation of the form @xmath201 where a , b and c are constants .
solutions for v for various values of r are given in table 2 .
once v has been determined , @xmath202 can be obtained from the continuity equation , and the pressure and proton temperature can be found from the adiabatic conditions . for @xmath196 ,
the inflow never becomes supersonic , but it very closely approaches sound speed ( which is half what the free - fall speed would be ) by the time the magnetosphere is reached . for the dilute plasma at the bondi radius ,
there is a relatively low frequency of electron - proton collisions .
the kinetic energy gained by falling electrons is not sufficient to keep them in thermal equilibrium with the ions .
approximately half of the gravitational energy released goes into heating the ions while the electrons remain relatively cold .
the electrons do gain some energy , which we can calculate , in collisions with protons .
as the plasma falls by @xmath203 , the time taken is @xmath204 .
an electron travels @xmath205 in this time .
the number of collisions with ions in this time is the distance travelled divided by the mean free path , @xmath206 , where @xmath207 is the coulomb cross - section .
@xmath208 is the electron energy , @xmath51 the electron temperature and @xmath209 is the coulomb logarithm , which we assume to have the value 30 . for collisions with protons which , on average , have much more momentum than the electrons , the fraction of proton energy , @xmath210 , that can be transferred is approximately @xmath211 .
then while falling dr , an electron gains energy @xmath212 .
taking @xmath213 for both the proton thermal speed and speed of descent and substituting for @xmath214 and @xmath215 , this integrates to yield @xmath216 and specifically , @xmath217 yields the values in table 2 .
an electron moving with the protons at speed @xmath218 in a magnetic field would emit cyclotron radiation at a rate @xmath219 . comparing this rate for the axial dipole field of a meco with the mass of sgr a * with @xmath220 above
, we find that the radiation rate would exceed the rate that electron energy is extracted from the protons for @xmath221 , which holds for @xmath222 , and cyclotron frequencies above @xmath223 . nearer the meco at smaller axial distances and higher frequencies ,
the electron energy extracted from the protons in the flow is essentially all that is available for radiation .
this permits a rough calculation of the luminosity generated in the polar inflow .
we approximate the polar inflow geometry within the magnetosphere as a cone of lateral radius @xmath224 at distance z from the magnetic pole . see figure 1 .
the number of electrons in the cone in axial thickness @xmath28 is @xmath225 .
thus the rate that they can radiate is about @xmath226 .
substituting for @xmath227 , @xmath228 , @xmath229 as before and using @xmath230 , @xmath231 from equations in table 1 , this reduces to @xmath232 , which integrates to give the luminosity for polar inflow to axial distance @xmath6 as @xmath233 this last result includes an additional factor of two to account for inflow into both magnetic poles . where the flow is optically thin above @xmath15 , the larmor frequency should strongly dominate the spectrum .
this frequency is a function of the magnetic field strength , which varies as @xmath234 .
alternatively , we can say that position @xmath6 in the polar axial inflow is correlated with the dominant radiation frequency as @xmath31 . substituting @xmath235 for @xmath6 in the expression just preceding eq .
32 , one obtains @xmath236 hence the spectral index in the optically thin sub - mm / nir would be @xmath237 for the meco model . | in previous work we have presented evidence for the existence of intrinsic magnetic moments in black hole candidates .
we later developed a general relativistic , magnetospheric eternally collapsing object ( meco ) model for black hole candidates and showed that the model is consistent with broad band spectral and luminosity characteristics and accounts for the radio / x - ray luminosity correlations of both galactic black hole candidates ( gbhc ) and active galactic nuclei ( agn ) .
since magnetic moments are forbidden attributes for black holes , the meco model has the potential to test whether black hole candidates are actual black holes .
we show here that the meco model has the advantage of being able to : a ) satisfy the luminosity constraints that have been claimed as proof of an event horizon for sgr a * , b ) reconcile the low bolometric luminosity with the expected bondi accretion rate for sgr a * by means of a magnetic propeller driven outflow , c ) account for the sgr a * nir and x - ray luminosities , the general characteristics of its broad band spectrum , and the sequence of flares in different spectral ranges as well as the pattern of its observed orthogonal polarizations .
we also include specific predictions for images that may be obtained in sub - millimeter to nir wavelengths in the near future .
high resolution images in radio frequencies should be elongated in an equatorial outflow plane , while high resolution images in shorter infrared wavelengths should be elongated along an orthogonal , magnetic polar inflow axis ( generally n - s ) .
since the emissions in these shorter wavelengths are confined to a narrow axial inflow cone , and radio frequencies are generated primarily at greater distances in the equatorial outflow , there would be no uniform background to provide a silhouette image of a dark central object .
additional future tests for the presence of an intrinsic magnetic moment for sgr a * will require global solutions for electron density and magnetic field distributions in a bondi accretion flow into a compact , rotating magnetic dipole .
these will provide for definitive tests in the form of detailed calculations of spectral and spatial luminosity distributions and polarization maps for direct comparison with high resolution images of sgr a*. _ key words : _ accretion , accretion disks black hole physics
galaxy : center gravitation infrared : general magnetic fields |
let @xmath2 be an @xmath3 integer symmetric matrix with characteristic polynomial @xmath4 .
the aim of this paper is to describe all such matrices @xmath2 that have the maximum modulus of their eigenvalues at most @xmath5 .
the significance of the bound @xmath5 is that , by a result of kronecker @xcite , every eigenvalue of such a matrix @xmath2 is then of the form @xmath6 , for some root of unity @xmath7 .
thus @xmath8 is a cyclotomic polynomial .
for this reason we call such integer symmetric matrices _ cyclotomic _ matrices . in 1970
smith @xcite classified all cyclotomic @xmath9-matrices with zeros on the diagonal , regarding them as adjacency matrices of graphs ( see figure [ f - except ] ) .
such graphs were called _ cyclotomic _
graphs in @xcite .
it turns out that a full description of cyclotomic matrices is conveniently stated using more general graphs .
so if we allow the off - diagonal elements of our matrix to be chosen from the set @xmath10 , we obtain a _ signed _ graph ( see @xcite,@xcite ) , a non - zero @xmath11th entry denoting a ` sign ' of @xmath12 or @xmath13 on the edge between vertices @xmath14 and @xmath15 .
further , for a general symmetric @xmath10 matrix , where now the diagonal entries may be nonzero , we obtain what we call a _ charged _ signed graph ; we regard a nonzero @xmath16th entry of @xmath2 as corresponding to a ` charge ' on its @xmath14th vertex .
if none of the edges of a charged signed graph in fact have sign @xmath12 , then we have a charged ( unsigned ) graph .
however , a graph is also a signed graph , and a signed graph is also a charged signed graph .
the notion of a charged signed graph is a convenient device for picturing and discussing symmetric integer matrices with entries in @xmath10 .
these are the most important matrices in our description of general cyclotomic matrices . in this paper
we extend smith s result to cyclotomic charged signed graphs ( theorem [ t - lcsg ] ) , and then , with little further work , to all cyclotomic matrices ( theorem [ t - ism ] ) . along the way we find all cyclotomic signed graphs ( theorem [ t - csg ] ) . as a consequence
, we can also describe all cyclotomic charged graphs ( theorem [ t : unsigned_charged ] ) and all cyclotomic matrices whose entries are non - negative ( theorem [ t : nonnegative_matrix ] ) . having obtained our results for the closed interval @xmath0 $ ] , it is then very natural to consider restricting the eigenvalues to the open interval @xmath1 . we give a complete classification of symmetric integer matrices with eigenvalues in this restricted set ( theorem [ t - ismopen ] ) . as in the case of the closed interval , there are corresponding results for cyclotomic signed graphs ( theorem [ t : open_signed ] ) , cyclotomic charged signed graphs ( theorem [ t : open_charged_signed ] ) , cyclotomic charged graphs ( theorem [ t : open_charged ] ) and cyclotomic matrices whose entries are non - negative ( theorem [ t : nonnegative_matrix_open ] ) . having dealt with the general cyclotomic case , this is a relatively straightforward problem .
there is a connection here with the theory of finite reflection groups and their coxeter graphs , and we conclude with a discussion of this . in @xcite , cyclotomic graphs were used to construct salem numbers and pisot numbers .
the original motivation for this current work was that it provides one of the ingredients necessary to extend the work in @xcite .
but we think that our results may be of independent interest . throughout the paper , a _
subgraph _ of the ( charged , signed ) graph under consideration will always mean a vertex - deleted subgraph , that is , an induced subgraph on a subset of the vertices .
in order to state our results , we need some preliminaries . the matrix @xmath2 will be called _ indecomposable _ if and only if the underlying graph is _
connected_. ( in the underlying graph , vertices @xmath14 and @xmath15 are adjacent if and only if the @xmath11th entry of @xmath2 is nonzero . )
if @xmath2 is not indecomposable , then there is a reordering of the rows ( and columns ) such that the matrix has block diagonal form with more than one block , and its list of eigenvalues is found by pooling the lists of the eigenvalues of the blocks . for our classification of cyclotomic matrices , it is clearly sufficient to consider indecomposable ones .
a repeatedly useful tool for us is cauchy s interlacing theorem ( for a short proof , see @xcite ) .
[ l : interlacing ] let @xmath2 be a real symmetric matrix , with eigenvalues @xmath17 . pick any row @xmath14 , and let @xmath18 be the matrix formed by deleting row @xmath14 and column @xmath14 from @xmath2 .
then the eigenvalues of @xmath18 interlace with those of @xmath2 : if @xmath18 has eigenvalues @xmath19 , then @xmath20 in view of this lemma , if @xmath2 is cyclotomic , then so is any matrix obtained by deleting from @xmath2 any number of its rows , along with the corresponding columns : we then speak of the smaller matrix as being _ contained in _ the larger one ( the smaller graph is an induced subgraph of the larger graph ) .
we call an indecomposable cyclotomic matrix ( or its graph ) _ maximal _ if it is not contained in a strictly - larger indecomposable cyclotomic matrix : the corresponding cyclotomic graph is not an induced subgraph of a strictly larger connected cyclotomic graph .
we shall see that every non - maximal indecomposable cyclotomic matrix is contained in a maximal one .
it is therefore enough for us to classify all maximal indecomposable cyclotomic matrices .
when we consider matrices that have all their eigenvalues in the open interval @xmath1 , we shall see that it is no longer always true that every such matrix is contained in a maximal one : there is an infinite family of indecomposable exceptions .
denote by @xmath21 the orthogonal group of @xmath3 signed permutation matrices .
then conjugation of a cyclotomic matrix by a matrix in @xmath21 gives a cyclotomic matrix with the same eigenvalues .
we say that two @xmath3 cyclotomic matrices are _ strongly equivalent _ if they are related in this way .
further , we say that two indecomposable cyclotomic matrices @xmath2 and @xmath22 are merely _ equivalent _ if @xmath22 is strongly equivalent to @xmath2 or @xmath23 .
this notion then extends easily to decomposable cyclotomic matrices . both of these notions are equivalence relations on the set of all cyclotomic matrices . for indecomposable cyclotomic matrices ,
the equivalence classes for the weaker notion are the union of one or two strong equivalence classes , depending on whether or not @xmath23 is in the same strong equivalence class as @xmath2 .
it is clearly sufficient to classify all cyclotomic matrices up to equivalence . for a charged signed graph ,
the notions of strong equivalence and equivalence of course carry over via the adjacency matrix .
now @xmath21 is generated by diagonal matrices of the form @xmath24 and by permutation matrices .
conjugation by these diagonal matrices corresponds to reversing the signs of all edges incident at a certain vertex @xmath25 ; we call this _ switching at @xmath25_. conjugation by a permutation matrix merely means that we can ignore vertex labels ; we therefore do not label the vertices of our graphs .
thus for unlabelled charged signed graphs , strong equivalence classes are generated only by such switching operations .
the concept of switching , and signed switching classes , appeared earlier for signed graphs in @xcite .
equivalence of charged signed graphs is generated both by switching , and by the operation of reversing all the edge signs and vertex charges of a component of a graph .
since most of our graphs will in fact be signed graphs , we avoid clutter by drawing edges with sign @xmath13 as unbroken lines , and edges with sign @xmath12 as dashed lines - - - - - - . for vertices ,
those of charge @xmath26 will be drawn , @xmath27 , respectively , with the vertices @xmath27 without a charge being called _ neutral _ vertices .
[ t - csg ] every maximal connected cyclotomic signed graph is equivalent to one of the following : 1 . for some @xmath28 , the @xmath29-vertex toral tesselation @xmath30 shown in figure [ f - csg_general ] ; 2 .
the @xmath31-vertex signed graph @xmath32 shown in figure [ f - csg_14 ] ; 3 .
the @xmath33-vertex signed hypercube @xmath34 shown in figure [ f - csg_16 ] .
further , every connected cyclotomic signed graph is contained in a maximal one .
in particular , @xmath35 of case ( i ) gives an octahedron @xmath36 , shown in figure [ f - oct ] , while a more typical example @xmath37 is shown in figure [ f - torus ] .
[ t - lcsg ] every maximal connected cyclotomic charged signed graph not included in theorem [ t - csg ] is equivalent to one of the following : 1 .
for some @xmath38 , one of the two @xmath29-vertex cylindrical tesselations @xmath39 shown in figure [ f - lcsg_general ] ; 2 .
one of the three sporadic charged signed graphs @xmath40 shown in figure [ f - lcsg_sporadic ] ; further , every connected cyclotomic charged signed graph is contained in a maximal one .
in particular , @xmath41 of case ( i ) gives two charged tetrahedra @xmath42 , shown in figure [ f - lcsg_tet ] .
we remark that all the maximal cyclotomic graphs of theorems [ t - csg ] and [ t - lcsg ] are ` visibly ' cyclotomic : their adjacency matrices @xmath2 all satisfy @xmath43 , so all their eigenvalues are @xmath44 .
the exact multiplicity of these eigenvalues is given in table [ ta-1 ] at the end of the paper .
our most general result is readily deduced from the previous two theorems .
[ t - ism ] every maximal indecomposable cyclotomic matrix is equivalent to one of the following : 1 .
the adjacency matrix of a maximal connected charged cyclotomic signed graph ( given by theorems [ t - csg ] and [ t - lcsg ] ) ; 2 . the @xmath45 matrix @xmath46 or the matrix @xmath47 .
+ further , every indecomposable cyclotomic matrix is contained in a maximal one .
a signed graph @xmath48 is called _ bipartite _ if its vertices can be split into two disjoint parts such that every edge of @xmath48 joins a vertex in one part to a vertex in the other ( @xcite ) .
the eigenvalues of @xmath48 are then symmetric about 0 , counted with multiplicity ; we record this fact as a lemma .
[ l : bipartite ] let @xmath48 be a bipartite signed graph with @xmath49 vertices .
then @xmath50 one can mimic a standard proof for graphs ( as in @xcite ; this result first appeared in a chemistry paper @xcite ) , or simply note that if one changes the signs of all edges incident with vertices in one part then @xmath51 is unchanged , yet every edge has then changed sign so that @xmath52 is changed to @xmath53
. it will be convenient to extend the definition of bipartite to cover any charged signed graph such that changing the sign of every edge and charge produces a graph that is strongly equivalent to the original . for ( neutral )
signed graphs , this captures the usual definition of being bipartite .
the extension of lemma [ l : bipartite ] holds true for this larger class of bipartite charged signed graphs , with the same proof . a _ cycle _ of length @xmath54 in a charged signed graph @xmath48 is a list of distinct vertices @xmath55 such that there is an edge in @xmath48 between @xmath56 and @xmath57 ( @xmath58 ) and between @xmath59 and @xmath60 . a charged signed graph without cycles
is called a ( charged signed ) _
forest_. a connected forest is called a
_ tree_. any charged signed forest is equivalent to one for which all the edges are positive . an easy induction on the number of vertices : for the inductive step consider removing a leaf ( a vertex with exactly one neighbour ) , unless there are no edges . for detecting noncyclotomic integer symmetric matrices , the following trivial and obvious sufficient condition can be useful .
[ l : non - cyc ] let @xmath2 be an @xmath3 integer symmetric matrix . if either @xmath61 or @xmath62 , then @xmath2 is not cyclotomic .
[ l : smallcases ] up to equivalence , the only indecomposable @xmath13-by-@xmath13 or @xmath5-by-@xmath5 cyclotomic matrices are @xmath63 of these , the only maximal ones are @xmath46 and @xmath64 .
this is an easy computation , using lemma [ l : non - cyc ] to constrain the matrix entries .
for example , to show that @xmath64 is maximal , suppose that @xmath65 is cyclotomic .
to achieve @xmath66 and @xmath67 requires both @xmath68 and @xmath69 , giving @xmath70 , so that @xmath2 is not indecomposable .
[ l : reducetochargedsignedgraphs ] apart from matrices equivalent to either @xmath46 or @xmath64 , any indecomposable cyclotomic matrix has all entries from the set @xmath71 .
in other words , it is the adjacency matrix of a cyclotomic charged signed graph .
let @xmath72 be an indecomposable cyclotomic matrix , not equivalent to either @xmath46 or @xmath64 .
suppose first that some diagonal entry of @xmath2 had modulus at least 2 , say @xmath73 . by interlacing ( lemma [ l : interlacing ] ) ,
the 1-by-1 matrix @xmath74 is cyclotomic , and
then by lemma [ l : smallcases ] it equals @xmath75 and is maximal , so equals @xmath2 , giving a contradiction .
next suppose that some off - diagonal entry @xmath76 had modulus at least @xmath5 . by interlacing ,
the 2-by-2 matrix @xmath77 is cyclotomic , and by lemma [ l : smallcases ] this must equal @xmath78 , and is maximal , so equals @xmath2 .
again we have a contradiction .
thus no entry of @xmath2 has modulus greater than 1 .
we conclude that , apart from two ( up to equivalence ) trivial examples , all indecomposable cyclotomic matrices are the adjacency matrices of connected cyclotomic charged signed graphs .
thus theorem [ t - ism ] follows from theorems [ t - csg ] and [ t - lcsg ] , and we can restrict our attention to charged signed graphs .
let @xmath2 be the adjacency matrix of a cyclotomic charged signed graph with @xmath49 vertices . in particular , @xmath2 has all eigenvalues at least @xmath79 .
hence @xmath80 is positive semi - definite .
this implies that we can find vectors @xmath81 , ,
@xmath82 in real @xmath49-dimensional space such that @xmath80 is their gram matrix : the @xmath11-entry of @xmath80 is the dot product of @xmath83 and @xmath84 . the dimension of the space spanned by the @xmath83 might of course be smaller than @xmath49 .
a particularly simple case is that of a signed graph , where there are no charges .
then the diagonal entries of @xmath80 all equal 2 , so that the vectors @xmath83 all have length @xmath85
. moreover the lines spanned by the @xmath83 meet each other with angles @xmath86 or @xmath87 . in the language of @xcite
we have represented our signed graph in a line system , and if the graph is connected then the line system is indecomposable . if we change the sign of one of our gram vectors , then the line that it spans is unchanged , and the new gram matrix is equivalent to the old one : we have just changed the sign of all edges incident with the vertex that corresponds to our gram vector . since we are working up to equivalence , we can fix ( at our discretion ) the direction of each line in our system .
indecomposable line systems have been classified .
every such line system is contained in a maximal one .
it follows that every cyclotomic connected signed graph is contained in a maximal one .
moreover we can hunt for these by looking inside the maximal indecomposable line systems .
these are @xmath88 ( @xmath89 ) and @xmath90 , which we now describe .
fix @xmath91 , and let @xmath92 , ,
@xmath93 be an orthonormal basis for @xmath94 .
the signed graph @xmath95 has @xmath96 vertices , represented by the vectors @xmath97 adjacency of unequal vertices is given by the dot product of the corresponding vectors , which always equals one of 0 , 1 , @xmath12 .
if @xmath2 is the adjacency matrix of @xmath88 , then @xmath80 is the gram matrix of the set of vectors .
let @xmath92 , , @xmath98 be an orthogonal basis for @xmath99 , where , in contrast to the previous subsection , each @xmath100 has length @xmath85 .
the signed graph @xmath101 has 120 vertices , represented by the vectors @xmath92 , , @xmath98 and 112 vectors of the form @xmath102 where @xmath103 is one of the 14 strings @xmath104 ( the referee has pointed out that these strings are the supports of the nontrivial words in the extended binary hamming code of length @xmath105 . ) as for @xmath88 , adjacency of unequal vertices
is given by the dot product ( one of 0 , 1 , @xmath12 ) .
as a notational convenience , the vertices of @xmath90 will be written as strings of digits , some of them overlined .
single digits @xmath106 refer to the basis vectors @xmath92 , , @xmath98 .
strings of four digits , with any of the last three overlined , refer to the vectors @xmath107 , with overlining indicating a minus sign . for example , @xmath108 indicates the vector @xmath109 .
we sum up this discussion with the following result . for the proof one
trivially adapts to signed graphs the argument for graphs in chapter 3 of @xcite , noting that the fact that we can have negative edges makes the argument significantly easier .
[ t : dnore8 ] up to equivalence , the only ( neutral ) connected signed graphs that have all their eigenvalues in @xmath110 are the connected subgraphs of @xmath88 ( @xmath91 ) and of @xmath90
. signed graphs with all their eigenvalues in @xmath110 have been studied earlier by vijayakumar @xcite , singhi and vijayakumar @xcite and ray - chaudhuri , singhi and vijayakumar @xcite .
in this section we prove theorem [ t - csg ] , and so classify all cyclotomic signed graphs .
the plan is as follows .
first we find all the connected cyclotomic signed graphs that contain triangles ( triples of vertices with each pair being adjacent ) .
then , in view of proposition [ t : dnore8 ] , it suffices to consider triangle - free subgraphs of @xmath88 and @xmath90 .
we find all maximal triangle - free subgraphs of @xmath88 , and observe the remarkable fact that they are all cyclotomic .
we then find all maximal triangle - free subgraphs of @xmath90 : these are not all cyclotomic , and so we need to search among their subgraphs for any new maximal connected cyclotomic signed graphs that had not already been found as subgraphs of some @xmath95 .
[ l : addtotriangle ] suppose that @xmath48 is a cyclotomic signed graph that contains a triangle on vertices @xmath25 , @xmath111 , @xmath112 ( the signs of the three edges being arbitrary ) .
if @xmath113 is a fourth vertex in @xmath48 then @xmath113 is a neighbour of an even number of @xmath25 , @xmath111 , @xmath112 .
direct computation of the small number of cases .
one finds that if @xmath113 is a neighbour of one or three of @xmath25 , @xmath111 , @xmath112 then the subgraph induced by @xmath25 , @xmath111 , @xmath112 , @xmath113 is not cyclotomic , contradicting @xmath48 being cyclotomic , by interlacing . if @xmath113 is a neighbour of exactly two of @xmath25 , @xmath111 , @xmath112 , then the subgraph induced by @xmath25 , @xmath111 , @xmath112 , @xmath113 is not always cyclotomic , and the next lemma describes the extra condition on the signs of the edges that is required for a cyclotomic graph .
[ l : parity ] if @xmath48 is a cyclotomic signed graph containing two triangles that share an edge , then one triangle has an even number of negative edges , and the other has an odd number of negative edges .
if two triangles share an edge and the parities of the numbers of negative edges in the two triangles are equal , then one quickly checks that a suitable equivalence will make all the edges on both triangles positive .
but then the subgraph induced by the two triangles is not cyclotomic ( it has @xmath114 as an eigenvalue ) , and by interlacing neither is @xmath48 .
[ c : triplysharededge ] if @xmath48 is a cyclotomic signed graph , then no three triangles can share a single edge .
[ c : tetrahedron ] if @xmath48 is a cyclotomic signed graph , then it does not contain a tetrahedron as an induced subgraph .
this latter corollary also follows from lemma [ l : addtotriangle ] .
[ l : triangle ] if @xmath48 is a connected cyclotomic signed graph that contains a triangle , then it is equivalent to a subgraph of the signed octahedron @xmath36 of figure [ f - oct ] .
suppose that @xmath48 is a connected cyclotomic signed graph that contains a triangle , on vertices @xmath59 , @xmath115 , @xmath116 . by a suitable equivalence
, we may suppose that the three edges of this triangle are all positive .
if @xmath48 contains no other vertices then we are done .
otherwise suppose that @xmath117 is another vertex of @xmath48 , joined to @xmath59 , say . by lemma [ l : addtotriangle ] , @xmath117 is adjacent to exactly one other of the @xmath56 .
relabelling if necessary , we suppose that @xmath117 is adjacent to @xmath59 and @xmath115 . if @xmath48 contains no other vertices then we are done . otherwise @xmath48 contains a fifth vertex @xmath118 , adjacent to at least one of @xmath59 , @xmath115 , @xmath116 , @xmath117 . by lemma [ l : addtotriangle ]
, @xmath118 is adjacent to two vertices on one of the triangles @xmath119 , @xmath120 , and hence is adjacent to one of @xmath59 or @xmath115 . by corollary [ c : triplysharededge ] , @xmath118 can not be adjacent to both @xmath59 and @xmath115 . without loss of generality ,
@xmath118 is adjacent to @xmath59 . by lemma
[ l : addtotriangle ] ( using triangles @xmath119 and @xmath120 ) , @xmath118 is also adjacent to both @xmath116 and @xmath117 . if @xmath48 contains no other vertices then we are done .
otherwise @xmath48 contains a sixth vertex , @xmath121 , adjacent to one of @xmath115 , @xmath116 , @xmath117 , @xmath118 ( it can not be adjacent to @xmath59 , or else by lemma [ l : addtotriangle ] it would be adjacent to one of the others , producing three triangles sharing an edge , contrary to corollary [ c : triplysharededge ] ) .
applying lemma [ l : addtotriangle ] repeatedly , we see that @xmath121 must be adjacent to all of @xmath115 , @xmath116 , @xmath117 , @xmath118 .
we now have a subgraph of @xmath48 that is equivalent to the signed octahedron pictured in the lemma ( by lemma [ l : parity ] the parity of the number of negative edges on faces sharing an edge must differ , and up to equivalence one sees that there is just one choice of signs ) .
finally , @xmath48 can have no more vertices , as each existing triangle shares each of its edges with another : we can not adjoin a new vertex in a way that is compatible with both lemma [ l : addtotriangle ] and corollary [ c : triplysharededge ] .
[ c : degreebound ] in a cyclotomic signed graph @xmath48 , each vertex has degree at most @xmath122 .
if @xmath48 contains a triangle then it is equivalent to a subgraph of the signed octahedron , and hence has maximal degree at most 4 .
we may therefore assume that @xmath48 is triangle - free .
if @xmath48 has a vertex @xmath25 of degree at least 5 , then @xmath25 has neighbours @xmath59 , , @xmath118 say ( and possibly others ) , and since @xmath48 is triangle - free there are no edges between any pair of @xmath59 , , @xmath118 . by computation the starlike subgraph induced by @xmath25 , @xmath59 , , @xmath118 is not cyclotomic ( up to equivalence all the edges are positive , so there is only one case to compute ) .
this contradicts @xmath48 being cyclotomic , by interlacing .
after lemma [ l : triangle ] , our search for connected cyclotomic signed graphs can be restricted to triangle - free connected cyclotomic signed graphs .
after proposition [ t : dnore8 ] we can hunt for these triangle - frees as subgraphs of one of the @xmath95 , or of @xmath90 . here
we deal with the @xmath88 , classifying all the maximal triangle - free subgraphs .
fortunately for us ( in view of our ultimate goal ) these subgraphs are all cyclotomic . for @xmath123
so that @xmath124 define the _ conjugate _ vertex @xmath125 to be @xmath126 . if @xmath127 , then we say that @xmath25 _ includes _ @xmath100 and @xmath128 .
note that @xmath25 and @xmath125 have the same neighbours in @xmath88 .
[ l - conjugates ] let @xmath48 be a maximal triangle - free subgraph of @xmath88 .
if @xmath25 is a vertex of @xmath48 , then so is @xmath125 .
if @xmath129 is a triangle in @xmath48 , then so is @xmath130 .
hence if @xmath48 contained @xmath25 but not @xmath125 we could add @xmath125 to the vertex set and get a larger triangle - free signed graph , contradicting the maximality of @xmath48 .
[ l - atmost4 ] let @xmath48 be a maximal triangle - free subgraph of @xmath88 .
each @xmath100 is included in at most four vertices of @xmath48 .
if @xmath100 is included at all , then take a vertex @xmath25 including @xmath100 and @xmath128 ( @xmath131 ) .
suppose first that there exists a vertex @xmath111 in @xmath48 that includes @xmath100 and @xmath132 for some other @xmath133 ( @xmath134 , @xmath135 )
. then if @xmath112 is a vertex of @xmath48 that includes @xmath100 and @xmath136 ( @xmath137 ) we must have either @xmath138 or @xmath139 , or else @xmath140 would be a triangle .
hence @xmath100 is included exactly four times , in @xmath25 , @xmath111 , @xmath125 , @xmath141 .
if no such @xmath111 exists , then @xmath100 is included in exactly two vertices , @xmath25 and @xmath125 .
[ l - exactly4 ] let @xmath48 be a maximal triangle - free subgraph of @xmath88 .
the maximum degree of @xmath48 is at most 4 .
moreover if a vertex @xmath25 in @xmath48 has distinct neighbours @xmath142 and @xmath143 with @xmath144 , then @xmath25 has four neighbours , @xmath142 , @xmath143 , @xmath145 , @xmath146 .
take any vertex @xmath25 in @xmath48 . by relabelling , and moving to @xmath125 if necessary
, we can suppose that @xmath147 .
let @xmath111 be a neighbour of @xmath25 . again after relabelling , and so on
, we can suppose that @xmath148 .
then @xmath141 is also a neighbour of @xmath25 .
if @xmath25 has a third neighbour @xmath112 , then , in the same way , we can suppose that @xmath149 .
note that @xmath112 can not include @xmath150 , by lemma [ l - atmost4 ] .
then @xmath151 is a fourth neighbour of @xmath25 .
by lemma [ l - atmost4 ] again , there can be no more neighbours , as these would have to include either @xmath92 or @xmath150 , both of which have been included four times already ( in @xmath25 , @xmath125 , @xmath112 , @xmath151 and in @xmath25 , @xmath125 , @xmath111 , @xmath141 respectively ) .
recall that a path @xmath152 in @xmath48 is a sequence of distinct vertices @xmath56 in @xmath48 with @xmath56 adjacent to @xmath57 for @xmath153 .
[ l - maxpathiscycle ] let @xmath48 be a maximal connected triangle - free subgraph of @xmath95 , where @xmath154 .
let @xmath155 be a path in @xmath48 , maximal subject to no @xmath56 equalling any @xmath156 .
then * @xmath59 and @xmath157 are adjacent , so that the induced subgraph on the vertices of @xmath158 is a cycle . *
@xmath159 is a path in @xmath48 disjoint from @xmath158 , and @xmath48 is the subgraph spanned by @xmath158 and @xmath160 .
first suppose that @xmath59 and @xmath157 are not adjacent .
by lemma [ l - conjugates ] , @xmath160 is a subgraph of @xmath48 .
no vertex in @xmath158 can have more than two neighbours in @xmath158 , else together with its neighbours in @xmath160 it would have more than four neighbours in @xmath48 , contradicting lemma [ l - exactly4 ] . without loss of generality ,
@xmath161 , @xmath162 , , @xmath163 , @xmath164 . now for @xmath165 , @xmath56 has neighbours @xmath166 , @xmath57 , @xmath167 , @xmath168 , so has no other neighbours in @xmath48 , by lemma [ l - exactly4 ] . by maximality of @xmath158 , @xmath59 and @xmath157
have no neighbours in @xmath48 that are not in @xmath158 or @xmath169 , so @xmath158 and @xmath160 span a component of @xmath48 , and hence span @xmath48 .
but then we could add @xmath170 to @xmath48 without introducing triangles , contradicting maximality of @xmath48 .
thus @xmath59 and @xmath157 are adjacent , and without loss of generality @xmath161 , @xmath171 , , @xmath163 , @xmath172 . now each element of @xmath173 has four neighbours in @xmath173 , so no others , and again @xmath158 and @xmath160 span the whole of @xmath48 .
the proof of lemma [ l - maxpathiscycle ] establishes the first sentence of the next result .
[ justabove ] every maximal connected triangle - free signed graph that is a subgraph of some @xmath88 ( @xmath174 ) is equivalent to one with vertex set of the form @xmath175 for some @xmath176 in the range @xmath177 . moreover ,
every such graph is cyclotomic , and is a maximal connected cyclotomic signed graph . it remains to prove that such a graph @xmath48 is cyclotomic ( maximality as a connected cyclotomic signed graph follows from corollary [ c : degreebound ] ) . for @xmath178 one
gets this by computation ( or an easy adaptation of the following argument ) . for @xmath179 , note that if @xmath25 and @xmath180 are distance 2 apart in @xmath48 , then there are exactly two 2-paths from @xmath25 to @xmath111 , one along edges of the same sign , and one along edges of opposite sign .
there are four 2-paths from @xmath25 to @xmath125 , two along edges of the same sign , and two along edges of opposite sign .
hence ( with @xmath2 the adjacency matrix of @xmath48 ) all off - diagonal entries of @xmath181 are zero .
since each vertex has degree 4 , we deduce that @xmath182 . hence all the eigenvalues are either @xmath5 or @xmath79 , so @xmath48 is cyclotomic .
a nice representative of the equivalence class of the maximal cyclotomic signed graph , denoted @xmath183 in the theorem , described in proposition [ justabove ] is obtained by replacing the vertex @xmath184 by @xmath185 .
then one of the @xmath49-cycles ( say @xmath186 ) has all positive edges , and the other ( @xmath187 ) has all negative edges .
the linking edges of the form @xmath188 ( interpreted cyclically ) are all positive , and those of the form @xmath189 are all negative .
one gets a nice picture if the two cycles are viewed as the ends of a cylinder .
alternatively , the graph can be drawn on a torus without crossings , wrapping each cycle round the torus in such a way that it can not be shrunk to a point ( as in figure [ f - torus ] ) .
the search for triangle - free subgraphs of @xmath90 ( up to equivalence ) was done by computer , using moderately intelligent backtracking .
a lexicographical ordering was given to the 120 vertices , and a set of equivalences of @xmath90 was precomputed ( each as an explicit permutation of the 120 vertices ) , as follows .
we can change the sign of any @xmath100 : for any @xmath190 , we can swap the roles of @xmath14 and @xmath191 , which preserves all dot products . then flip the sign of any vector that is no longer a vertex of @xmath90 to induce an equivalence of @xmath90 .
if @xmath48 is a signed subgraph of @xmath90 , then applying this process gives an equivalent ( but perhaps different ) subgraph .
some , but not all , permutations of @xmath192 induce a permutation of the lines spanned by the vertices of @xmath90 ( and hence induce a permutation of the vertices of @xmath90 ) . for any string @xmath103 that appears as a vertex
, we can apply elements of the klein 4-group acting on @xmath193 to induce a permutation of the vertices of @xmath90 .
for example , if we apply @xmath194 to the vertex @xmath195 we get the vector @xmath196 , which spans the same line as the vertex @xmath197 , so the image of @xmath195 under @xmath194 is @xmath197 .
note that such a transformation might not be an isomorphism of signed graphs ( since some of the vertices may be switched ) but will be an equivalence .
again , applying this to a subgraph of @xmath90 will give an equivalent subgraph .
also , if @xmath103 is a vertex of @xmath90 , then we can perform a change of basis by the following four swaps : @xmath198 , @xmath199 , @xmath200 , @xmath201 .
this induces an equivalence on @xmath202 .
( it is enough to check that this works for @xmath203 , and then use the previous symmetries to reduce to this case . )
starting with @xmath204 being empty , the search grew @xmath204 by adding the smallest possible vertices ( with respect to the chosen ordering ) whilst ( i ) maintaining triangle - freeness , and ( ii ) checking that none of the above equivalences of @xmath90 would map the enlarged @xmath204 to a lexicographically earlier set .
the use of equivalences was hugely powerful in cutting down on the number of sets @xmath204 considered by rejecting most sets at an early stage .
when no more vertices could be added , the set @xmath204 was tested for maximality , and maximal triangle - frees were written to a file . then
backtracking was done to find the next candidate for @xmath204 .
the following twenty inequivalent maximal triangle - free subgraphs were found .
* @xmath205 + this is two copies of the toral tesselation @xmath206 .
* @xmath207 + this comprises two isolated vertices plus @xmath208 . *
@xmath209 + this is @xmath210 . * @xmath211 + this is the hypercube @xmath34 .
* @xmath212 + this is an isolated vertex plus @xmath32 . * @xmath213 + this is @xmath214 . * @xmath215 + this is a square plus @xmath216 . *
@xmath217 + 10 vertices , 2 cyclotomic components ( both are 5-cycles ) . by corollary
[ c : degreebound ] , @xmath32 and @xmath34 are maximal . in the remaining cases ,
the larger component was noncyclotomic .
* @xmath218 + 12 vertices , 1 component , 29 maximal cyclotomic subgraphs ( maximal in the sense that no larger subgraph of * g9 * is cyclotomic ) .
* @xmath219 + 12 vertices , 1 component , 13 maximal cyclotomic subgraphs . * @xmath220 + 11 vertices , 2 components ( one being a single vertex ) , 15 maximal cyclotomic subgraphs . * @xmath221 + 12 vertices , 1 component , 15 maximal cyclotomic subgraphs .
* @xmath222 + 12 vertices , 1 component , 19 maximal cyclotomic subgraphs . * @xmath223 + 12 vertices , 1 component , 17 maximal cyclotomic subgraphs . * @xmath224 + 12 vertices , 1 component , 37 maximal cyclotomic subgraphs . * @xmath225 + 11 vertices , 1 component , 44 maximal cyclotomic subgraphs . * @xmath226 + 11 vertices , 2 components : @xmath227 plus a 9-vertex component ; 36 maximal cyclotomic subgraphs . * @xmath228 + 12 vertices , 1 component , 3-regular , 45 maximal cyclotomic subgraphs . * @xmath229 + 12 vertices , 2 components : @xmath227 plus a 3-regular 10-vertex component , equivalent to the petersen graph ( switch at vertex @xmath122 to get it ) 57 maximal cyclotomic subgraphs . *
@xmath230 + 10 vertices , 1 component , 23 maximal cyclotomic subgraphs . for each of the noncyclotomic components listed above , it was checked by computer that
none of their cyclotomic subgraphs are maximal cyclotomic graphs .
this completes the proof of theorem [ t - csg ] .
let @xmath48 be a cyclotomic subgraph of @xmath90 that is not equivalent to a subgraph of any @xmath231 , with vertices given by vectors @xmath232 ,
, @xmath233 , contained in 8-dimensional real space .
then @xmath234 , obtained from @xmath48 by changing the signs of all edges and charges , is also cyclotomic .
now @xmath234 is equivalent to @xmath48 , so can not be represented in any line system @xmath231 , so must be represented in the line system @xmath90 , and hence the vertices of @xmath234 can be represented as vectors @xmath81 ,
, @xmath82 , where for each @xmath14 either @xmath83 or @xmath235 is in the signed graph @xmath90 .
we can view the concatenated vectors @xmath236 $ ] , , @xmath237 $ ] as elements of 16-dimensional real space , a subset of the 28800 vectors @xmath238 $ ] where @xmath239 , @xmath240 .
moreover , since the @xmath83 represent @xmath234 , we have @xmath241 for all @xmath242 .
this implies that @xmath243\cdot[{\bf v}_j,{\bf w}_j ] = 0\ ] ] for all @xmath242 : our concatenated vectors @xmath236 $ ] , , @xmath237 $ ] are pairwise orthogonal .
since these vectors lie in 16-dimensional space , we must have @xmath244 , as is confirmed by the examples computed in section [ s : e8tfree ] .
conversely , suppose that we take any orthogonal subset @xmath245 $ ] , , @xmath237 $ ] of the 28800 vectors considered above , with the constraint that @xmath232 , , @xmath233 are distinct .
then the signed graph @xmath48 with vertices @xmath232 , , @xmath233 ( and adjacency of unequal vertices given by the dot product ) is cyclotomic , for both @xmath48 and @xmath234 are represented in the line system @xmath90 , with @xmath81 , , @xmath82 spanning the lines that represent @xmath234 .
the maximal cyclotomic graphs classified by smith are shown in figure [ f - except ] .
the @xmath49-cycle @xmath246 and the graph @xmath247 are subgraphs of @xmath183 , while the sporadic examples are all subgraphs of the hypercube @xmath34 . unlike in the signed case , however , the maximal unsigned graphs are not visibly cyclotomic .
[ l]@xmath248 [ l]@xmath249 [ l]@xmath250 and @xmath251 .
the number of vertices is @xmath13 more than the index .
( from @xcite).,title="fig : " ] we can deduce smith s classification as a corollary of theorem [ t - csg ] , by checking that these graphs are the only maximal ( unsigned ) subgraphs of the signed graphs of the theorem . a useful fact to use in this check is that the graphs @xmath252 and @xmath253 , since they have @xmath5 as an eigenvalue , can not be a proper subgraph of any such graph .
this is because otherwise the graph would have an eigenvalue greater than @xmath5 see @xcite .
we also note in passing that the classification of all graphs having all their eigenvalues in the open interval @xmath1 follows from smith s result .
such a graph is either a subgraph of @xmath254 or of some @xmath255 for @xmath256 ( figure [ f - p ] ) . here , @xmath254 is @xmath257 ( figure [ f - except ] ) with its rightmost vertex removed ( same as @xmath258 in figure [ f - u ] ) , and @xmath255 is @xmath247 with a leaf removed . see also theorem [ t : nonnegative_matrix_open ] below for a generalisation of this result .
we now embark upon the trickier task of proving theorem [ t - lcsg ] , and so classifying all cyclotomic charged signed graphs .
the addition of charges means that we can no longer appeal to proposition [ t : dnore8 ] , although the gram matrix approach will still prove extremely powerful . by interlacing
, every subgraph of a cyclotomic charged signed graph is cyclotomic .
we can therefore exclude as subgraphs any that are not cyclotomic .
in particular , the following eight non - cyclotomic charged signed graphs @xmath259 of figure [ f - x ] ( or anything equivalent to any of them ) can not be subgraphs of any cylotomic charged signed graph .
certain cyclotomic charged signed graphs have the property that if one tries to grow them to give larger connected cyclotomic graphs then one always stays inside one of the maximal examples on the following list : @xmath260 , @xmath261 , @xmath262 , @xmath263 , @xmath264 , @xmath265 , @xmath266 , @xmath36 . the process of proving that a cyclotomic graph has this property is in principle simple , although perhaps tedious , to carry out .
starting from the given graph , one considers all possible ways of adding a vertex ( up to equivalence ) such that the graph remains connected and cyclotomic .
check that the resulting graphs are ( equivalent to ) subgraphs of one of graphs on this list .
repeat with all the larger graphs found . if the checks in this process are always valid , then , since the process terminates , the original graph is suitable for exclusion . by this technique ,
the six cyclotomic graphs @xmath267 of figure [ f - y ] ( and anything equivalent to them ) can be excluded from future consideration .
let @xmath48 be a charged signed graph .
we define the _ charged subgraph _ of @xmath48 to be the subgraph induced by all its charged vertices , and the _ neutral subgraph _ of @xmath48 to be the subgraph induced by all its neutral vertices .
the components of the charged ( respectively neutral ) subgraph of @xmath48 will be called the _ charged components _ of @xmath48 ( respectively the _
neutral components _ of @xmath48 ) .
our next task will be to show that the charged components of a cyclotomic charged signed graph are tiny , provided that @xmath48 does not contain @xmath268 , @xmath269 , or any equivalent subgraph .
[ l : chargedcomponents ] let @xmath48 be a cyclotomic charged signed graph that does not contain any subgraph equivalent to @xmath268 or @xmath269 of section [ s : excludedii ] .
then each charged component of @xmath48 contains at most two vertices , necessarily of the same charge .
the last phrase is clear , since @xmath268 is excluded as a subgraph .
moreover the exclusion of @xmath268 forces every charged component to have all charges of the same sign , which by equivalence we may assume to be all positive . since graphs @xmath270 and @xmath271 of section [ s : excludedi ] are not cyclotomic , and @xmath269 is excluded by assumption , no charged component of @xmath48 can have as many as three vertices . [
l : localgeometry1 ] let @xmath48 be a cyclotomic charged signed graph .
suppose that @xmath48 contains two nonadjacent neutral vertices @xmath25 and @xmath111 that have a charged vertex @xmath112 as a common neighbour .
then @xmath25 and @xmath111 have the same neighbours .
adjacency being unchanged by equivalence , we may suppose that the charge on @xmath112 has negative sign , and that the edges joining @xmath25 and @xmath111 to @xmath112 are positive .
the subgraph induced by @xmath25 , @xmath111 , @xmath112 is then . since @xmath48 is cyclotomic , all eigenvalues of its adjacency matrix @xmath2 are in @xmath110 , so @xmath80 is the gram matrix of some set of vectors .
let @xmath272 , @xmath273 , @xmath274 be the gram vectors corresponding to @xmath25 , @xmath111 , @xmath112
. since @xmath25 and @xmath111 are neutral , @xmath272 and @xmath273 have length @xmath85 . since @xmath112 has a negative charge , @xmath274 has length 1 .
the angle between @xmath272 and @xmath274 is @xmath275 , the angle between @xmath273 and @xmath274 is @xmath275 , and the angle between @xmath272 and @xmath273 is @xmath87 .
hence @xmath272 , @xmath273 , @xmath274 are coplanar , with @xmath274 in the direction of @xmath276 . by consideration of their lengths
we have @xmath277 now let @xmath278 be any other vertex of @xmath48 , with corresponding gram vector @xmath279 .
taking dot products with ( [ xvw ] ) gives @xmath280 the left hand side of ( [ ydotx ] ) is an even integer , hence the parities of the two integers on the right must agree . hence @xmath278 is adjacent to @xmath25 if and only if it is adjacent to @xmath111 . [
l : localgeometry2 ] let @xmath48 be a cyclotomic charged signed graph containing adjacent charged vertices @xmath25 and @xmath111 , where the signs on the charges for @xmath25 and @xmath111 agree . then @xmath25 and @xmath111 have the same neighbours
. adjacency is preserved by equivalence , so we may suppose that the charges on @xmath25 and @xmath111 are both negative , and that the edge between @xmath25 and @xmath111 is positive .
in the usual way , let @xmath272 and @xmath273 be gram vectors corresponding to @xmath25 , @xmath111 .
these have length 1 , and the angle between them is zero , so @xmath281 , although @xmath282 .
hence @xmath25 and @xmath111 have the same neighbours .
we now show that if a connected cyclotomic charged signed graph does not contain a subgraph equivalent to any of the excluded subgraphs of section [ s : excludedii ] , then it has a single neutral component . as a first step ,
we show that certain charged vertices can be deleted without disconnecting the graph .
[ l : removecharge1 ] suppose that a connected cyclotomic charged signed graph @xmath48 has two adjacent charged vertices @xmath25 and @xmath111 , with the charges on @xmath25 and @xmath111 having the same sign .
then the vertex @xmath111 can be deleted without disconnecting @xmath48 . by lemma [ l : localgeometry2 ]
every neighbour of @xmath25 is a neighbour of @xmath111 ( and vice versa ) .
let @xmath112 and @xmath278 be any distinct vertices in @xmath48 , with neither of them being @xmath111 .
we must show that there is a walk in @xmath48 from @xmath112 to @xmath278 that does not pass through @xmath111 .
certainly there is a path @xmath283 in @xmath48 from @xmath112 to @xmath278 ( @xmath284 , @xmath285 ) .
suppose that this path contains @xmath111 , say @xmath286 .
if either @xmath166 or @xmath57 is @xmath25 , then we can simply remove @xmath111 from the path , since @xmath25 shares its neighbours .
otherwise we can replace @xmath111 by @xmath25 in the path ( producing a walk , but perhaps no longer a path ) , again since @xmath25 and @xmath111 share their neighbours .
[ l : removecharge2 ] let @xmath48 be a connected cyclotomic charged signed graph , with more than three vertices , that contains no subgraph equivalent to either @xmath268 or @xmath287 of section [ s : excludedii ] .
suppose that @xmath48 contains two nonadjacent neutral vertices @xmath25 and @xmath111 that share a common charged neighbour @xmath112 ( as in lemma [ l : localgeometry1 ] ) .
then @xmath112 can be deleted from @xmath48 without disconnecting the graph . by the hypothesis on the number of vertices in @xmath48 ,
there is some fourth vertex @xmath278 in @xmath48 that is adjacent to one of @xmath25 , @xmath111 , @xmath112 .
first we dispose of the cases where @xmath278 is adjacent to @xmath112 . if @xmath278 has a charge , then since @xmath268 is an excluded subgraph , @xmath278 and @xmath112 have charges of the same sign
. then lemma [ l : removecharge1 ] shows that @xmath112 can be removed without disconnecting @xmath48 . if @xmath278 is neutral , then since subgraphs equivalent to @xmath287 are excluded , and a subgraph equivalent to @xmath288 of section [ s : excludedi ] is impossible , @xmath278 can not be adjacent to either @xmath25 or @xmath111 .
but then @xmath48 would contain a subgraph equivalent to @xmath289 of section [ s : excludedi ] , which is not possible .
we may now suppose that @xmath278 is not adjacent to @xmath112 , and more strongly may suppose that @xmath25 and @xmath111 are the only neighbours of @xmath112 . by lemma [ l : localgeometry1 ] , @xmath25 and @xmath111 share all their neighbours .
in particular , @xmath278 is adjacent to both @xmath25 and @xmath111 .
let @xmath290 and @xmath291 be any vertices in @xmath48 other than @xmath112 .
it is enough to show that there is a walk in @xmath48 from @xmath290 to @xmath291 that does not pass through @xmath112 .
certainly there is a path @xmath283 from @xmath290 to @xmath291 ( @xmath292 , @xmath293 ) .
suppose that @xmath112 is on this path : say @xmath294 .
we know that @xmath166 and @xmath57 each equal one of @xmath25 and @xmath111 .
we can therefore replace @xmath112 by @xmath278 in our path to produce the desired walk .
the requirement that @xmath48 has more than three vertices is clearly necessary : if @xmath25 , @xmath111 , @xmath112 are the only vertices in @xmath48 then deleting @xmath112 disconnects @xmath48 .
[ l : singleneutralcomponent ] let @xmath48 be a connected cyclotomic charged signed graph that does not contain a subgraph equivalent to @xmath268 , @xmath287 or @xmath269 of section [ s : excludedii ] .
suppose further that @xmath48 has at least four vertices .
then @xmath48 contains a single neutral component : all charged vertices can be deleted without disconnecting @xmath48 . by lemma
[ l : chargedcomponents ] , all charged components have at most two vertices , and do not equal @xmath268 . by lemma
[ l : removecharge1 ] , we can remove a charged vertex from any charged component that has two vertices , without disconnecting @xmath48 .
we are thus reduced to charged components containing only one vertex .
if a charged vertex is a leaf , it can be removed without disconnecting @xmath48 .
if a charged vertex has two neutral neighbours , then since subgraphs equivalent to @xmath288 and @xmath287 are excluded we can appeal to lemma [ l : removecharge2 ] to see that this vertex can be removed without disconnecting @xmath48 .
no charged vertex can have three or more neutral neighbours , or @xmath48 would contain a subgraph equivalent to one of @xmath288 , @xmath289 or @xmath287 .
let @xmath48 be a connected cyclotomic charged signed graph that contains at least four vertices , at least one of which is charged , but does not contain any of the excluded subgraphs @xmath268 ,
, @xmath269 .
then lemma [ l : singleneutralcomponent ] tells us that @xmath48 has a single neutral component , @xmath295 say . by interlacing ,
@xmath295 is cyclotomic , and from the classification of all cyclotomic signed graphs we know that @xmath295 is ( equivalent to ) a subgraph of one of @xmath231 ( for some @xmath54 ) , @xmath32 or @xmath34 .
we treat each of these cases in turn .
we may suppose that @xmath54 is minimal such that @xmath231 contains a subgraph equivalent to @xmath295 .
cases with @xmath296 can be dealt with exhaustively by growing each possible @xmath295 in all possible ways , adding only charged vertices , and checking that each maximal connected cyclotomic charged signed graph ( maximal subject to the neutral component being @xmath295 ) is contained in some @xmath297 or @xmath298 .
we may therefore suppose that @xmath299 .
working up to equivalence , we identify @xmath295 with some subgraph of @xmath231 ( which has vertices @xmath300 for @xmath301 , where @xmath92 , , @xmath302 is an orthonormal set of vectors ) . from our knowledge of the structure of
cyclotomic signed graphs , we see that by relabelling and changing signs of basis vectors ( thereby inducing an equivalence ) , we can suppose that @xmath295 contains @xmath303 , @xmath304 , , @xmath305 , and that all other vertices of @xmath295 are of the form @xmath306 ( for some @xmath14 in the range @xmath307 ) , or @xmath308 .
now suppose that @xmath111 is a charged vertex in @xmath48 .
since ( i ) @xmath48 is connected , ( ii ) @xmath268 is an excluded subgraph , and ( iii ) adjacent charged vertices that have the same charge share all their neighbours ( lemma [ l : localgeometry2 ] ) , we deduce that @xmath111 is adjacent to one or more vertices in @xmath295 .
we treat first the case where @xmath111 has charge @xmath12 .
we have represented ( a graph equivalent to ) @xmath295 by a set of gram vectors , where adjacency of unequal vertices is given by the dot product , and we can extend this to ( a graph equivalent to ) @xmath309 , where @xmath111 is represented by the gram vector @xmath310 if @xmath311 is in the span of @xmath92 , , @xmath302 , then we may set @xmath312 ; otherwise we need an extra dimension for @xmath311 , and take @xmath313 of length 1 and orthogonal to all of @xmath92 , , @xmath302 .
since @xmath111 has charge @xmath12 , @xmath311 has length @xmath13 .
we consider two subcases .
case 1 ( which we shall prove to be impossible ) : @xmath295 contains one or both of @xmath308 , so that @xmath295 contains a cycle of length @xmath54 containing no pair of conjugate vertices . case 2 : @xmath295 contains neither of the vertices @xmath308 . in case 1
, @xmath295 contains at least one cycle of length @xmath54 containing no pair of conjugate vertices .
suppose that @xmath111 were adjacent to at least two vertices on such a cycle , say @xmath112 and @xmath278 ( and perhaps others )
. since subgraphs of @xmath48 equivalent to @xmath287 have been excluded , and @xmath48 can not contain a subgraph equivalent to @xmath288 , the vertices @xmath112 and @xmath278 are not adjacent . by lemma [ l : localgeometry1 ] , every neighbour of @xmath112 is a neighbour of @xmath278 .
but in a cycle of length at least 5 containing unadjacent vertices @xmath112 and @xmath278 and containing no pair of conjugate vertices , there will be a neighbour of @xmath112 that is not a neighbour of @xmath278 . still in case 1 , suppose next that @xmath111 is adjacent to exactly one vertex in some cycle of length @xmath54 containing no pair of conjugate vertices .
then @xmath48 would contain a subgraph equivalent to @xmath314 , giving a contradiction . to kill off case 1 , we now consider the remaining subcase where @xmath111 is adjacent to none of the vertices in the cycle @xmath315 , @xmath316 , , @xmath305 , @xmath308
. then @xmath295 must contain at least one more vertex , and after some relabelling and equivalence we can assume that @xmath111 is adjacent to @xmath317 , with a positive edge .
then @xmath318 where the ` @xmath319 ' might be ` @xmath320 ' if @xmath321 .
this gives @xmath322 and hence @xmath323 , giving a contradiction .
we now move to case 2 , where @xmath295 contains the path formed by the vertices @xmath303 , @xmath316 ,
, @xmath305 , and all other vertices in @xmath295 are of the form @xmath324 ( for some @xmath14 in the range @xmath307 ) .
if @xmath111 were adjacent to more than one vertex in our path , say @xmath112 and @xmath278 , then as in case 1 we would have @xmath112 and @xmath278 unadjacent , implying that they share all their neighbours , giving a contradiction .
if @xmath111 were not adjacent to any vertex in our path , then it would be adjacent to some @xmath324 , and from @xmath325 we would get at least five distinct @xmath15 such that @xmath326 , contradicting @xmath327 .
we are reduced to the case where @xmath111 is adjacent to exactly one vertex in our path .
since @xmath314 is excluded as a subgraph , this neighbour of @xmath111 must be an endvertex of our path .
relabelling , we can suppose that @xmath111 is attached to @xmath303 by a positive edge , but to none of @xmath316 ,
, @xmath305 . if @xmath295 also contained @xmath328 , then @xmath111 would necessarily be adjacent to it , or else @xmath48 would contain a subgraph equivalent to @xmath329 . moreover , as @xmath316 is joined to @xmath317 by a negative edge , exclusion of subgraphs equivalent to @xmath330 implies that @xmath111 must then be connected to @xmath328 by a positive edge . to sum up ,
if the minimal value of @xmath54 is at least 5 , then we can assume that @xmath295 contains the vertices @xmath303 , @xmath304 ,
, @xmath305 , and that all other vertices are of the form @xmath324 .
any negatively charged vertex @xmath111 in @xmath48 is adjacent to one of @xmath331 or @xmath332 .
if both of @xmath333 are in @xmath295 and @xmath111 is adjacent to one of them , then it is adjacent to both ; similarly for @xmath334 .
the excluded graph @xmath330 constrains the signs of the edges that connect @xmath111 to @xmath295 . in short , @xmath309 is equivalent to a subgraph of one of the @xmath297 or @xmath298 . by equivalence ,
similar remarks hold for positively - charged vertices in @xmath48 .
if more than one charged vertex in @xmath48 is adjacent to the same vertex in @xmath295 , then the exclusion of subgraphs equivalent to @xmath335 and @xmath336 implies that these charged vertices are adjacent to each other ; the exclusion of subgraph @xmath268 implies that they all have the same sign ; lemma [ l : localgeometry2 ] implies that there are at most two such . we conclude that @xmath48 is equivalent to a subgraph of one of the @xmath297 or @xmath298 .
we shall show that @xmath295 is in fact equivalent to a subgraph of @xmath231 for some @xmath54 , so that we are reduced to the previous case . recalling previous notation ,
the vertices of @xmath34 are labelled 1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 , @xmath337 , @xmath338 , @xmath339 , @xmath340 , @xmath341 , @xmath342 , @xmath343 , @xmath344 ( a trivial relabelling of * g4 * ) .
these are vectors in 8-dimensional real space , with adjacency of unequal vectors given by the dot product .
each vector has length @xmath345 .
our restrictions on @xmath48 imply that it has no triangles except perhaps involving two charged vertices and one neutral vertex .
note that @xmath34 is bipartite , with parts @xmath346 , @xmath347 , @xmath338 , @xmath339 , @xmath340 , @xmath341 , @xmath342 , @xmath343 , @xmath348 .
there is an equivalence of @xmath34 that interchanges these two parts , induced by the orthogonal map with matrix @xmath349 with respect to @xmath92 , , @xmath98 .
let @xmath111 be a charged vertex in @xmath48 .
arguing as before , @xmath111 is adjacent to at least one vertex in @xmath295 .
first we treat the case where @xmath111 is adjacent to at least two vertices in @xmath295 , say @xmath112 and @xmath278 . now @xmath112 and @xmath278 can not be adjacent in @xmath48 , or we would have a forbidden triangle equivalent to @xmath288 or @xmath287 .
then by lemma [ l : localgeometry1 ] the vertices @xmath112 and @xmath278 share all their neighbours ( and they must have at least one neighbour in @xmath295 or @xmath295 would not be connected ) .
it follows that @xmath112 and @xmath278 are either both in @xmath350 or both in @xmath351 . working up to equivalence , and swapping @xmath350 and @xmath351 as above if necessary
, we may suppose that @xmath112 and @xmath278 are both in @xmath350 .
we may also suppose that @xmath111 is negatively charged , so that if we extend our set of gram vectors representing @xmath295 ( some subset of the vectors / vertices in @xmath352 ) to a set of gram vectors representing @xmath309 , the vector @xmath311 representing @xmath111 will have length 1 .
we may write @xmath353 , where @xmath354 ( length @xmath85 , orthogonal to @xmath92 , , @xmath98 ) is included in case we need an extra dimension to make room for @xmath311 . since @xmath327 , we have @xmath355 if @xmath112 and @xmath278 correspond to @xmath14 and @xmath15 in our labelling of the vertices of @xmath34 , then from @xmath356 we have @xmath357 , and hence all other @xmath358 are zero . there are now essentially two cases ( up to equivalence ) : @xmath359 and @xmath360 .
indeed there are self - equivalences of @xmath34 induced by elements of the klein @xmath122-group acting on any of the six ` missing ' quartets @xmath361 , @xmath362 , @xmath363 , @xmath364 , @xmath365 , @xmath366 ( one then needs to apply appropriate further transpositions of the form @xmath367 , and some changes of signs of certain vertices , to map @xmath34 to itself ) .
we see that any vertex in @xmath350 can be mapped to @xmath13 by a self - equivalence of @xmath34 , and that with @xmath13 fixed , any vertex in @xmath368 can be mapped to @xmath5 .
in the case @xmath359 , since @xmath13 and @xmath5 are not adjacent , lemma [ l : localgeometry1 ] implies that they have the same neighbours in @xmath48 , and hence also in @xmath295 , whence @xmath339 , @xmath340 , @xmath341 , @xmath342 @xmath369 .
one of @xmath337 and @xmath338 has dot product @xmath370 with @xmath273 , and hence must be excluded from @xmath295 ( or else together with @xmath111 and 1 ( or 2 ) we would have a forbidden triangle ) .
hence the vertices in @xmath295 are a subset of @xmath371 , where @xmath372 depending on the signs of @xmath373 and @xmath374 .
then @xmath295 is readily seen to be equivalent to a subgraph of @xmath375 . for the other essentially distinct case , @xmath360
, similar reasoning shows that @xmath295 is a subset of @xmath376 , contradicting the connectedness of @xmath295 .
we are left with the possibility that @xmath111 is adjacent to exactly one vertex in @xmath295 .
let us temporarily call a signed charged graph @xmath377 _ friendly _ if it is cyclotomic , contains exactly one charged vertex @xmath111 , the vertex @xmath111 is joined to exactly one neutral vertex , and the neutral vertices in @xmath377 form a single component . in our current case
, @xmath309 is friendly
. it will be enough to show that any friendly graph with neutral component equivalent to a subgraph of either @xmath34 or @xmath32 is contained in a larger friendly graph ( where the neutral component of the larger friendly graph might or might not be equivalent to a subgraph of either @xmath34 or @xmath32 ) .
for then we can grow our friendly graph @xmath309 to a larger friendly graph @xmath378 with @xmath379 not equivalent to a subgraph of either @xmath34 or @xmath32
. then @xmath379 must be equivalent to a subgraph of some @xmath231 , and hence the same is true for @xmath295 .
a computer search checked that all friendly graphs with up to @xmath31 neutral vertices are contained in larger friendly graphs . as an indication of the work involved , some @xmath380 friendly graphs with @xmath381 vertices ( @xmath31 neutral vertices ) were considered ; these would not all have been inequivalent , as it proved more efficient to perform a fast but imperfect weeding out of equivalent graphs , allowing some repeats through .
the search could have been pushed further , but it was easier simply to check that there are no friendly graphs with @xmath381 or @xmath33 neutral vertices for which the neutral component is equivalent to a subgraph of @xmath34 .
the argument here is very similar to that for @xmath34 , but in fact slightly simpler , as @xmath32 has fewer vertices .
analogously , we have @xmath382 , @xmath383 . in the ` unfriendly ' case
we find that the vertices in @xmath295 are ( after a suitable equivalence ) a subset of @xmath384 where @xmath385 then @xmath295 is equivalent to a subgraph of @xmath386 .
this completes the proof of theorem [ t - lcsg ] .
sections [ s : open],[s : charged ] and [ s : nonneg ] are devoted to results for matrices and graphs under further restrictions .
these follow more or less straightforwardly from theorems [ t - csg ] and [ t - lcsg ] .
we consider first restricting to eigenvalues in the open interval @xmath1 ( section [ s : open ] ) , deferring the proofs to section [ s : openproofs ] .
then we consider charged ( unsigned ) graphs , treating both the open and closed intervals ( section [ s : charged ] ) .
finally we treat symmetric matrices that have non - negative integer entries ( section [ s : nonneg ] ) .
having classified all integer symmetric matrices having all their eigenvalues in the interval @xmath0 $ ] , a natural question is what happens if we restrict the eigenvalues to the open interval @xmath1 . from our knowledge of the closed interval case
, we can immediately restrict to cyclotomic signed graphs and cyclotomic charged signed graphs , and need only consider subgraphs of the maximal ones .
[ t : open_signed ] up to equivalence , the connected signed graphs maximal with respect to having all their eigenvalues in @xmath1 are the eleven @xmath105-vertex sporadic examples @xmath387 shown in figure [ f - u ] , and the infinite family @xmath388 of @xmath29-cycles with one edge of sign @xmath12 , for @xmath389 , shown in figure [ f - o ] .
further , every connected cyclotomic signed graph having all its eigenvalues in @xmath1 is either contained in a maximal one , or is a subgraph of one of the signed graphs @xmath390 of figure [ f - q ] for @xmath391 .
we note in passing that the graphs @xmath392 can all be obtained from the cube @xmath393 by deleting certain edges .
not every choice of edge - deletion produces a @xmath392 , however . for instance
no edge - deleted subgraph of @xmath393 containing an induced subgraph equivalent to @xmath394 can have all its eigenvalues in @xmath1 .
next we have a corresponding result for charged signed graphs .
[ t : open_charged_signed ] up to equivalence , the connected charged signed graphs maximal with respect to having all their eigenvalues in @xmath1 , and not covered by the theorem [ t : open_signed ] above , are the eight @xmath122-vertex sporadic examples @xmath395 shown in figure [ f - v ] , and the infinite family @xmath396 of @xmath49-vertex charged paths of figure [ f - p ] for @xmath154 .
further , every connected cyclotomic charged signed graph not covered by the previous theorem is contained in such a maximal one .
we can combine the previous two theorems , translated into matrix language , to obtain the following .
[ t - ismopen ] every indecomposable cyclotomic matrix maximal with repect to having all its eigenvalues in the open interval @xmath1 is equivalent to the adjacency matrix of one of the graphs @xmath397 , @xmath398 , @xmath399 , @xmath400 ( given by theorems [ t : open_signed ] and [ t : open_charged_signed ] ) .
further , every indecomposable cyclotomic matrix having all its eigenvalues in @xmath1 is either contained in a maximal one , or is contained in the adjacency matrix of one of the signed graphs @xmath390 of figure [ f - q ] for @xmath391 .
we now restrict our attention to cyclotomic charged graphs , looking for those that are maximal with respect to having all their eigenvalues in @xmath0 $ ] .
for such a graph @xmath48 we need to define @xmath401 as the graph whose edges are the same as those of @xmath48 , with the same signs , but with the charges on vertices being the opposite of those on @xmath48 .
for example , graphs @xmath402 and @xmath403 are shown in figure [ f - v ] .
it is clear that when @xmath48 is a tree , @xmath401 is equivalent to @xmath48 .
one difference for this kind of maximality is that it is not a property of equivalence classes of charged graphs : two of them may be equivalent with one of them maximal and the other not .
for instance , one consequence of the next result is that for the maximal graphs @xmath404 and @xmath405 the graphs @xmath406 and @xmath407 , being subgraphs of @xmath408 , are not maximal .
however , we have the following . [
t : unsigned_charged ] the maximal connected cyclotomic charged graphs not covered by smith s result ( i.e. not graphs ) , are the sporadic examples @xmath409 from figure [ f - mccs ] , along with @xmath410 , @xmath411,@xmath412 , and the seven families @xmath413 , @xmath414 , @xmath415 , @xmath416 , @xmath417 and @xmath418 , @xmath419 from figure [ f - mccf ] .
further , every connected cyclotomic charged graph is contained in such a maximal one .
the proof of this theorem is by inspection of the maximal connected cyclotomic charged signed graphs of theorem [ t - lcsg ] to find their maximal connected charged ( unsigned ) subgraphs .
we next have a corresponding result for eigenvalues in the open interval @xmath1 .
[ t : open_charged ] the connected charged ( unsigned ) graphs maximal with respect to having all their eigenvalues in @xmath1 , are the graph @xmath258 , the charged graphs @xmath403 , @xmath402 , @xmath420 , from figure [ f - v ] , and @xmath396 of @xmath49-vertex charged paths of figure [ f - p ] for @xmath154 .
further , every connected cyclotomic charged graph not covered by theorem [ t : open_signed ] is contained in one of the above graphs . note that for @xmath256 the graph @xmath255 ( figure [ f - p ] ) is a subgraph of some @xmath390 .
so it has all its eigenvalues in @xmath1 and is covered by theorem [ t : open_signed ] .
it is not contained in any charged graph maximal with respect to having all its eigenvalues in @xmath1 .
in this section we record our results for non - negative cyclotomic matrices , i.e. , those integer symmetric matrices that are cyclotomic and have only non - negative entries .
[ t : nonnegative_matrix ] up to conjugation by permutation matrices , the only maximal indecomposable non - negative cyclotomic matrices are the matrices @xmath46 and @xmath421 , adjacency matrices of @xmath422 , @xmath423 , @xmath424 , @xmath425 , @xmath426 ( figure [ f - except ] ) along with the two families @xmath416 and @xmath417 ( figure [ f - mccf ] ) .
further , every indecomposable non - negative cyclotomic matrix is contained in such a maximal one .
this result is readily deduced from theorems [ t - ism],[t : unsigned_charged ] and smith s results ( figure [ f - except ] ) .
[ t : nonnegative_matrix_open ] up to conjugation by permutation matrices , the only indecomposable non - negative cyclotomic matrix maximal with respect to having all its eigenvalues in @xmath1 is the adjacency matrix of @xmath258 ( figure [ f - u ] ) .
further , every indecomposable non - negative cyclotomic matrix is either contained in the adjacency matrix of @xmath258 or in the adjacency matrix of either @xmath427 or @xmath255 ( figure [ f - p ] ) for some @xmath49 .
to prove theorem [ t : open_signed ] , we first we show that the two infinite families @xmath428 , @xmath390 have their eigenvalues in the open interval .
suitable sets of gram vectors , for the two cases , are : for @xmath428 , the columns of the @xmath429 matrix @xmath430 , where @xmath431 for @xmath390 , the columns of the @xmath432 matrix @xmath433 , where @xmath434 note that both of these sets of columns are easily seen to be linearly independent .
hence , in each case , for the adjacency matrix @xmath2 of these signed graphs , @xmath80 is non - singular , so @xmath79 is not an eigenvalue .
since these families comprise bipartite graphs ( in the extended sense ) , @xmath5 is not an eigenvalue .
the families are cyclotomic , being subgraphs of @xmath435 for some @xmath49 , so we are done .
now we find the remaining graphs . for subgraphs of the sporadic graphs
@xmath32 and @xmath34 , we know that these are subgraphs of @xmath90 , and so can be embedded in @xmath436 , with @xmath80 nonsingular .
hence such a subgraph can have at most @xmath105 vertices .
these can be found by exhaustive search ; the maximal ones are @xmath387 and @xmath437 .
there remain the subgraphs of the infinite families .
we observe that : * an hour - glass , equivalent to an unsigned square , has @xmath5 as an eigenvalue ; * the classical @xmath249 graphs ( see figure [ f - except ] ) have @xmath5 as an eigenvalue .
hence the subgraph can contain at most one pair of conjugate vertices .
so it is either a path , a cycle , some @xmath390 or @xmath438 , defined to be @xmath439 with its two leaves identified .
a path is a subgraph of some @xmath390 , while a cycle must be equivalent to some @xmath428 , for otherwise it is equivalent to a cycle with all positive edges for which @xmath5 is an eigenvalue . for @xmath438 , we can delete one of its pair of conjugate vertices to obtain a graph equivalent to a cycle with all positive edges .
this completes the proof of theorem [ t : open_signed ] .
the proof of theorem [ t : open_charged_signed ] is similar .
we can assume that the charged graphs we seek do indeed have at least one charged vertex .
the relevant subgraphs of @xmath260 , @xmath261 and @xmath440 are found by exhaustive search . for the subgraphs of @xmath297 and @xmath298
, we see by the same argument as above that the neutral component can contain at most one pair of conjugate vertices .
hence the neutral component is a path , or some @xmath390 .
two adjacent charges of the same sign have one of @xmath44 as an eigenvalue , so each charged component has exactly one charge .
putting charges of the same sign at each end of a path would give one of @xmath44 as an eigenvalue , as one can see by writing down an obvious eigenvector . putting a charge on either end of some @xmath390
gives one of @xmath441 as an eigenvalue . to see this it suffices to consider adding a negative charge to one end , with corresponding column vector @xmath442 to add to @xmath433 , and adding a row of zeroes to @xmath433 to make it square , giving a singular matrix , and hence @xmath79 as an eigenvalue
this leaves @xmath443 ( and its subgraphs ) as the only possibilities . for @xmath443 ,
the columns of the @xmath3 matrix @xmath444 , where @xmath445 are easily seen to be linearly independent .
again , for the adjacency matrix @xmath2 of this bipartite charged signed graph , @xmath80 is non - singular , so @xmath79 is not an eigenvalue , and hence neither is @xmath5 .
this completes the proof of theorem [ t : open_charged_signed ] .
theorem [ t : open_charged ] then follows easily .
table [ ta-1 ] gives the reciprocal polynomials of the maximal connected cyclotomic charged signed graphs that appear in our results .
all are maximal in the sense explained where they appear , apart from the @xmath390 which , as we have seen , do not belong to any connected cyclotomic charged signed graph maximal with respect to having all eigenvalues in @xmath1 .
note , however , that the polynomials associated to @xmath446 and @xmath260 will need changes of variable @xmath447 , @xmath448 when going from one equivalent , but not strongly equivalent , graph to another .
c | c | c charged signed graph & characteristic polynomial & associated cyclotomic polynomial + @xmath30 & @xmath449 & @xmath450 + @xmath32 & @xmath451 & @xmath452 + @xmath34 & @xmath453 & @xmath454 + @xmath446 & @xmath455 & @xmath456 + @xmath457 & @xmath449 & @xmath458 + @xmath260 & @xmath459 & @xmath460 + @xmath261 , @xmath262 & @xmath461 & @xmath462 + table [ ta-2 ] gives the reciprocal polynomials of the cyclotomic signed graphs of theorems [ t : open_signed ] and [ t : open_charged_signed ] , shown in figures [ f - u ] , [ f - q ] , [ f - v ] , [ f - o ] and [ f - p ] . in the table , @xmath463 denotes the @xmath49th cyclotomic polynomial .
c | c charged signed graph & associated cyclotomic polynomial + @xmath428 & @xmath464 + @xmath390 & @xmath465 + @xmath393 & @xmath466 + @xmath467 & @xmath468 + @xmath469 & @xmath470 + @xmath471 & @xmath472 + @xmath258 & @xmath473 + @xmath474 , @xmath475 & @xmath476 + @xmath477 , @xmath478 & @xmath479 + @xmath480 & @xmath481 + @xmath482 & @xmath483 + @xmath402 & @xmath484 + @xmath403 , @xmath485 & @xmath486 + @xmath487 , @xmath488 & @xmath489 + @xmath490 , @xmath491 & @xmath492 + @xmath493 , @xmath494 & @xmath495 + @xmath496 & @xmath497 + @xmath498 & @xmath499 + @xmath500 & @xmath501 + for a single sporadic graph with adjacency matrix @xmath2 , the reciprocal polynomial @xmath8 can be easily calculated .
for the infinite families , more work is required . here , for convenience
, we use the same notation for a graph and its associated cyclotomic polynomial . for computing formulae for families of associated cyclotomic polynomials , a standard tool will be to use induction on the determinant @xmath502 , where @xmath2 is the adjacency matrix of the graph under consideration . in this way it is easy first to compute the the @xmath49-vertex ( unsigned ) path @xmath496 ,
giving @xmath503 , as in the table ( see also @xcite ) .
then expansion by the first row of the determinant gives @xmath504 .
also @xmath443 is readily calculated , again expanding in the same way . for @xmath428 ,
determinant expansion firstly along the top row , and then down the left rows of the resulting determinants gives @xmath505 , and hence the result . for @xmath390 , the formulae for @xmath439 and @xmath506 can be proved by induction , using the determinant , in a similar way to that for @xmath507 .
these can then be used as the base cases for an inductive proof of the @xmath390 formula . for @xmath30 ,
label its top vertices @xmath508 and the bottom vertices @xmath509 , with @xmath5 the conjugate vertex to vertex @xmath13 .
then @xmath510 is an eigenvector of @xmath30 with eigenvalue @xmath79 , and @xmath511 is an eigenvector of @xmath30 with eigenvalue @xmath5 , both associated to the hourglass @xmath512 $ ] . from the symmetry of @xmath30 that acts by @xmath513 on its vertices , we get two eigenvectors , with eigenvalues @xmath79 and @xmath5 for each of the hourglasses @xmath514 , [ 5,6,7,8],\dots,[2k-1,2k,1,2]$ ] .
these eigenvectors are independent , so that @xmath30 has characteristic polynomial @xmath515 , which , on putting @xmath516 , gives the result .
for @xmath297 , label the vertices as for @xmath30 .
for the hourglasses @xmath514 $ ] , @xmath517 $ ] ( those without charged vertices ) , we get the same eigenvectors as for @xmath30 , with the same eigenvalues .
the hourglasses @xmath512 $ ] and @xmath518 $ ] give the same eigenvectors as for @xmath30 with eigenvalue @xmath79 .
for the hourglass @xmath512 $ ] , however , we also get two independent eigenvectors @xmath519 and @xmath520 with eigenvalue @xmath5 , and from the hourglass @xmath518 $ ] we get two more independent eigenvectors @xmath521 and @xmath522 with eigenvalue @xmath5 . thus @xmath297 has characteristic polynomial @xmath455 , giving the result . for @xmath298 , note that this is bipartite in the extended sense , so that the eigenvalues @xmath5 and @xmath79 have equal multiplicities .
given the root system @xmath523 of a finite reflection group , one classically looks for a subset @xmath524 that is a _ simple system _ , namely one that is a basis for the @xmath525-span of @xmath523 and such that every element of @xmath523 is a linear combination of elements of @xmath524 with all coefficients weakly of the same sign .
the coxeter graph of a simple system is determined by the reflection group , and provides a means of classifying finite reflection groups .
if we have a signed graph with all eigenvalues in @xmath1 then ( as we have seen ) its vertices can be associated with a linearly independent set @xmath526 of vectors , and the reflection group generated by the hyperplanes orthogonal to those vectors is a finite reflection group .
the closure of @xmath526 under this reflection group is a root system @xmath523 : in the language of @xcite we are taking the star closure of the lines spanned by the elements of @xmath526 .
our set @xmath526 will not generally be a simple system for @xmath523 , but it will be a basis for the @xmath525-span of @xmath523 .
the unsigned version of our graph ( making all edges positive ) is the coxeter graph of @xmath526 .
the neutral signed graphs of theorem [ t : open_signed ] therefore provide a classification of all coxeter graphs coming from bases for the @xmath525-span of root systems contained in either @xmath88 @xmath527 or @xmath90 .
for example , one can generate @xmath90 using eight reflections whose coxeter graph is the cube @xmath393 of figure [ f - u ] .
robin chapman has pointed out that , up to equivalence , the signed graph @xmath32 of figure [ f - csg_14 ] can be defined as follows : label the vertices @xmath529 and , working modulo @xmath530 , for each @xmath14 join @xmath14 to each of @xmath531 , @xmath532 and @xmath533 by positive edges , and join @xmath14 to @xmath534 by a negative edge .
the representation of @xmath32 in the figure is based on this observation .
kronecker , leopold . :
zwei s " atse " uber gleichungen mit ganzzahligen coefficienten , j. reine angew . math . * 53 * ( 1857 ) , 173175 . see also werke .
vol . 1 , 103108 , chelsea publishing co. , new york , 1968
. vijayakumar , g. r. ; singhi , n. m. some recent results on signed graphs with least eigenvalues @xmath538 .
coding theory and design theory , part i , 213218 , i m a vol .
appl . , 20 , springer , new york , 1990 . smith john h. some properties of the spectrum of a graph . 1970 combinatorial structures and their applications ( proc .
calgary internat .
conf . , calgary , alta . , 1969 ) ,
pp . 403406 , gordon and breach , new york .
vijayakumar , g. r. ; singhi , n. m. some recent results on signed graphs with least eigenvalues @xmath538 .
coding theory and design theory , part i , 213218 , i m a vol .
appl . , 20 , springer , new york , 1990 . | we completely describe all integer symmetric matrices that have all their eigenvalues in the interval @xmath0 $ ] . along the way we classify all signed graphs , and then all charged signed graphs , having all their eigenvalues in this same interval .
we then classify subsets of the above for which the integer symmetric matrices , signed graphs and charged signed graphs have all their eigenvalues in the open interval @xmath1 . |
consider a system of @xmath0 charges in a spherical container of radius @xmath1 .
an external potential centered at the origin exerts an attractive central force .
the hamiltonian is @xmath2where @xmath3 and @xmath4 are the position and velocity of charge @xmath5 .
the repulsive interaction potential between particles @xmath5 and @xmath6 is denoted by @xmath7 where @xmath8 . the external confinement " potential @xmath9 is the same for all particles , and @xmath10 is the wall potential that is zero inside the container and infinite otherwise .
charged systems are of direct relevance e.g. to dusty plasmas and ions in traps . for neutral systems
@xmath11 is the interaction of each particle with a uniform neutralizing background , corresponding to an ocp or jellium with an attractive trap at the origin .
an example that has been studied recently is jellium with a point positive ion of charge number @xmath12 at the origin @xcite . at equilibrium the electron density
is enhanced ( partial confinement ) near the origin and approaches a uniform limit far from the ion , for sufficiently large @xmath1 , due to charge neutrality .
an opposite extreme is a system of charges in a strong trap such that they are localized in a finite domain away from the wall and their density vanishes outside this domain ( complete confinement ) .
there is a continuous mapping between these two limiting cases , controlled by the relative strengths of the repulsion between particles and the strength and form of the confining potential .
the discussion of these cases here will be limited to classical statistical mechanics .
in the absence of a confining potential the equilibrium density may be uniform ( charge neutral ) or nonuniform ( charged ) . in general the confining potential will induce a nonuniform equilibrium density in any case .
formally it is determined from the exact yvon - born - green equation @xcite@xmath13where @xmath14 is the joint density for two charges and @xmath15 is the density of the uniform neutralizing background .
a reasonable approximation is the hypernetted chain approximation ( hnc ) @xcite,@xmath16 = 0 .
\label{3.3}\ ] ] the solutions to this equation are quite rich , reflecting the competition between the attraction of @xmath17 and the renormalized repulsion represented by @xmath18 .
they have been described in some detail for the case of the positive ion in jellium @xcite . here
we report on some preliminary results for the qualitatively different case of weakly coupled charges @xmath19 in a harmonic trap @xmath20 with @xmath21 .
there is then a competition between coulomb repulsion and attraction by the central trap , with charge density enhanced near the wall or near the center , respectively .
these effects are exactly balanced when @xmath22 , leading to a uniform density at all temperatures . here
@xmath23 is the mie plasma frequency ( @xmath24 ) and @xmath25 is the frequency associated with the harmonic trap . more generally , when @xmath26 the charges are increasingly drawn away from the wall as the temperature is lowered , approaching at @xmath27 a uniform density for @xmath28 and zero density for @xmath29 @xcite . in the opposite case of @xmath30
the coulomb repulsion dominates and the particle density is enhanced at the wall . as the temperature is lowered the density becomes sharp at the walls - a coulomb explosion " that is restrained by the external walls .
this behavior is illustrated in fig .
1a ) for @xmath31 and in fig .
1b ) for @xmath32 , at @xmath33 , @xmath34 , and @xmath35 .
consider a perturbation of the nonuniform equilibrium state by an external potential of the form @xmath36 , where @xmath37 is the phase function representing the particle density . then
, to linear order in @xmath38 the response of the average particle density to this perturbation is @xcite @xmath39with the response function@xmath40 the brackets @xmath41 denote an equilibrium ensemble average .
the second equality follows from the conservation law @xmath42 , and the stationarity of the equilibrium ensemble .
the response function @xmath43 provides an appropriate object for studying the possible modes of excitation that are supported by the system .
an alternative equivalent instrument is the dielectric function @xmath44 defined in analogy with electrodynamics@xmath45
the evaluation of the response function is a difficult many - body problem .
the conditions of interest include both strong confinement and possibly strong charge correlations , so there is no small parameter available for simplifications . instead
, a non perturbative markov kinetic theory has been discussed recently in this context @xcite .
it is based on approximating the formal generator for dynamics in the single particle phase space by its exact form at @xmath46 .
this leads to a mean field theory of the linear vlasov form , but with both the confining potential and the charge - charge potential renormalized by the initial equilibrium correlations .
the analysis of reference @xcite can be extended in a straightforward way to the system considered here with the result@xmath47where @xmath48 is the response function for confined particles without the interparticle interactions ( @xmath49 ) @xmath50@xmath51 is the maxwellian , and where the generator @xmath52 for the effective single particle dynamics is@xmath53the renormalized confinement potential is determined from the equilibrium density by @xmath54 . also , @xmath55 is @xmath56where the renormalized charge - charge potential , @xmath57 , is defined in terms of the direct correlation function by @xmath58 .
further progress requires evaluation of @xmath59 whose dynamics is determined from the effective single particle charge dynamics in the presence of the confinement potential
. this can be quite complex over the whole range of weak to strong confinement , and has been discussed in some detail for the case of a central positive ion in jellium @xcite .
the autocorrelation function for the total force on the charges was studied as a function of the charge number on the central ion .
it was found that a simple representation of the single particle trajectories in terms of bound and free states provided an accurate analytical description from weak to strong ion - electron interaction .
preliminary studies of the harmonic trap suggest the possibility of a simplification in that case as well .
as noted above , the low temperature limit leads to a uniform density inside a sphere of radius @xmath60 . since the effective potential @xmath61 for the dynamics
is determined from this potential , @xmath62 becomes the correlation function for free particles inside a sphere .
an exact evaluation of the force autocorrelation function in terms of all response frequencies is possible in this case @xcite .
this research was supported by the nsf / doe partnership in basic plasma science and engineering under the department of energy award de - fg02 - 07er54946 .
9 talin b , calisti a , dufty j and pogorelov i 2008 phys .
e * 77 * , 036410 ; dufty j , pogorelov i , talin b and calisti a 2003 , j. phys .
a * 36 * , 6057 ; dufty j , talin b and calisti a 2004 , in _ theory of energy deposition _ , adv .
quant . chem . *
46 * , 293 . | the dynamics of fluctuations is considered for electrons near a positive ion or for charges in a confining trap .
the stationary nonuniform equilibrium densities are discussed and contrasted . the linear response function for small perturbations of this nonuniform state is calculated from a linear markov kinetic theory whose generator for the dynamics is exact in the short time limit . the kinetic equation is solved in terms of an effective mean field single particle dynamics determined by the local density and dynamical screening by a dielectric function for the non - uniform system . the autocorrelation function for the total force on the charges is discussed . |
hexagonal ymno@xmath0 is one of the most well - known examples of multiferroics , a material that possesses both ferroelectricity and magnetism .
the ferroelectric ( fe ) transition occurs at a relatively high temperature above 900 k. below the curie temperature the compound crystallizes in a hexagonal structure with the space group @xmath6 @xcite .
the mn@xmath7 ( @xmath8=2 ) ions form nearly triangular networks in @xmath9=0 and @xmath9=1/2 layers which stack in the abab sequence along the @xmath9 axis with a wide separation introduced by the intervening y and o ions @xcite .
this suggests a predominant two - dimensional ( 2d ) character in the @xmath10 plane . for this system the geometrical frustration of
the antiferromagnetic ( af ) spins on the triangular lattice results in the @xmath11 spin ordering below the nel temperature @xmath1 @xmath12 70 k. the curie - weiss ( cw ) temperature deduced from susceptibility measurements is approximately ten times higher than @xmath1 @xcite and the magnetic moment is lower than the expected value from the fully polarized mn@xmath7 @xcite .
this reduction in @xmath1 and magnetic moment is consistent with strong spin fluctuations due to geometrical frustration and/or low dimensionality .
the coexistence of novel magnetism and ferroelectricity suggests a peculiar interaction between the spins and the lattice degrees of freedom , although the low @xmath1 , compared to the ferroelectric transition , infers weak coupling between magnetism and ferroelectricity . although various crystallographic , optical and magnetic structural studies have been performed to characterize independently the fe and magnetic properties @xcite , only a few ones have been directed to understand their interrelation .
dielectric constant @xcite , specific heat @xcite and thermal conductivity @xcite experiments revealed a spin - lattice coupling whose understanding remains an open question .
as far as magnetism is concerned , neutron scattering studies @xcite have reported the presence of unconventional spin fluctuations both above and below @xmath1 that could be related to the anomalous increase of the thermal conductivity upon magnetic ordering @xcite .
however , the question of coexistence of long - range order with significant spin - fluctuations in the nel state for the triangular lattice remains .
the symmetry of the order parameter and the nature of the universality class associated with the @xmath11 spin structure are also not resolved up to now . to investigate the coupling between magnetism and the lattice degrees of freedom in ymno@xmath0 , we report an ultrasonic investigation of the elastic moduli of a single crystal in the temperature range 2 - 150 k , both above and below the spin ordering temperature @xmath1 .
although elastic anomalies at @xmath1 are observed on all studied moduli , an important coupling of the magnetic order parameter is only obtained with an in - plane strain field .
a landau free energy expansion is used to interpret these anomalies and a critical analysis of the temperature behavior just below @xmath1 is tentatively given .
small temperature hysteretic effects are observed indicating a domain structure .
a peculiar temperature profile above @xmath1 for two moduli is discussed in relation with the magnetic fluctuation issue .
high quality ymno@xmath0 single crystals are grown using the standard floating zone technique .
powder x - ray diffraction measurements confirmed that the crystal investigated here has the @xmath6 hexagonal structure at room temperature ( @xmath13=@xmath14=6.1380(1) , @xmath15=11.4045(1) ) .
a single crystal was then oriented with a back laue x - ray diffraction technique and cut with a wire - saw to obtain two sets of parallel faces respectively oriented perpendicular ( separation distance 3.29 mm ) and parallel ( 3.54 mm ) to the hexagonal @xmath15 axis .
the faces were subsequently polished to obtain a mirror like aspect .
we used a pulsed - echo ultrasonic technique in the reflection mode to measure the velocity and amplitude of longitudinal and transverses waves propagating along both directions .
the plane acoustic waves were generated with linbo@xmath0 piezoelectric transducers ( fundamental frequency 30 mhz and odd overtones ) bonded to one parallel face with silicon seal .
the technique consists in measuring the phase shift and the amplitude of the first elastic pulse reflected from the opposite parallel face .
doing so , this technique yields only velocity and attenuation variations ; an absolute value of the velocity is obtained by measuring the transit time between different reflected echoes .
the crystal was mounted in a variable temperature insert ( vti ) ; the useful temperature range was 2 - 150 k , the maximum value being imposed by phase transformation of the silicon seal around 200 k. a magnetic field up to 14 tesla could be applied along different directions . for a crystal having hexagonal symmetry
the elastic modulus matrix has only five non - zero independent components @xmath16 @xcite .
four of these components can be obtained by propagating ultrasonic waves parallel and perpendicular to the hexagonal @xmath15 axis and by measuring the velocity of these waves with various polarizations ( longitudinal and transverse ) . for propagation along @xmath15 ,
the longitudinal and transverse velocities are respectively given by @xmath17 and @xmath18 ; for the perpendicular direction , besides the longitudinal one , two transverse velocities ( polarization @xmath19 and @xmath20 ) are obtained : @xmath21 , @xmath22 and @xmath23 . by using the density @xmath24 g/@xmath25 and the experimental velocities ,
the four @xmath16 are obtained .
we notice that @xmath26 and @xmath27 are equal if the crystal is correctly oriented , a situation that was checked in our experiment .
the absolute values of the moduli at low temperatures are given in table i. .[table1]elastic moduli of ymno@xmath0 ( units @xmath28 ) at 4k . [ cols="^,^,^,^",options="header " , ]
we have then tried to analyze the variation of the elastic moduli @xmath3 , @xmath29 and @xmath4 relative to their value at @xmath1 by fitting the order parameter to @xmath30 near the critical temperature @xmath1 .
@xmath31 where @xmath32 is a constant depending on @xmath33 .
this analysis for @xmath3 and @xmath4 is surely more valid than for @xmath29 since the anomalies are at least 10 times larger at @xmath1 .
the results of this analysis are presented in figure 4 on a log - log scale .
not only the overall temperature dependence appears very similar for @xmath3 and @xmath4 , but the same exponent @xmath34 is obtained with @xmath1 = 72.43(5 ) .
surprisingly the same exponent is also found for @xmath29 , what means that the overall temperature profile of the background near @xmath1 is not important to determine the critical behavior .
also hysteresis effects were found to only affect @xmath1 ( fig.3 ) , not the exponent .
the value @xmath35 , which is obtained over only two decades of the reduced temperature , is neither consistent with chiral xy ( 0.25 ) and chiral heisenberg ( 0.30 ) universality classes or with xy ( 0.35 ) and heisenberg ( 0.36 ) ones @xcite .
although this appears to contradict our previous observation pertaining to the coupling constants , it is clear from fig.4 that the same physical phenomenon is at the origin of the elastic anomalies on @xmath3 and @xmath4 , and by extension on the other two moduli .
the af magnetic order parameter @xmath36 has been associated with 3d heisenberg or 3d xy symmetry classes from specific heat @xcite and muon - spin relaxation @xcite studies ; our study is however not able to resolve this issue .
finally , the sharp downward step observed on @xmath3 ( fig.3a ) could be consistent with the prediction of eq.(6 ) .
on the one hand , the observation of hysteresis in the vicinity of this peak could be consistent with weak ferromagnetism , although this has been observed only in scmno@xmath0 @xcite .
on the other hand , coupling between magnetic and electric domains in ymno@xmath0 has been reported by imaging with optical second harmonic generation @xcite ; it has been suggested that the small anomaly observed on the in - plane dielectric constant @xcite could result from such coupling effects on domain walls @xcite .
the small peak on @xmath3 at @xmath1 together with the observation of hysteresis could result from an interaction between in - plane compression waves and the domain walls .
variation of the attenuation @xmath37 of in - plane compression waves a function of temperature at two frequencies , 103 and 168 mhz.,width=321 ] previous neutron scattering studies @xcite reported the presence of unconventional spin fluctuations both above and below @xmath1 .
these fluctuations could explain the anomalous behavior of @xmath3 ( and by extension @xmath4 ) well above @xmath1 as shown in figure 3a .
since similar effects are not observed for @xmath38 and @xmath29 , fluctuations preferentially 2d - like associated to in - plane exchange constants are likely responsible .
the ultrasonic attenuation is generally more sensitive to fluctuations since it is a dissipative phenomenon .
these attenuation variations @xmath39 are difficult to obtain with precision on a wide temperature range and this is why they must be interpret with great care .
we present in figure 5 the variation of the attenuation associated to the @xmath3 modulus as a function of temperature at two frequencies .
as the temperature is decreased from high temperature , the attenuation generally decreases because the scattering of the acoustic waves by phonons is reduced . here
the attenuation shows a minimum at a rather high temperature near 110 k for both frequencies and it increases for lower temperatures , a behavior that is not consistent with decreasing phonon scattering . at 103 mhz ,
the af magnetic ordering produces a very sharp peak at @xmath1 but does not change significantly the monotonous increase of the attenuation that begun near 110 k and the attenuation increases down to the lowest temperature . at 168 mhz , the temperature profile from 150 k to 72 k is conserved ; however , the peak is more important at @xmath1 and the attenuation decreases rapidly below the nel temperature . at lower temperature , there is no saturation of the attenuation but a small increase below 30 k is observed .
these attenuation data strongly suggest that spin degrees of freedom , in other words magnetic fluctuations both above ( at least up to 110 k ) and below @xmath1 , dissipate acoustic energy . at @xmath1 ,
the sharp increase is due to the spontaneous strain yielding to eq.(6 ) and for which hysteresis is observed . below @xmath1 ,
spin degrees of freedom are lost when 3d af ordering occur and the attenuation decreases rapidly .
the data reveal an increase of attenuation at temperatures much higher than @xmath1 and a reduction below when 3d af magnetic ordering occurs .
although this reduction increases rapidly in amplitude with frequency , dissipation still occurs well below @xmath1 .
these observations are consistent with _ extraordinary spin - phonon interactions _ deduced from thermal conductivity data @xcite .
how these fluctuations are related to the domain structure remains an open question .
the presence of domains and fluctuations near @xmath1 could complicate the analysis of the critical behavior presented in fig.4 and affect the determination of the exponent @xmath5 .
relative variation of @xmath3 as a function of the magnetic field at two temperatures well below and above @xmath1.,width=321 ] the magnetic character of the peculiar behavior of @xmath3 above @xmath1 can also be revealed by a magnetic field analysis . up to 14 tesla
we did not find any measurable variation of @xmath1 whatever the orientation of the field relative to the @xmath15 axis was .
this probably implies that the amplitude of the magnetic order parameter below @xmath1 is highly rigid relative to the magnetic field .
nevertheless some magnetic field effects can be measured on @xmath3 over the 2 - 150 k temperature range .
these are presented in figure 6 as @xmath3 variations relative to the value at zero field .
the data are presented as a function of @xmath40 at two temperatures , 10 k well below @xmath1 and 100 k well above for a field applied parallel to the @xmath15 axis .
the field profile is very similar for both temperatures and the absolute value has not even decreased by a factor 2 when the temperature has been multiplied by ten .
these effects have thus to be related to the overall background : a magnetic field enhances the softening of @xmath3 both above and below @xmath1 confirming the existence of important magnetic fluctuations in this geometrically frustrated magnet , fluctuations that are likely enhanced by a magnetic field .
it has been suggested from heat capacity data @xcite that a single einstein contribution could adequately fit the low - temperature deviation , indicating that there is no evidence of a anomalous magnetic contribution for @xmath41 .
our data do not support this suggestion .
in conclusion , our ultrasonic experiments have identified an important coupling between the af magnetic order parameter and the lattice in hexagonal ymno@xmath0 . this coupling yield important stiffening anomalies below @xmath1 only for in - plane elastic deformations , compression and shear .
this implies that the coupling is established via the modulation of the in - plane exchange interactions .
the critical exponent @xmath5 for the af order parameter near @xmath1 is found higher than the expected value from xy and 3d heisenberg universality classes .
this situation favors a conventional antiferromagnetic long range ordering as suggested from other measurements .
however , the data confirm important spin fluctuations over a wide temperature range above and below the nel temperature .
finally , a sharp step appearing on @xmath3 at @xmath1 , for which hysteresis is observed , could be consistent with a weak coupling between ferroelectricity and magnetism at domain walls .
yakel , w.c.koehler , e.f .
bertaut and e.f .
forrat , acta crystallogr .
* 16 * , 957 ( 1963 ) .
t. katsufuji , m. masaki , a. machida , m. morimoto , k. kato , e. nishibori , m. takata , m. sakata , k. ohoyama , k. kitazawa and h. takagi , phys . rev .
b * 66 * , 134434 ( 2002 ) .
a. munoz , j.a .
alonzo , m.j .
martinez - lope , m.t .
casais , j.l . martinez and m.t .
fernandez - diaz , phys .
b. * 62 * , 9498 ( 2000 ) .
t. katsufuji , s. mori , m. masaki , y. morimoto , n. yamamoto and h. takagi , phys .
b * 64 * , 104419 ( 2001 ) .
m. bieringer and j.e .
greedan , solid state chem . * 143 * , 132 ( 1999 ) .
m. n. iliev , h .-
lee , v. n. popov , m. v. abrashev , a. hamed , r. l. meng , and c. w. chu , phys .
b * 56 * , 2488 ( 1997 ) .
m. tachibana , j. yamazaki , h. kawaji and t. atake , phys .
b * 72 * , 064434 ( 2005 ) .
tomuta , s. ramakrishnan , g.j .
nieuwenhuys and j.d .
mydosh , j. phys .
matter * 13 * , 4543 ( 2001 ) .
sharma , j.s .
ahn , n. hur , s. park , sung baek kim , seongsu lee , j .-
park , s. guha and s .- w .
cheong , phys .
lett . * 93 * , 177202 ( 2004 ) .
zhou , j.b .
goodenough , j.m .
gallardo - amores , e. moran , m.a .
alario - franco and r. caudillo , phys .
b * 74 * , 014422 ( 2006 ) .
lonkai , d. hohlwein , j. ihringer and w. prandl , appl .
phys . a * 74 * s843 ( 2002 ) .
sato , s .- h .
lee , t. katsufuji , , m. masaki , s. park , j.r.d .
copley and h. tanagi , phys .
b * 68 * , 014432 ( 2003 ) .
j. park , j .-
park , g.s .
jeon , y .- y .
choi , c. lee , w. jo , r. bewley , k.a .
mcewen and t.g .
perring , phys .
b * 68 * , 104426 ( 2003 ) . b. roessli , s.n .
gvasaliya , e. pomjakushina and k. conder , jetp letters ( in press ) , cond - mat0503209 .
e. dieulesaint and d. royer , ondes lastiques dans les solides , masson et cie , diteurs , paris 1974 , p.121 .
collins and o.a .
petrenko , can .
75 * , 605 ( 1997 ) .
y. trudeau , m.l .
plumer , m .poirier and a. caill , phys . rev . * 48 * , 12805 ( 1993 ) .
g. quirion , t. taylor and m. poirier , phys . rev .
* 72 * , 094403 ( 2005 ) .
g. quirion , x. han , m.l . plumer and m. poirier ,
lett . * 97 * , 077202 ( 2006 ) .
w. rewald , adv . phys . * 22 * , 721 ( 1973 ) .
t. lancaster , s.j .
blundell , d. andreica , m. janoschek , b. roessli , s.n .
gvasaliya , k. conder , e. pomjakushina , m.l .
brooks , p.j .
baker , d. prabhakaran , w. hayes and f.l .
pratt , cond - mat/0610467v2 .
m. fiebig , th .
lottermoser , d. frlich , a.v .
gottsev and r.v .
pisarev , nature * 419 * , 818 ( 2002 ) .
huang , y. cao , y.y .
sun , y.y . sue and c.w .
chu , phys .
b * 56 * , 2623 ( 1997 ) . j.h .
qiu an q. jiang , phys .
b * 73 * , 024406 ( 2006 ) . | we report an ultrasonic investigation of the elastic moduli on a single crystal of hexagonal ymno@xmath0 as a function of temperature .
stiffening anomalies in the antiferromagnetic nel state below @xmath1 = 72.4 k are observed on all the four elastic moduli @xmath2 .
the anomalies are the most important on @xmath3 and @xmath4 for in - plane elastic deformations ; this is consistent with a strong coupling of the lattice with the in - plane exchange interactions .
we use a landau free energy model to account for these elastic anomalies .
we derive an expression which relates the temperature profile of the anomaly to the order parameter ; the critical exponent associated to this parameter @xmath5 = 0.42 is not consistent with a chiral xy or 3d heisenberg universality class , but more in agreement with a conventional antiferromagnetic long range order . a tiny softening anomaly on @xmath3 for which hysteresis effects are observed
could be indicative of an interaction between ferroelectric and magnetic domains at @xmath1 .
moreover , magnetic fluctuations effects both above and below @xmath1 are identified through abnormal temperature and magnetic field effects . |
the spinless falicov - kimball model has two types of particles : spinless electrons and classical particles , which we refer to as ions .
the particles are on a lattice with the restriction that there is at most one ion at each lattice site .
the spinless electrons are fermions , so there is at most one electron per site .
the electrons can hop between nearest neighbor sites , but the ions can not .
there is an on - site interaction between electrons and ions .
the hamiltonian is @xmath5 where @xmath6 and @xmath7 are creation and annihilation operators for the electrons .
@xmath8 is the occupation number for the ions , i.e. , @xmath9 if there is a ion at @xmath10 and @xmath11 if there is not .
the sum over @xmath12 is over nearest neighbor bonds in the lattice .
( the factor of 4 in front of the @xmath13 is included for latter convenience . )
this paper is only concerned with the square lattice , although the model may be defined on any lattice .
we will only consider the neutral model in which the number of electrons is equal to the number of ions , and the interaction parameter @xmath13 will be large and positive . by a hole - particle transformation results for positive @xmath13
imply results for negative @xmath13 , but we will not bother to state them . a review of rigorous work on the falicov - kimball model may be found in @xcite . here
we mention only some of the work on the neutral model for large positive @xmath13 .
in one dimension it is expected that for large @xmath13 the ground state of the neutral model with rational density is the periodic arrangement of the ions which is `` most homogeneous . ''
( there is an explicit algorithm for determining the most homogeneous configuration . )
this was proved by lemberger for @xmath14 where @xmath15 depends on the denominator of the rational density @xcite . in any number of dimensions the ground state for density
@xmath16 is the checkerboard configuration for all @xmath17 @xcite . in two dimensions with large @xmath13
the ground states for densities @xmath18 and @xmath1 are known rigorously and are periodic @xcite . for densities between @xmath0 and @xmath16
there are partial results on the ground state @xcite , but there is no proof it is periodic for rational densities .
based on what is known in one and two dimensions and the methods used to obtain these results , it is natural to conjecture that in two dimensions the ground state for large @xmath13 is periodic for rational densities . in this paper
we prove that this conjecture is wrong for densities between @xmath1 and @xmath0 other than @xmath2 .
in this density range there is phase separation in the ground state .
the phases involved are the ground states for densities @xmath1 , @xmath2 and @xmath0 , which are shown in figure 1 .
the ground state for densities between @xmath1 and @xmath2 is made up of large regions of density @xmath1 and density @xmath2 ground states with the relative areas chosen to yield the desired density .
a similar statement holds for densities between @xmath2 and @xmath0 .
the precise theorem is as follows .
1 [ thmmain ] there are positive constants @xmath19 and @xmath20 such that for @xmath21 the following is true for @xmath22 by @xmath22 squares @xmath23 with @xmath24 ions and @xmath24 electrons .
let @xmath25 be the common density .
if the density @xmath26 is @xmath2 , @xmath22 is a multiple of 6 and we use periodic boundary conditions , then the ground state configurations of the ions are configuration b in figure 1 and its translates .
if the density @xmath26 is between @xmath1 and @xmath2 , then for every ground state configuration of the ions we can find a subset @xmath27 of @xmath23 which contains at most @xmath28 sites and is such that on each connected component of @xmath29 the configuration agrees with either configuration a or b in figure 1 ( up to a lattice symmetry ) . if the density @xmath26 is between @xmath2 and @xmath0 , then the same statement is true with `` a or b '' replaced by `` b or d. '' ( the statements for densities other than @xmath2 are true for any choice of boundary conditions . ) although @xmath30 is large , the key point is that the bound on the number of sites in @xmath27 contains @xmath22 while the number of sites in @xmath23 is @xmath31 .
so for large @xmath22 , @xmath27 is a tiny fraction of the total area .
we can think of @xmath27 as consisting of domain walls between large regions which contain one of the configurations shown in figure @xmath32 .
the phase separation we find in this range of densities for the neutral model should not be confused with the phase separation discussed by freericks and falicov @xcite .
their argument applied to the non - neutral model .
the expectation that the neutral model with rational density should have periodic ground states is based on the following intuition .
the attraction between electrons and ions is large , so each electron spends most of its time at a site with an ion .
now consider the kinetic energy of an electron .
if nearby sites have ions , then the electrons at those sites will restrict the movement of the electron we are considering .
so its kinetic energy is minimized by spreading out the ions as much as possible to maximize the space that each electron has to move in .
however , watson @xcite emphasized that there will typically be a mismatch between the lattice and the natural ion configuration in the absence of a lattice .
thus the lattice structure can frustrate the exclusion principle s attempts to put the ions in the `` most homogeneous '' configuration .
we should emphasize that this paper only covers a small interval of densities .
an important open question is whether the phase separation we find here holds for most densities , or whether there are intervals in which the rational densities have periodic ground states which are the most homogeneous in some sense .
we conclude the introduction by sketching the proof for densities in @xmath3 .
when @xmath13 is large and the model is neutral , the ground state energy of a given ion configuration may be expanded in powers of @xmath33 .
this yields an effective hamiltonian for the ions .
one can begin to study it by only keeping terms up to a certain order in @xmath33 .
( of course , to prove anything one must eventually consider all orders . )
watson @xcite showed that when the density is between @xmath1 and @xmath0 the ground states of the fourth order hamiltonian correspond to tilings of the plane by squares and diamonds in which the squares and diamonds have the dimensions of those found in figure 1 .
watson s result will play a crucial role in our proof .
the vertices in such a tiling can be one of four types which we label a , b , c or d following watson s notation .
the four types are shown in figure 3 .
note that in figure 1 configurations a , b or d contain only vertices of type a , b or d , respectively .
all configurations that correspond to a square - diamond tiling have the same energy through fourth order .
to determine the ground states for densities in @xmath3 we must go to higher orders in the perturbation series . following watson s treatment of a similar model , we write the hamiltonian as a function of the number of each type of vertex . at sixth order
these square - diamond tilings still all have the same energy . at eighth order vertices of type a ,
b and c have the same energy but vertices of type d have higher energy . for densities between @xmath1 and @xmath2
there are square - diamond tilings which contain no vertices of type d. to determine the ground state among all these configurations we must go to tenth order . here
we find that the energy of a type c vertex is higher than that of a type b vertex .
thus for densities in @xmath3 the ground state must be a square - diamond tiling with only type a and b vertices .
( note that the tiling with only type a vertices has density 1/5 , while the tiling with only type b vertices has density 2/9 .
so one can obtain any density in @xmath3 by a suitable mix of type a and b vertices . )
however , a type a vertex can not be adjacent to a type b vertex . thus we must separate the a and b vertices to minimize the energy .
to derive the perturbation theory it is convenient to change to `` spin '' variables for the ions .
let @xmath34 , so @xmath35 when there is an ion at @xmath10 and @xmath36 when there is not .
we then take @xmath37 this differs from the original hamiltonian by a term proportional to @xmath38 , but we will only consider problems in which the number of electrons is fixed , so such a term is constant .
there are no interactions between the electrons , so the hamiltonian is just the second quantized form of the single electron hamiltonian @xmath39 . the operator @xmath40 has matrix elements @xmath41 with @xmath42 if @xmath43 and @xmath44 otherwise .
the operator @xmath45 is diagonal with entries @xmath46 .
the ground state energy for @xmath47 electrons is the sum of the @xmath47 lowest eigenvalues of @xmath48 . to find the ground state for a particular density of electrons and ions we must minimize this energy over all @xmath45 with the desired ion density .
let @xmath49 be the ground state energy for the ion configuration @xmath45 with the number of electrons equal to the number of ions .
to expand @xmath49 in powers of @xmath33 , we begin by rewriting @xmath49 as in @xcite . if @xmath50 then the number of negative eigenvalues of @xmath48 is equal to the number of sites with @xmath51 , i.e. , the number of ions .
thus when the number of electrons equals the number of ions , we have @xmath52\ ] ] where @xmath53 are the eigenvalues of @xmath48 .
now @xmath54 , and if we keep the number of ions fixed then @xmath55 is a constant .
so we might as well redefine @xmath56 .
then we write this as @xmath57^{{1 \over 2 } } ) = -u tr(1+\delta)^{{1 \over 2}}\ ] ] with @xmath58 we have used the fact that @xmath59 .
now we derive the perturbation theory by following the treatment by gruber , jedrzejewski and lemberger @xcite .
a somewhat different derivation may be found in @xcite .
if @xmath13 is sufficiently large , then @xmath60 and we may expand @xmath61 in a power series in @xmath62 . since @xmath41 is nonzero only if @xmath10 and @xmath63 are nearest neighbors , when we take the trace of each term we generate nearest neighbor walks that end where they start .
grouping together terms with the same power of @xmath64 , we may write the result as @xmath65 @xmath66 is summed over finite subsets of the lattice , and @xmath67 .
the coefficient @xmath68 is nonzero only if there is a nearest neighbor walk with @xmath69 steps which ends where it begins and visits each site in @xmath66 .
( it may visit sites outside of @xmath66 as well . )
the coefficients @xmath68 are invariant under the usual lattice symmetries .
there is a constant @xmath20 such that for every site @xmath10
@xmath70 we need to compute this effective hamiltonian through tenth order , i.e. , through the @xmath71 terms .
there are a lot of terms at this order .
since we will consider relatively low densities , it will prove useful to go back to the occupation variables @xmath8 .
( recall that @xmath72 . )
equation ( [ eqa ] ) gives @xmath73 where @xmath74 .
the coefficients @xmath75 may be computed from the @xmath68 in a straightforward manner . in particular , they are nonzero only if @xmath66 is contained in the set of sites visited by a @xmath69 step nearest neighbor walk that ends where it begins .
the @xmath75 satisfy a bound like ( [ eqb ] ) . for the proof it is useful to split the effective hamiltonian into three parts .
@xmath76 @xmath77 contains the terms in eq .
( [ eqaa ] ) for @xmath78 and @xmath79 .
@xmath80 contains the terms with @xmath81 and @xmath82 , and @xmath83 contains the terms with @xmath84 .
the fourth order hamiltonian @xmath77 has been computed before @xcite . in the occupation variables
it may be written in the following form . @xmath85
\\ & & + u^{-3 } [ 64 \sum_{<xy>:|x - y|=\sqrt{2 } } { v}_x { v}_y + 16 \sum_{<xy>:|x - y|=2 } { v}_x { v}_y - 16 \sum_{x } { v}_x + \sum_{x \in { e_4 } } c_{4,x } { v}_x ] \\\end{aligned}\ ] ] @xmath86 is the collections of sets @xmath66 which appear at fourth order and contain a pair of sites @xmath87 with @xmath43 .
there are only a few such sets and we can compute their coefficients explicitly , but as we will see their actual values play no role in the proof . if @xmath88 and @xmath89 , then there is a nearest neighbor pair @xmath87 with @xmath90 .
this gives a contribution of @xmath91 to @xmath77 which is much larger than the order @xmath92 contribution from @xmath93 .
it was shown in @xcite that for configurations with density between @xmath1 and @xmath0 which minimize @xmath77 every 3 by 3 block of sites must look like one of the four configurations shown in figure 2 .
( we will include a proof of this later . )
watson @xcite made the following observation that will play an essential role in our proof .
= 20 pt = 20 pt figure 2 : for densities between 1/5 and 1/4 , fourth order perturbation theory implies that in a ground state every 3 by 3 square must be one of the above cases , up to lattice symmetries .
[ lema ] if every 3 by 3 block equals one of those shown in figure 2 ( up to a lattice symmetry ) , then the configuration corresponds to a tiling of the lattice by squares and diamonds in which every vertex looks like one of the four types shown in figure 3 up to a lattice symmetry .
* proof @xcite : * note that the 3 by 3 blocks we are using can overlap .
thus the condition that every one of these blocks is one of those shown in figure 2 puts many constraints on the configuration .
we start with a site at which there is an ion . using the fact that every 3 by 3 block is one of those shown in figure 2 we work outwards from the initial ion and determine all possible configurations in the neighborhood of the inital ion .
we find that the configuration must look like one of the four cases shown in figure 3 .
width5pt it is important to note that the proof of the lemma is local .
if we have a configuration in which some 3 by 3 blocks are not one of those shown in figure 2 , then we can still conclude that in the parts of the lattice where the 3 by 3 blocks are one of those shown in the figure the configuration must be given by a square - diamond tiling .
we will say that an ion is type a , b , c or d if the configuration in the neighborhood of the ion agrees with that shown in figure 3 up to a lattice symmetry .
note that in regions where the configuration does not correspond to a square - diamond tiling there will be ions that are not any of these four types .
the tenth order hamiltonian @xmath80 contains too many terms to list here .
however , we can organize the hamiltonian so that we only need the actual values of the coefficients of a modest number of them .
the values of the coefficients of the terms in @xmath80 are important only in regions where @xmath77 is minimized . by the lemma the configuration corresponds to a square - diamond tiling in such regions . in these tilings
two ions can not be separated by a distance @xmath94 , @xmath95 or @xmath96 .
so we define @xmath97 to be the collection of sets that appear at tenth order or lower and contain a pair of sites @xmath87 with @xmath98 equal to @xmath99 or @xmath100 .
for @xmath101 the coefficients @xmath102 are given in table 1 .
[ thma ] let @xmath103 be the number of 3 by 3 blocks in which the configuration is not one of those shown in figure 2 .
let @xmath104 and @xmath105 be the number of type a , b , c and d ions respectively ( figure 3 ) . consider an @xmath22 by @xmath22 lattice so there are @xmath106 sites , and let @xmath26 be the ion density so there are @xmath107 ions .
then there are polynomials @xmath108 in @xmath64 and functions @xmath109 of @xmath64 such that @xmath110 @xmath111 @xmath112 @xmath113 @xmath114 @xmath115 denotes a quantity whose absolute value may be bounded by a constant times @xmath116 .
* proof : * the inequality for @xmath77 was proved in @xcite .
we include a short proof for completeness .
let @xmath117 be a 3 by 3 block of sites .
( so it contains 9 sites . )
let @xmath118 be the center of @xmath117 .
define @xmath119 \\\end{aligned}\ ] ] where @xmath120 is the number of translates of @xmath66 that are contained in @xmath117 .
now consider @xmath121 where the sum is over all 3 by 3 blocks .
( so some of them overlap . ) a pair @xmath122 with @xmath43 is contained in 6 different blocks , a pair with @xmath123 in 4 different blocks , and a pair with @xmath124 in 3 different blocks . using these facts
we find @xmath125 so to complete the proof we must show that @xmath126 vanishes if @xmath117 is one of the blocks shown in figure 2 and is at least @xmath127 otherwise . first note
that if the block contains a pair of nearest neighbor ions then the second order part of @xmath77 is at least @xmath128 .
since the fourth order part is smaller by a factor of @xmath129 , this shows that @xmath130 for all such configurations if @xmath13 is large enough .
now suppose that the block does not contain any nearest neighbor ions .
note that this implies that @xmath131 for all @xmath88 .
so we can easily compute @xmath126 for such configurations .
we find that for all 3 by 3 blocks that do not contain a nearest neighbor pair of ions and which are not in figure 2 , @xmath126 is at least @xmath127 .
@xmath126 vanishes on the configurations in figure 2 .
this completes the proof of ( [ inequalfour ] ) . to prove ( [ inequaltena ] ) and ( [ inequaltenb ] ) we first consider configurations which correspond to a square - diamond tiling .
if @xmath132 then @xmath131 in these configurations .
we will consider two @xmath66 s to be equivalent if they are related by a translation , reflection and or rotation .
for @xmath101 there are 16 equivalence classes , listed in table 1 . for each equivalence class
we want to write the number of @xmath66 in the class with @xmath133 in terms of @xmath134 .
consider the second equivalence class .
it contains those @xmath66 s of the form @xmath135 with @xmath124 .
the number of such @xmath66 with @xmath89 in figure 3 is 0,2,2 , or 4 for a , b , c , or d , respectively .
however , this overcounts the number of @xmath66 with @xmath89 .
each @xmath66 is counted 4 times .
so the number of @xmath66 s in the second equivalence class with @xmath133 is @xmath136 .
these coefficients @xmath137 along with the coefficients for all the other equivalence classes are given in table 1 .
we should emphasize that the overcounting factor is not always 4 .
it varies from equivalence class to equivalence class , and in one case within the equivalence class .
using table 1 we find that for configurations which correspond to a square - diamond tiling , @xmath138 the quantities @xmath104 and @xmath105 are not all independent .
each ion corresponds to a vertex in the square - diamond tiling , so @xmath139 by considering the areas associated with each of the four types of vertices in figure 3 , we see that @xmath140 using these two equations we can eliminate @xmath141 and @xmath142 from our expression for @xmath80 .
the result is ( [ inequaltena ] ) .
if we use the two equations to eliminate @xmath142 and @xmath105 , then the result is ( [ inequaltenb ] ) .
this proves ( [ inequaltena ] ) and ( [ inequaltenb ] ) for configurations that correspond to a square - diamond tiling . now consider a configuration that does not .
the number of sites in the region where it does not correspond to such a tiling is at most @xmath143 .
thus the terms in @xmath80 that intersect this region can contribute at most @xmath144 . outside of this region
we can apply the above argument .
there will be errors at the boundary of the region in which there is a square - diamond tiling since there will be incomplete squares and diamonds , but the contribution of these errors is also @xmath144 .
equations ( [ abcdn ] ) and ( [ abcdv ] ) are not true for a general configuration , but the difference between the right and left side of these equations is bounded by a constant times @xmath103 .
this completes the proof of ( [ inequaltena ] ) and ( [ inequaltenb ] ) .
finally , we must prove ( [ inequalinfinitya ] ) and ( [ inequalinfinityb ] ) .
we only give the proof of ( [ inequalinfinitya ] ) .
a similar argument proves ( [ inequalinfinityb ] ) , or it may be obtained from ( [ inequalinfinitya ] ) by showing @xmath145 is bounded by a constant times @xmath103 . to compare a term @xmath146 in @xmath83 with the corresponding term for the configurations in figure 1 , we need to know if @xmath66 is contained in a region where the configuration agrees with one of the configurations in figure 1 .
so we make the following definitions . for a lattice site @xmath10 and positive integer @xmath147 , let @xmath148 be the set of sites such that the @xmath149 distance from @xmath10 to @xmath63 is at most @xmath147 .
given a configuration , let @xmath150 be the set of sites with a type a ion such that @xmath151 is a subset of configuration a in figure 1 .
@xmath152 is defined analogously using configuration b. define @xmath153 so that @xmath154 since @xmath131 if @xmath11 , we can restrict the sum over @xmath10 to sites with an ion .
let @xmath155 denote the sites with an ion .
we start by comparing @xmath83 with the sum over a subset of the terms in @xmath83 : @xmath156 @xmath157 using ( [ eqb ] ) .
we need to estimate the size of @xmath158 .
let @xmath159 .
then @xmath151 is not a subset of configuration a or b in figure 1 .
so either ( i ) @xmath151 is not a subset of any square - diamond tiling , ( ii ) @xmath151 is a subset of a square - diamond tiling which contains a type c or type d vertex , or ( iii ) @xmath151 is a subset of a square - diamond tiling which contains both type a and type b vertices . in case ( i ) @xmath151 must intersect a 3 by 3 block that is not one of those in figure 2 .
note that it is not possible for a tiling to contain both type a and type b vertices , but no type c or d , so case ( iii ) never happens .
thus we can associate with each site in @xmath158 either a `` bad '' 3 by 3 block or a type c or type d vertex .
the number of sites that are associated with the same 3 by 3 block or type c or type d vertex is bounded by @xmath160 for some constant @xmath161 .
thus @xmath162 thus ( [ hamdif ] ) is @xmath163 let @xmath164 be the energy per site from @xmath83 when the entire finite lattice contains configuration a in figure 1 , and let @xmath165 be the same quantity for configuration b in figure 1 .
let @xmath166 if @xmath167 then the definition of @xmath150 implies that @xmath168 thus @xmath169 since @xmath170 , ( [ setbound ] ) implies @xmath171 .
this shows that ( [ adif ] ) is @xmath172 .
* proof of theorem [ thmmain ] : * first we prove the statement about the ground state for density @xmath2 .
configuration b has @xmath175 and only type b vertices , so theorem [ thma ] immediately implies it is a ground state .
any other ground state must have @xmath175 and no type c or d vertices .
the condition @xmath175 implies the configuration is a square diamond tiling . the only tiling which contains only a and b vertices and has density @xmath2 is configuration b in figure 1 .
now we turn to the proof of the statements for densities between @xmath0 and @xmath1 .
we will construct a trial configuration with relatively low energy and then use the inequalities in theorem [ thma ] to prove theorem [ thmmain ] .
we give the proof for the case of densities in @xmath3 .
the proof for densities in @xmath4 is similar .
divide the @xmath22 by @xmath22 square into two rectangles and put configuration a in figure 1 on one side and configuration b on the other side .
the relative areas of the two rectangles are chosen to give the desired density .
the number of 3 by 3 blocks which do not agree with one of those in figure 2 is bounded by a constant times @xmath22 .
thus @xmath176 is @xmath177 .
@xmath178 and @xmath105 are both zero , so @xmath179 is @xmath180 and @xmath181 is @xmath182 .
any ground state must have energy no greater than that of our trial configuration , so theorem [ thma ] implies that in a ground state @xmath183 which may be rearranged as @xmath184 { m}+ [ 224 u^{-7 } + 12240 u^{-9 } + { o}(u^{-11 } ) ] n_d \\ & & + [ 2340 u^{-9 } + { o}(u^{-11 } ) ] n_c \le { o}(u^{-1 } ) l \\\end{aligned}\ ] ] if @xmath13 is large enough this implies that @xmath185 is bounded by a constant times @xmath186 .
now take @xmath27 to be the union of all the 3 by 3 blocks which do not agree with one of the configurations in figure 2 together with the union over all type c and type d ions of the region associated with the ion shown in figure 3 . in @xmath29
the configuration must be a square diamond tiling with no c or d vertices .
note that type a and type b vertices can not be adjacent in the tiling , i.e. , separated by a distance @xmath187 .
however , this does not quite insure that each component contains only type a or type b vertices .
for example one can have a component that consists of two large regions which are connected only by a single line of sites .
then one can have type a vertices on one side of the narrow connection and type b on the other side . to eliminate this ,
redefine @xmath27 to be the original @xmath27 plus all sites within a distance @xmath161 of the original set .
if @xmath161 is chosen large enough we eliminate the above problem and every connected component of @xmath29 will be a square - diamond tiling with only type a or type b vertices .
width5pt * acknowledgements : * ideas from @xcite play a crucial role in this paper .
the author thanks prof .
watson for sending him a copy of this paper prior to its publication .
this work was supported in part by nsf grant dms-9623509 .
t. kennedy , some rigorous results on the ground states of the falicov - kimball model . _ _ rev . math .
* 6 * 901 - 925 ( 1994 ) .
also in _ the state of matter _ , michael aizenman and huzihiro araki ( eds . ) world scientific , 1994 . | the falicov - kimball model consists of spinless electrons and classical particles ( ions ) on a lattice .
the electrons hop between nearest neighbor sites while the ions do not .
we consider the model with equal numbers of ions and electrons and with a large on - site attractive force between ions and electrons . for densities @xmath0 and @xmath1 the ion configuration in the ground state had been proved to be periodic .
we prove that for density @xmath2 it is periodic as well .
however , for densities between @xmath0 and @xmath1 other than @xmath2 we prove that the ion configuration in the ground state is not periodic .
instead there is phase separation . for densities in @xmath3
the ground state ion configuration is a mixture of the density @xmath1 and @xmath2 ground state ion configurations . for the interval
@xmath4 it is a mixture of the density @xmath2 and @xmath0 ground states .
epsf 0.0 in 0.0 in 0.0 in |
the gravitational collapse of a spherically - symmetric massless scalar - field has two possible end states . either the scalar field dissipates away leaving a flat spacetime or a black - hole forms .
numerical simulations of this model problem @xcite have revealed an unexpected critical behavior when the initial conditions are close to a critical case @xmath5 ( @xmath6 is some parameter which characterizes the strength of the initial scalar field , and @xmath7 is the threshold value ) .
more precisely , choptuik has found a power - law dependence of the black - hole mass on critical separation @xmath0 of the form : @xmath8 for @xmath9 , and a discrete echoing with a period , @xmath3 , ( a discrete self - similar behavior ) for @xmath5 .
subsequently similar critical behavior have been observed for other collapsing fields : axisymmetric gravitational wave packets @xcite , spherically symmetric radiative fluids @xcite and complex scalar fields @xcite . in all these model problems the critical exponent @xmath2 turned out to be close to the value originally found by choptuik 0.37 , suggesting a universal behavior .
however , maison @xcite has shown that for fluid collapse models with an equation of state given by @xmath10 the critical exponent strongly depends on the parameter k. in this work we conjecture the existance of a small periodic correction , @xmath11 $ ] , to the power - law dependence of the black - hole mass .
@xmath12 is periodic and its period , @xmath1 , is universal and it depends on the previous universal parameters as @xmath4 .
our analytical argument predicts the existence of the fine - structure periodic term and its expected period .
the argument is based upon the final stage of a super critical evolution : from the moment when the deviation from the exact self - similar critical evolution becomes larger than some given value ( and the evolution is no longer self- similar ) up to the horizon formation .
we then provide a numerical evidence that verifies the existence of the conjectured periodic term , the universality of its period and the relation @xmath4 .
our numerical formalism is based on the characteristic scheme of goldwirth and piran @xcite to which we have added an expansion near the origin which is essential to achieve the extremely high accuracy needed for these computations .
the evolution equations , our algorithm and numerical methods , and the discretization and error analysis are all described in a previous paper @xcite , and will not be repeated here .
we consider the spherical collapse of a massless scalar field .
choptuik has shown that for a critical parameter @xmath7 there is a critical solution which has an infinite discrete self - similar behavior .
the critical solution by itself does not yield the black - hole mass scaling relation in which we are interested .
we perturb , therefore , the critical initial conditions .
this leads to a dynamical instability - a growing deviation from the critical evolution toward either sub - critical dissipation or supercritical black - hole formation .
let @xmath13 be a function of @xmath14 , the time coordinate of an observer at rest at the origin , that characterize the solution along the outgoing null geodesic that leaves the origin at @xmath14 .
the function , @xmath15 , could be , for example , the maximal value of @xmath16 along this geodesic .
following evans and coleman @xcite and maison @xcite we describe the run - away of the perturbed solution from the critical evolution , ( described by @xmath17 ) as a power - law : @xmath18 where the critical solution reaches the zero - mass singularity at @xmath19 .
the prefactor @xmath20 satisfies : @xmath21 .
we assume that the range of validity of the perturbation theory is restricted to some maximal deviation , @xmath22 , from the exact critical evolution , i.e. the evolution is approximately self - similar until @xmath23 . from here on the evolution is outside the scope of the perturbation theory - there is sub - critical dissipation of the field or supercritical black - hole formation . in either case , the evolution from this stage onwards looses its self - similar character .
we chose now @xmath9 so that the perturbed initial conditions develop into a black - hole .
the time @xmath24 required in order to reach the maximal deviation is given simply by the relation : @xmath25 of course , a larger initial perturbation requires a shorter time to reach this value .
we define now the logarithmic time , @xmath26 $ ] , in which the critical solution is periodic .
the logarithmic time , @xmath27 , which corresponds to the loss of self - similarity , is given by @xmath28 where @xmath29 depends on @xmath22 , @xmath30 and @xmath31 . the index @xmath31 denotes the family of initial conditions considered .
we conjecture now that the logarithmic time until the horizon formation , equals to @xmath27 plus a periodic term @xmath32 $ ] : @xmath33 \ . \label{tbh}\ ] ] the period , @xmath1 , is universal and it depends on the previous universal parameters according to : @xmath34 consider now two different initial - conditions , which lead to n and n+1 echoes respectively ( until the deviation from the critical self - similar evolution reaches @xmath22 ) .
these solutions are related to each other by an exact scale transformation with a factor @xmath35 .
the final stages of these two supercritical evolutions , from the stage when the deviation from the exact critical evolution reaches @xmath22 , ( and the evolution ceases to be periodic in @xmath36 ) , up to the horizon formation , are equal up to a scaling transformation .
the periodic nature of the function f arises from this final stage : the period of the function f is the amount that should be added to the quantity @xmath37 , in order to reduce the number of echoes by one .
this will reduce @xmath38 by @xmath3 . from eq .
[ tbh ] this amount to a period @xmath39 in @xmath40
. this conjecture is verified by numerical simulations of four families of initial data ( two neutral and two charged ) . in all those families
we have found that @xmath41 as a function of @xmath37 was well fit by a straight line with a slope @xmath42 . on top of this straight line
there was a small modulation .
the deviation from a straight line is shown in fig .
1 . which provides a numerical evidence for the existence of the periodic term f in eq .
we see that the function f is indeed periodic , with a universal period @xmath43 .
we have to relate , now the exponent @xmath44 to the critical exponent @xmath45 ( which describes the power - law dependence of the black - hole mass ) and then generalize choptuik s scaling - relation by proving that one should also add a periodic term to choptuik s mass scaling relation . we write @xmath38 in the form : @xmath46 where @xmath47 is the initial logarithmic time required for the system to settles down to a periodic behavior in @xmath36 , and @xmath48
is the number of echoes .
we assume that @xmath47 is independent of @xmath0 . using eq .
[ tbh ] we obtain : @xmath49 where @xmath50 is a family - dependent constant .
we define @xmath51 as the mass after @xmath48 echoes ( note that this is not the final black hole mass ) .
since @xmath52 decreases in each echo by a factor @xmath53 we have , using eq .
[ ndelta ] : @xmath54 from which it follows that @xmath55 . to obtain , @xmath56
, the final black - hole mass one should multiply @xmath51 by a periodic function @xmath57 $ ] which measures the change of mass , from the stage when the evolution is no longer periodic in @xmath36 , until the horizon forms .
the function @xmath58 depends only on the field configuration at the moment when the deviation from the exact self - similar evolution reaches @xmath22 ( and the evolution is no longer self - similar ) .
thus , @xmath58 is expected to have the same value each time the system completes another echo , i.e. each time n increases by unity .
( 6 ) we find that the function @xmath57 $ ] is expected to have a period of @xmath59 .
thus , we obtain @xmath60 \ , \ ] ] where @xmath61 is a family - dependent constant and @xmath62 $ ] is a periodic function with a _ universal _
period , @xmath1 . fig .
2 depicts this periodic fine structure for our four families of solutions mentioned earlier . in all four families
we obtain the basic power law behavior with @xmath63 .
2 displays the deviation of @xmath64 from this straight line as a function of @xmath37 .
the agreement between the four families shows that the fine structure is indeed universal with the expected period .
one may worry , of course , whether this fine structure is real or could it arise from some numerical errors . in a previous paper @xcite
we have established the stability and convergence of our code with numerous tests .
still becuase of the importance of this issue we demonstrate here the physical character of this fine structure .
3 depicts the deviations of @xmath64 from a straight line as a function of @xmath37 for five different calculations with 100 , 200 , 400 , 800 and 1600 grid points for the same initial data .
the same features were found on all grids even though the grid sizes differ by a factor of 16 .
the five curves overlap and all show the same periodic behaviour . numerical convergence
( the 800 and 1600 curves are nearer than the 100 and 200 curves , for example ) is clearly seen .
we have studied the spherical gravitational collapse of a massless scalar - field , both for the uncharged case and for the charged configurations .
our main interest was the supercritical @xmath65 feature of choptuik s solution , i.e. the power - law dependence of the black - hole mass on the critical separation .
we have shown the existence of a fine - structure above this power - law dependence in the form of a periodic term with a universal period , @xmath1 .
we are not aware of such fine - structure periodic term in any other phase - transitions in statistical mechanics .
our periodic term with its period strongly depends on the discrete echoing character of the critical solution .
this discrete self - similarity has been seen only in the works of choptuik@xcite , abrahams and evans @xcite ( collapse of axisymmetric vacuum gravitational field ) and in our work concerning the gravitational collapse of a charged ( complex ) scalar field @xcite .
abrahams and evans @xcite have found an echoing period of @xmath66 .
thus , using our analytical argument , we conjecture that a careful analysis will reveal a periodic fine - structure ( to the power - law behavior ) , with a period of @xmath67 , in the model problem of the collapse of axisymmetric gravitational wave packets .
99 s. hod , msc .
thesis , the hebrew university , jerusalem ( 1995 ) .
choptuik , phys .
70 * , 9 , ( 1993 ) .
a.m. abrahams and c.r .
evans , phys .
rev.lett . * 70 * , 2980 , ( 1993 ) .
evans and j.s .
coleman , phys .
lett . * 72 * , 1782 , ( 1994 ) .
s. hod and t. piran , preprint ( 1996 ) .
d. maison , phys .
* b366 * , 82 , ( 1996 ) .
d.s . goldwirth and t. piran , phys . rev . * d36 * , 3575 , ( 1987 ) . | we conjecture ( analytically ) and demonstrate ( numerically ) the existence of a fine - structure above the power - law behavior of the mass of black - holes that form in gravitational collapse of spherical massless scalar field @xcite .
the fine - structure is a periodic function of the critical - separation @xmath0 .
we predict that the period @xmath1 is _ universal _ and that it depends on the previous universal parameters , the critical exponent , @xmath2 , and the echoing period @xmath3 as @xmath4 .
= -1 cm = -1 cm |
for smooth , non - singular lenses , it is well - known that the total number of images is odd and that the number of even parity images exceeds the number of odd parity images by one ( e.g. , burke 1981 ; schneider , ehlers & falco 1992 , chap . 5 ) . when the source lies within both the tangential and radial caustics , it is lensed into 5 images . when the source lies outside one caustic but inside the other ,
then it is lensed into 3 images .
if the surface density at the galaxy centre is cusped and that cusp is stronger than isothermal , then the situation changes .
there is no radial caustic and there are either 4 or 2 images depending on whether the source lies inside or outside the remaining tangential caustic ( evans & wilkinson 1998 ) .
in fact , almost all the known sixty or so gravitational lens systems are 2 or 4 image configurations ( see pospieszalka et al .
1999 for details of the gravitational lensing database which maintains a list of candidates ) .
there are two systems in which the presence of a weak , central image has been claimed .
apm08279 + 5255 is an ultraluminous broad absorption line quasar .
originally , two images were detected serendipitously by irwin et al .
( 1998 ) in the optical .
subsequently , ibata et al .
( 1999 ) found convincing evidence for a central third image using higher resolution infrared imaging .
the source of mg1131 + 0456 is a radio galaxy .
one extended radio component is lensed into a ring , another is lensed into two images
. there seems to be a weak central image at the center of the ring , and this may be the third or fifth image of parts of the background radio source ( chen & hewitt 1993 ) ; however , this could also be emission from the lensing galaxy .
the almost total absence of the central image from the known lens systems sets strong constraints on the core radius and the steepness of the lensing potential ( e.g. , narayan , blandford & nityananda 1984 , wallington & narayan 1993 , rusin & ma 2000 ) .
the aims of this paper are twofold . on the theoretical side ,
our aim is to find expressions for the sums of the magnifications of any subset of the images produced by multiply lensed quasars or galaxies , at least within the framework of a class of flexible and popular models .
we do this by extending the analysis of our earlier work ( hunter & evans 2001 , henceforth paper i ) to lensing potentials stratified on similar concentric ellipse but this time with cores . in doublet ( or quadruplet ) systems , there are one ( or two ) even parity images , one ( or two ) odd parity images , together with the missing central image .
we provide contour integrals which find the sums of the magnifications of the even parity images , the odd parity images and the central image separately . on the astrophysical side ,
our aim is to set constraints on the cusp profile and core size of lensing galaxies by requiring that the central image be too weak to be detectable .
for the 2 and 4 image systems , we investigate the permissible core radius as a function of the steepness and the flattening of the lensing potential and the external shear .
the paper is arranged as follows .
section 2 develops the contour integral representation needed for lensing potentials that are elliptically stratified .
section 3 specialises the analysis to models in which the lensing potential is a power - law of the elliptic radius combined with external shear of arbitrary orientation .
the conditions for triple and quintuple imaging are found in terms of the depth and shape of the potential , as well as the position of the source .
the contour integrals for the sums of the magnifications of the two even parity , the two odd parity and the missing central image are evaluated .
section 4 uses monte carlo simulations to set limits on the lensing potential by requiring the central image to be unobservable .
finally , our conclusions are given in section 5 .
this section develops contour integrals for computing the magnifications of images assuming only that the lensing potential is stratified on similar concentric ellipses . in this paper
, we always assume that the lensing potential @xmath11 is stratified on similar concentric ellipses with constant axis ratio @xmath1 with @xmath12 so that @xmath13 for a thin lens with potential @xmath11 , the lens equation is ( e.g. , schneider et al . ,
section 5.1 ) @xmath14 where ( @xmath15 ) are cartesian coordinates of the source . here
@xmath16 and @xmath17 allow for a constant external shear in an arbitrary direction .
we always assume that the lens is not circularly symmetric ( that is , either @xmath18 or @xmath19 or @xmath20 ) .
complex numbers often simplify calculations in lensing theory ( e.g. , bourassa , kantowski & norton 1973 , bourassa & kantowski 1975 , witt 1990 , rhie 1997 ) . as in paper
i , we shall find it helpful to use @xmath21 the lens equation ( [ eq : original ] ) then becomes @xmath22 where @xmath23,\qquad q = \ffrac{1}{2 } [ 1- q^2 + \gamma_1(1+q^2 ) + 2{\rm i}q\gamma_2 ] .\ ] ] it and its conjugate can be written in matrix form as @xmath24 where @xmath25 , from which it follows that @xmath26 when we form the real quantity @xmath27 from eqn .
( [ eq : zzetarel ] ) , we find that the solutions of the lens equation also satisfy the real equation @xmath28 ^ 2 = 0.\ ] ] we designate @xmath29 as _ the imaging equation_. its real and positive solutions for @xmath30 provide the image positions @xmath31 for a given source location @xmath32 .
once a solution for @xmath30 is found , the image positions are given by equation ( [ eq : zzetarel ] ) .
any solution of the original lens equation ( [ eq : original ] ) gives a solution of the imaging equation .
however , the imaging equation may have negative real or complex solutions which do not correspond to true images .
the lens equation defines a map from the complex @xmath33lens plane to the @xmath34source plane .
the jacobian of this mapping is @xmath35 where the second form has been evaluated using eqn ( [ eq : lenseq ] ) .
the eigenvalues and eigenvectors of @xmath36 determine the distortion of images .
if the eigenvalues are both positive , the image is direct ; if the eigenvalues are both negative , the image is doubly inverted .
these are the even parity cases . if one eigenvalue is positive , the other negative , then the image is inverted and the parity is odd .
the reciprocal of the determinant of the jacobian of the mapping from the real @xmath37lens plane to the @xmath38source plane corresponds physically to the signed magnification of an image ( schneider et al .
1992 , chapter 5 ) . with our choice ( [ eq : complexcoords ] ) of complex coordinates
, we find @xmath39 where @xmath40 is the absolute value of the magnification and @xmath41 is the parity of the image located at @xmath42 . for certain positions , @xmath43 vanishes and
the magnification is infinite .
these are the critical points and lines .
the caustics are the images of the critical points and curves under the lens mapping ( [ eq : original ] ) . evaluating the determinant of ( [ eq : jmatrix ] )
gives @xmath44 differentiating the imaging equation ( [ eq : defnk ] ) for @xmath45 partially with respect to @xmath30 , holding @xmath32 and @xmath46 fixed , and then using the lens equation ( [ eq : lenseq ] ) to express @xmath32 and @xmath46 in terms of @xmath31 and @xmath47 gives @xmath48.\ ] ] consequently , we get the following compact expression for the signed magnification of an image as @xmath49 our contour integral representation relies on the special structure of equation ( [ eq : defmag ] ) .
images correspond to simple zeros of @xmath50 and @xmath51 is the residue of @xmath52 at a simple zero .
we continue the right hand side of equation ( [ eq : defmag ] ) into the complex @xmath30-plane and write the sum of the signed magnifications of the images as the contour integral @xmath53 here , @xmath54 is a contour in the complex @xmath30-plane which excludes @xmath55 and encloses only the simple poles corresponding to whichever visible images we wish to analyse .
let us note that our analysis here shows how the methods first developed in paper i can do more than we achieved there .
we here exploit them in two new ways .
we first relax the restriction that the lensing potential be scale - free and allow it to have a core .
cored potentials produce either one , three , or five images depending on the strength of the potential , the size of the core and the position of the source .
second , we derive formulae for sums of signed magnifications for separate image pairs , and for the central image . hence , the present results are more detailed and informative than those of paper i , most of which were for sums over the four bright images , consisting of two pairs of opposite parity , and which therefore partially cancel .
we now specialize the analysis to the case of power - law galaxies with a core radius @xmath3 , which have the lensing potential @xmath56 these models were first studied by blandford & kochanek ( 1989 ) in the context of gravitational lensing ( see also kassiola & kovner 1993 ; witt 1996 ; witt & mao 1997 , 2000 ; evans & wilkinson 1998 and paper i ) .
they are the projections of three - dimensional power - law galaxies familiar in galactic astronomy and dynamics ( evans 1993 , 1994 ) .
for example , they have been used to model the nearby elliptical galaxy m32 ( van der marel et al . 1994 ) , the inner parts of the galactic bulge ( evans & de zeeuw 1994 ) , as well as the dark halo of our own galaxy in the interpretation of both the sagittarius stream and the microlensing results ( alcock et al .
1997 ; ibata & lewis 1998 ) .
the convergence ( or surface density in units of the critical density ) is @xmath57 it is easy to see that @xmath58 where the parameter @xmath59 and has the range @xmath60 .
the positive parameter @xmath61 , to which the magnitude of the lensing potential is proportional , can be removed by a rescaling .
specifically , we scale all lengths by @xmath62 .
the net effect is to set @xmath63 in eqs ( [ eq : lensing_pot])-([eq : lensing_pottwo ] ) .
the dependence of our results on @xmath61 , which is needed in applications , can be recovered by multiplying all powers of @xmath3 , @xmath31 and @xmath32 and their conjugates by the same powers of @xmath62 . in this section ,
we establish the conditions for the numbers and types of images , and the forms of the caustics .
it is convenient to use the variable @xmath64 .
it reduces to the same variable @xmath65 used in paper i in the limit of no core ( @xmath66 ) . when there is a core , the physically relevant range of @xmath65 is restricted to @xmath67 . in terms of the variable @xmath65 ,
@xmath68 the imaging equation ( [ eq : defnk ] ) requires the balance @xmath69 for those real and positive values of @xmath65 at which the images occur .
the values of @xmath70 and @xmath71 depend only on the flattening and the shear , and are defined by @xmath72 it follows from the definitions ( [ eq : pqdefn ] ) that @xmath73 and @xmath74 provided that @xmath75 . hence ,
both @xmath70 and @xmath71 are positive with @xmath76 and @xmath77 .
[ fig : chrisone ] shows how image positions can be found graphically by plotting separately the two sides of equation ( [ eq : tbalance ] ) .
the full curve represents the right hand side of that equation .
it depends only on the parameters of the lens .
the left hand side depends also on the position of the source .
the three dashed curves display it for three different source positions .
the lowest dashed curve intersects the full curve curve five times and there are five images . when the source is sufficiently close to the center of the galaxy , the left hand side remains small until it rises to its asymptote at @xmath78 .
a requirement for quintuple imaging is that @xmath70 must lie to the left of this asymptote , and hence that the core radius must satisfy @xmath79 .
images disappear in pairs either when the two curves touch in the @xmath80 interval , which corresponds to a crossing of the tangential caustic , or when they touch in the @xmath81 interval , which corresponds to a crossing of the radial caustic .
if the former happens first , then the three images that remain form a `` core triplet '' ; if the latter happens first , then the three images form a `` naked cusp triplet '' ( e.g. , kassiola & kovner 1993 ) .
the intermediate dashed curve in fig .
[ fig : chrisone ] represents a naked cusp triplet case of a source which lies outside the radial caustic but inside the tangential one .
the topmost dashed curve is for a source which lies outside both caustics , and gives a single image .
there is now just a single image corresponding to a single crossing on @xmath82 .
the three different source positions are shown in fig . [
fig : christwo]a . the maximum number of images diminishes as the core radius increases .
if the core radius satisfies @xmath79 , then quintuple imaging is possible , and there are two caustics . for core radii in the range @xmath83 , the vertical asymptote of the dashed curve in fig .
[ fig : chrisone ] lies between @xmath70 and @xmath71 .
there are then three images for the source sufficiently close to the center of the galaxy that it lies within the tangential caustic , and there is no radial caustic . for still larger core radii satisfying @xmath84 ,
then only a single image can occur and there are no caustics . appendix a justifies the statements concerning images and caustics , and gives equations for determining the caustics , and an approximate formula for the radial caustic for small @xmath3 and @xmath85 .
image positions and points on caustics must generally be determined numerically .
cusps are an exception because they occur in pairs when either @xmath86 or @xmath87 is a triple root of the imaging equation .
the sole caustic is tangential and of lips type for the range @xmath88 as in fig .
[ fig : christwo]c because it has only the @xmath86 pair of cusps .
when @xmath89 and quintuple imaging is possible , there are two pairs of cusps and three possible configurations of caustics . either both pairs of cusps lie on the tangential caustic , which is of astroidal shape , or else both caustics are lips - shaped with the radial caustic lying within the tangential and oriented oppositely to it ( schneider et al . 1992 ,
section 8.6.1 ) .
this double lips configuration occurs when @xmath90 as in fig .
[ fig : christwo]a .
the @xmath86 pair of cusps are then naked .
the stage at which those cusps become naked is described by the coupled pair of equations ( [ eq : nakedness ] ) , which must be solved numerically .
the results are plotted in fig .
[ fig : christhree ] .
the tangential caustic has a pair of naked cusps in the regions of parameter space below the curve for the appropriate value of @xmath91 . above that curve , the tangential caustic has four cusps and lies wholly within the radial caustic as in fig .
[ fig : christwo]b .
below that curve , but where @xmath92 is sufficiently large that the condition ( [ eq : twolips ] ) is violated , the two caustics intersect as in fig .
[ fig : christwo]d , and two of the four cusps of the tangential caustic are naked .
we note that our results extend previous calculations in the literature .
for example , kassiola & kovner ( 1993 ) give the conditions for multiple imaging in the isothermal case ( @xmath93 ) in the absence of shear ( @xmath94 ) , while evans & wilkinson ( 1998 ) give the results for the scale - free cases ( @xmath95 ) with on - axis shear only ( @xmath96 ) .
changing the variable in the contour integral ( [ eq : tautheorem ] ) for the magnification to @xmath65 gives @xmath97 we write the denominator term as @xmath98 where we have labelled its two parts as @xmath99 the component @xmath100 is the same as in paper i , whereas @xmath101 differs from the @xmath102 of paper i only through its extra @xmath103 factor , which is unity when there is no core . in paper
i , we were interested primarily in sums over the four bright images
. however , the contour integral ( [ eq : ttheorem ] ) , like the corresponding result in equation ( 23 ) of paper i , applies equally well to any subset of the images provided that the contour @xmath54 is chosen to enclose only the zeros of @xmath104 which correspond to that subset .
if @xmath105 for instance , we define a contour @xmath106 to be a simple closed contour in the complex @xmath65plane which encircles the two direct images of even parity .
these approach @xmath86 as @xmath107 , and so the contour @xmath106 also encloses @xmath71 which lies in between , but it excludes all the other images , as well as @xmath108 , @xmath87 , @xmath109 and any complex zeros of @xmath103 .
then , we obtain the formula @xmath110 we require that @xmath106 keep a finite distance from @xmath70 , @xmath71 , and every zero of @xmath103 .
then , with @xmath111 bounded above , and @xmath112 , @xmath113 , and @xmath114 bounded from below , it follows that @xmath115,\end{aligned}\ ] ] and hence @xmath116 on @xmath106 for sufficiently small @xmath117 .
we have here used the result @xmath118[\bar\lambda(t_2-t)+\bar\nu(t_1-t ) ] = |\lambda|^2(t_2-t)^2+|\nu|^2(t_1-t)^2,\ ] ] where @xmath119 and @xmath120 are the following two linear combinations of the complex source coordinates @xmath121 which have the property @xmath122 . with @xmath116
, we can now expand the @xmath123 term in the integral ( [ eq : directmags ] ) as a geometric series to get @xmath124^j \over ( t - t_1)^{2j+1}(t - t_2)^{2j+1}(1-a^2t^b)^{j+1}}dt.\end{aligned}\ ] ] the restrictions that we imposed earlier on @xmath106 guarantee that this series converges for sufficiently small @xmath117 .
every integral in it can be evaluated from its residue at @xmath86 alone because that is the only singularity within the contour @xmath106 .
the leading order term is the residue of the simple pole at @xmath86 , which is positive because both parities are even .
the full expansion for the sum of the two magnifications is @xmath125 here , @xmath126 is a correction term which vanishes if the source is exactly aligned with the center of the lensing galaxy .
more generally , it takes the form : @xmath127 where we have defined @xmath128 in fact , we can obtain closed form expressions for all of these integrals without any further residue calculus by simple partial differentiation because @xmath129.\nonumber\end{aligned}\ ] ] if the conditions for quintuple imaging are satisfied , then there is a pair of inverted images which approach @xmath87 as @xmath130 .
we can perform a similar analysis for a simple closed contour @xmath131 in the complex @xmath65-plane which encloses the images and @xmath70 , but not @xmath108 , @xmath86 , @xmath109 or any complex zero of @xmath103 .
the result is that @xmath132 where the correction term @xmath133 is @xmath134 here , we have defined @xmath135\nonumber\\ & = & r_2(t_2,t_1,a^2;j , k,\ell).\nonumber\end{aligned}\ ] ] in fact , the coefficients @xmath136 and @xmath137 can be expressed as finite sums , as is demonstrated in appendix b. neither series ( [ eq : directseries ] ) nor ( [ eq : invertedseries ] ) remains convergent at the tangential caustic where one of the direct and one of the inverted images merge ; the magnifications of the merging images become infinite and one can not then construct the @xmath131 and @xmath106 contours .
however , the sum of series ( [ eq : directseries ] ) and ( [ eq : invertedseries ] ) can remain convergent at the tangential caustic because the infinities of the signed magnifications of the two merging images cancel .
the series ( [ eq : invertedseries ] ) ceases to converge at the radial caustic where an inverted and a doubly inverted image merge , but the series ( [ eq : directseries ] ) for the direct images converges there if , as in fig .
[ fig : christwo]a , it lies inside the tangential caustic .
let us note some interesting special cases of the preceding formulae . in the scale - free limit ( @xmath95 )
when @xmath138 or @xmath139 , then remarkably the correction terms have the property that @xmath140 . as first discovered by witt & mao ( 2000
; see also paper i ) , the sum of the four signed magnifications is then an invariant completely independent of the source position @xmath141 even though this result is not exact for other values of @xmath91 , it is often an extremely good approximation . if the core radius is non - zero , then the sum of the signed magnifications becomes @xmath142 , \label{eq : approx}\ ] ] and this approximation remains remarkably accurate inside the radial caustic .
an example is the cored isothermal lens of fig .
[ fig : christwo]b .
the sum of the four signed magnifications is constant to within tenths of one percent over all of the region inside the tangential caustic except very close to the cusps .
however , the smallness of the dependence of that sum on the source position is due to a near - cancellation of the correction terms in the two image sums .
[ fig : chrisfour ] shows how much the sum of the magnifications of the two direct images varies over the same region .
numerically , the leading coefficients in expansion ( [ eq : ctwoseries ] ) for @xmath126 are @xmath143 and @xmath144 .
the sums @xmath145 for the same sets of indices are -0.01 .
though numerical values vary , @xmath145 is typically at least two orders of magnitude less than @xmath146 for lenses with small cores and with an inner tangential caustic . in section b.3
, we show how near - cancellation can occur more generally from the contour integral formula ( [ eq : tautheorem ] ) , and is not a peculiarity of power - law galaxies . however , this behavior is not universal .
the sum of the four signed magnifications has large variations over the region inside the inner radial caustic of fig .
[ fig : christwo]a , while the sum of the magnifications of the two direct images changes little in the middle third of that region .
that middle third lies well inside the tangential caustic at which the direct image sum becomes large .
the separate sums of the pairs contain more information than their signed combination .
physically speaking , it is the total magnification ( that is , the sum of the magnifications ) which is more interesting , but this varies considerably with source position .
the minimum magnification @xmath147 in the scale - free limit @xmath148 is @xmath149 this corresponds to the case when the source is aligned with the center of the lensing galaxy .
it is a good approximation only when the source offset is small . for non - zero core radius
, it becomes @xmath150.\ ] ] in the next section , we discuss another nearly constant four - image sum which reduces to the sum of the unsigned magnifications when the source is aligned with the center of the lensing galaxy .
the contour integral method can also be used to calculate configuration moments , that is , sums over the images of products of the signed magnifications with position coordinates .
those sums are obtained by adding the complex position coordinates from ( [ eq : zzetarel ] ) to the sum ( [ eq : tautheorem ] ) to obtain @xmath151 we expand for moments in the same manner as for the magnifications , with the result that @xmath152^{m+j } [ \barlam(t_2-t)+\barnu(t_1-t)]^{n+j}dt \over[(t - t_1)(t - t_2)]^{2j+m+n+1}(1-a^2t^b)^{j+1}}.\ ] ] after the two numerator factors are expanded binomially , we see that every integral is of the same form as ( [ eq : rtwodefn ] ) , and so , provided @xmath153 and @xmath154 , we obtain @xmath155 the corresponding sum for the indirect pair is @xmath156 in either case the same @xmath157 and @xmath146 functions as arise with the magnifications , and which are evaluated in appendix b , are all that is needed . whereas the present paper is mostly concerned with magnifications , our modeling in paper i made use of configuration moments too .
the analysis of this paper , and especially that of appendix section b.3 , shows that four - image sums of moments may also be far more uniform than the separate sums ( [ eq : dirconfigsum ] ) and ( [ eq : invconfigsum ] ) because of cancellations .
however it is possible to mitigate the cancellations which arise from signed magnifications by using specially contrived configuration moments . as an example , consider the zeroth order @xmath158 , @xmath159 moment [ see eqn .
( 48 ) of paper i ] in which the magnifications are weighted by the complex exponential @xmath160 , where @xmath161 .
the angle @xmath162 for each image is @xmath163this factor exactly compensates for the different signs of the magnifications when the source offset is zero and the two kinds of images lie on perpendicular lines through @xmath164 .
we obtain the @xmath165 terms directly from the integral ( [ eq : dirconfigint ] ) and its pair because the binomial expansions used to obtain equations ( [ eq : invconfigsum ] ) and ( [ eq : dirconfigsum ] ) are not valid when @xmath166 .
this gives us two @xmath167 integrals whose sum gives @xmath168,\ ] ] because @xmath169 .
the sum ( [ eq : newmagsum ] ) is minus the sum in equation ( [ eq : amincore ] ) when there is no off - axis shear and @xmath170 is real because @xmath171 on the major @xmath172-axis where the inverted images initially lie and @xmath173 on the minor @xmath174-axis .
off - axis shear rotates the configuration and @xmath175 is then the complex factor which makes @xmath176 real .
this four - image sum can remain nearly constant as the source offset changes because the changing angular positions of the images counteracts their changing magnitudes .
for the cored isothermal lens for which fig .
[ fig : christwo]b and fig .
[ fig : chrisfour ] are plotted for example , the @xmath158 , @xmath159 moment of equation ( [ eq : newmagsum ] ) displays the same near constancy over the region inside the tangential caustic as does the sum of the four signed magnifications .
estimates of @xmath1 and the orientation of the lens must be combined with observed positions to evaluate the sum in ( [ eq : newmagsum ] ) , but the result should be real when multiplied by @xmath175 .
there is no fifth central image in the absence of a core for @xmath85 .
when there is a core , the central image is given by the root of the imaging equation ( [ eq : newimaging ] ) for which @xmath177 and @xmath178 as @xmath130 .
we can obtain formulae for its signed magnification involving another set of @xmath179-functions , defined by @xmath180 where @xmath181 is a loop enclosing @xmath109 but not @xmath70 or @xmath71 or @xmath182 or any other zeros of @xmath103 .
these functions can be calculated by evaluating residues at @xmath109 .
this is a simple pole when @xmath165 , and we then get @xmath183^{\ell+1}[1-a^{2/b}t_2]^{k+1}}.\ ] ] the signed magnification of the central image is given by @xmath184[1-a^{2/b}t_2 ] } + c_a(t_1,t_2,a^2;\lambda,\nu).\ ] ] where the correction term @xmath185 has the form @xmath186 the lowest order term in ( [ eq : central ] ) is positive if @xmath187 when this image is doubly inverted , negative if @xmath188 and the image is simply inverted , and positive if @xmath189 and it is the only and direct image .
when @xmath3 is small , as in our applications , so that @xmath190 and the central image occurs for a large value of @xmath65 , equations ( [ eq : central ] ) and ( [ eq : centralc ] ) can be approximated by @xmath191 these special cases of integrals ( [ eq : centralintegral ] ) can be evaluated by residues ( change to @xmath192 as integration variable ) and give a series expansion for the magnification of the central image as @xmath193 , \quad \alpha_j={\gamma \left[{2(j+1 ) \over b}+1 \right ] \over j!\gamma \left[{2(j+1 ) \over b}+1-j \right]}.\ ] ] the series is especially simple for @xmath139 when it is binomial and gives @xmath194 there are also two cases for which the series is hypergeometric and can be summed explicitly .
they are @xmath138 , for which @xmath195 ^ 2},\ ] ] and @xmath196 , for which @xmath197.\ ] ] the reason for the simplifications is that they are most easily obtained directly from the basic magnification formula ( [ eq : ttheorem ] ) after its integrand is approximated for large @xmath65 to give @xmath198}.\ ] ] the root for the central image and its residue can be calculated directly without any need for series expansion for the three special values of @xmath91 .
more generally , series expansion is needed .
this analysis makes it clear that the basic requirement for the approximation ( [ eq : muasum ] ) to be valid is that @xmath65 be large at the central image .
it is evident that the series ( [ eq : muasum ] ) shows @xmath199 to be an increasing function of @xmath200 for @xmath201 because all its coefficients @xmath202 are then positive .
numerical evaluation , as in fig .
[ fig : chrisfive ] , shows that @xmath199 remains an increasing function for larger @xmath91 .
those plots extend to @xmath203 of the radius of convergence .
the series converge for @xmath204 as follows from using stirling s formula to approximate the gamma functions and standard convergence tests .
hence the range of usefulness of the series depends considerably on the value of @xmath91 . for @xmath85 and @xmath3 small ,
the radius of convergence is large , and the series converges out to the radial caustic as approximated by equation ( [ eq : radcausticsmalla ] ) .
the magnification of course becomes large as the caustic , at @xmath205 for @xmath139 and @xmath206 for @xmath138 , is approached . for @xmath207 on the other hand , the radius of convergence is caused by the breakdown of the approximation that the central image occurs for large @xmath65 and , as with equation ( [ eq : muabfour ] ) , @xmath199 remains finite as the condition ( [ eq : radconv ] ) is violated . that radius of convergence in @xmath117 is proportional to the positive power @xmath208 .
consequently , the approximation ( [ eq : muasum ] ) is useful for only a small part of the region within the radial caustic when the core radius @xmath3 is small , and @xmath199 , which is growing with source offset on a short length - scale , can continue to become considerably larger before the radial caustic is approached .
numerical tests shows that the approximation ( [ eq : muasum ] ) works well where the theory predicts that it will ; over much of the region within the radial caustic for @xmath85 , but only in limited central regions for @xmath207 .
it underestimates @xmath199 because of the neglect of denominator terms such as those at the leading order in equation ( [ eq : central ] ) .
that underestimate can be by 10 % or more for @xmath209 , but accuracy increases substantially for smaller @xmath3 .
here , we study the problem of missing central images from the standpoint of the structure of the lensing galaxy . in optical lenses ,
the experimental constraint is weak , as any central image can be masked by emission from the lensing galaxy .
radio - loud lenses provide much stronger constraints , as they have typically been probed with high dynamic range radio maps . the _ cosmic lens all - sky survey _ ( class ) is the largest statistically homogeneous search for gravitational lenses ( e.g. , myers 1999 ) .
the survey sample contains @xmath210 flat - spectrum radio sources . in the best cases , such as b0218 + 357 ( biggs et al .
1999 ) and b1030 + 074 ( xanthopoulos et al . 1998 ) , the magnification ratio of the faintest to the brightest image @xmath211 is constrained to @xmath212 from the absence of a detectable central image in the map . as a typical detection limit in the below calculations , we adopt @xmath213 for central images .
this is appropriate for radio lenses , but not for optical lenses .
we perform monte carlo simulations using the power - law galaxies ( [ eq : lensing_pot ] ) . sources are placed randomly within the outermost caustic for choices of @xmath4 , @xmath1 and core radius @xmath3 .
the imaging equation ( [ eq : tbalance ] ) is solved numerically to find the roots @xmath65 of the images .
we then evaluate the magnifications using ( [ eq : defmag ] ) , and work out the ratio of the brightest to the faintest image .
repeating this many times gives us the raw probabilities of observing a central image .
figs [ fig : wynzero]-[fig : wyntwo ] show the probability of observing a central image as a function of the slope of the projected density ( @xmath214 ) for the case of the quintuplets , the core triplets and the naked cusp triplets respectively .
the panels show how the raw probabilities depend on the core radius @xmath3 , while the different lines in each panel show different flattenings .
the models have a singular density profile if @xmath215 , and are very nearly singular when @xmath3 is small .
we describe the parameter @xmath216 as the slope index , as it controls the fall - off of the projected density . for the isothermal cusp , @xmath217 and so the projected density @xmath218 .
models steeper than isothermal have @xmath219 , models shallower than isothermal have @xmath220 .
there are a number of things to notice in fig .
[ fig : wynzero ] . as the core radius @xmath3 is diminished , the rgime in which the central image is visible shrinks .
visible central images occur only for @xmath207 when the central image is stronger as the simple estimate @xmath221 from 3.4 predicts .
the full curves for the flattened ( @xmath222 ) models lie above the dashed curves for the rounder ( @xmath223 ) models .
this is due to a stronger dimming of the brightest image with increasing flattening which outpaces the dimming of the central image .
all magnifications share a @xmath224-dependence which comes from the basic magnification formula ( [ eq : ttheorem ] ) .
the brightest image is direct for vanishing shear and zero source offset , and has magnification @xmath225 .
hence we estimate the ratio of the central image to the brightest image at zero source offset to be @xmath226 .
this ratio increases with increasing flattening . for small @xmath3 ,
the magnification of the central image does not vary greatly with increasing source offset within the quintuplet region for @xmath85 .
this region inside the tangential caustic is but a small part of that within the radial caustic , and hence the abscissa of fig .
[ fig : chrisfive ] varies little .
conversely , the brightest image strengthens considerably with increasing source offset as it gets closer to the tangential caustic .
the result is that the ratio @xmath227 of central to brightest magnifications decreases with increasing source offset for @xmath85 . for @xmath207 on the other hand
, @xmath199 grows so strongly across the quintuplet region that it can outpace the growth of @xmath228 .
the net result is that the flatter the potential , and the larger the core radius , then the greater is the likelihood that the central image is bright enough to be visible .
[ fig : wynone ] shows the raw probability of observing a central image for core triplets .
the behavior of the curves as a function of slope index @xmath216 has the following explanation . if @xmath216 is too large , then the central image is highly demagnified and so the raw probability is vanishingly small .
as the slope index @xmath216 diminishes , the central image becomes brighter and the probability rises quickly to @xmath229 .
the brightest image is generally the remaining direct one .
it dims with increasing source offset whereas the central image brightens .
the tangential caustic also grows in size as @xmath1 decreases for constant slope index .
that is the main reason why central images are more visible with increased flattening for @xmath207 ; the average source offset in the core triplet region between the tangential and radial caustics is then larger , and hence @xmath199 is significantly larger .
as @xmath216 decreases further , the astroidal tangential caustic grows , and the area in the source plane generating core triplets diminishes and eventually vanishes , as the tangential caustic becomes larger than the radial caustic as in fig .
[ fig : christwo]c .
the smaller @xmath1 is , the sooner this happens and , as fig .
[ fig : wynone ] shows , the size of the core has little effect on the stage at which the core triplet region disappears .
the second and third terms of the inequality ( [ eq : twolips ] ) become equal at that stage and , in the absence of shear , @xmath216 is then @xmath230 ( see appendix a ) . if the core radius @xmath231 , then even models steeper than isothermal ( @xmath6 ) can provide observable central images . for smaller core radii ( @xmath232 ) , the visible central images are confined to models less steep than isothermal .
notice that the constraints on the maximum possible steepness of the lensing potential provided by the missing image are stronger for doublets than quadruplets once @xmath233 . for small core radii ,
the fifth image in quintuplet systems is significantly more demagnified than the third image in triplet systems .
however , for larger core radii , it is the quadruplets that provide the stronger restriction .
[ fig : wyntwo ] shows the raw probability of observing a central image for naked cusp triplets .
naked cusps are much less abundant for elliptic potentials as opposed to elliptic densities ( kassiola & kovner 1993 ) .
almost as soon as naked cusps appear , all three images are of roughly similar brightness and they are all detectable . hence , the raw probability of observing a central image shows a swift transition from nearly @xmath234 to nearly @xmath229 as soon as naked cups become possible .
the monte carlo results are consistent with the transitions given in fig .
[ fig : christhree ] .
equations ( [ eq : nakedness ] ) predict that the transitions occur at the values @xmath235 for @xmath209 and @xmath236 for @xmath237 .
there are no strong observational candidates for naked cusp triplets , and so we must conclude that fig . [
fig : wyntwo ] sets a firm lower limit on the slope index @xmath216
. this must be larger than the critical value which permits naked cusps ( given in fig .
[ fig : christhree ] ) , otherwise naked cusps would be common . to compare with data from surveys ,
we must allow for the amplification bias .
lens systems with a high total magnification @xmath238 are preferentially included in a flux - limited survey ( e.g. , turner 1980 , turner , ostriker & gott 1984 ) .
the flux distribution of the sources in class is well described by @xmath239 , where @xmath240 is the number of sources with flux greater than @xmath241 ( rusin & tegmark 2001 ) . for a flux limited sample ,
the probabilities that take into account amplification bias are @xmath242^{1.1}\ ] ] where @xmath243 denotes the area enclosed by the caustics in the source plane for which the central image passes the threshold ( e.g. , rusin & ma 2001 ) .
[ fig : wynfour ] is the analogue of fig .
[ fig : wynone ] when amplification bias is taken into account .
only the rightmost branch of the curve is plotted , as this is the most relevant for constraining the core size and the slope index .
notice that the effects of the amplification bias cause only slight changes in the shapes of the curves .
once the core radius falls below @xmath244 , then irrespective of the flattening , the lensing potential is constrained to be at least as steep as isothermal to ensure that the fraction of triplets with an observable central image remains low .
[ fig : wynfive ] shows , this conclusion remains valid even in the presence of shear .
shear has little direct effect on the magnitude of the central image as equation ( [ eq : muasum ] ) , which is independent of shear , predicts .
however , shears of the order of 0.1 cause the inner tangential caustic to grow significantly in size for the @xmath216 values at which the curves of fig .
[ fig : wynfive ] rise sharply , while the radial caustic merely tilts a little .
the greater visibility of the central image is again because the average source offset in the diminished core triplet region has become larger , now as a result of shear .
the total shear produced by internal misalignments , large - scale structure and neighboring galaxies is typically constrained by @xmath245 .
what restriction is implied on the core radius @xmath3 from the data on radio lenses ?
the class survey found no triples , but 7 doublets .
there is perhaps only 1 definite triple system ( apm08279 + 5255 ) out of a total of @xmath246 doublets and triplets on pospieszalka et al.s gravitational lensing database ( which contains both radio and optical lenses ) .
so , the probability of detecting a triple with an observable central image is certainly very low . in this paper
, we take it to be @xmath247 for radio lenses ( i.e. , at the @xmath248 threshold ) . using fig .
[ fig : wynfour ] , the top left panel showing curves for @xmath249 suggests that detectable central images are common for isothermal ( or nearly so ) galaxies .
the smallest core radius @xmath3 seemingly compatible with the missing central images is @xmath244 if the galaxy is isothermal . at a typical lens redshift of @xmath250 and a typical source redshift @xmath251
, this corresponds to a physical size of @xmath252 pc ( using formula ( 11 ) of young et al .
( 1980 ) and formula ( 2.4 ) of wallington & narayan ( 1993 ) ) .
the result holds good for a flat , matter - dominated friedman - robertson - walker universe with a hubble constant @xmath253 .
suppose instead we use the currently popular cosmological model in which @xmath254 of the mean energy density required to make space flat is in the form of material for which gravity acts repulsively and the remaining @xmath255 is carried by collisionless massive particles of some type .
then the physical size of the core radii of isothermal galaxies is @xmath256 pc .
these results are good for @xmath217 . from the top right panel of fig .
[ fig : wynfour ] , the dimensionless core radius can be increased to @xmath257 if the galaxy has a larger slope index ( say @xmath258 , which we will argue shortly is appropriate for giant ellipticals ) . in the same two cosmological models and with the same assumptions as to typical source and lens redshifts ,
this gives physical sizes of the core radius of @xmath259 pc and @xmath260 pc respectively , consistent with almost all the central images being absent .
most of the optical depth to strong lensing resides in giant elliptical galaxies .
faber et al .
( 1997 ) analyse _ hubble space telescope _
surface photometry of nearby ellipticals .
they provide convincing evidence that giant elliptical galaxies have cuspy cores with steep outer power - law profiles and shallow inner profiles separated by a break radius .
this is in contrast to low luminosity ellipticals , which have power - law surface brightness profiles .
there are 26 giant ellipticals with cuspy cores presented in table 2 of faber et al .
the mean value of the asymptotic outer slope of the surface brightness is @xmath261 with a standard deviation of @xmath262 .
this corresponds to the slope index or @xmath216 in the notation of this paper .
the mean break radius is @xmath263 pc , which corresponds to only a few tens of millarcseconds at a typical lens redshift of @xmath264 .
the bright images of strong lenses are therefore probing the steep outer part of the cuspy core profile .
the light profile is much steeper than isothermal ; however , the projected mass may be less steep depending on the distribution of dark matter .
central images are absent because break radii are small in cuspy core galaxies and so the steep slope continues to small radii .
in fact , only 2 out of the 26 galaxies listed by faber et al .
( 1997 ) have a sufficiently shallow profile for the central image to stand any chance of being visible at the @xmath265 threshold .
this provides an explanation of why most central images are unobservable at the current thresholds .
other than a central black hole , there is little evidence for dark matter inside the inner 10 kpc of early - type galaxies that can not be simply assigned to the stellar mass ( e.g. , gerhard et al .
2001 ) . beyond 10 kpc or so , there probably is dark matter , although there are few hard facts on its distribution in ellipticals because of the absence of tracers at large radii .
the dynamical evidence refers to the mass within spheres , whereas lensing is concerned with the mass within cylinders .
dark matter at large radii may alter the slope of the projected mass distribution in the inner few kpc .
however , provided the slope remains isothermal or steeper and the break radius remains small , then the central image remains unobservable .
[ fig : wynsix ] shows the percentage of triplets with a visible central image as a function of the threshold for four different slope indices .
notice that for models with slopes steeper than isothermal , flattening typically makes the central image less visible , the reverse of the case when the slope is weak .
the latter case shows the same trends as in the @xmath266 panel of fig .
[ fig : wynone ] .
the radial caustic is large when @xmath267 and the tangential caustic occupies a much smaller part of the space within it .
both @xmath199 and @xmath228 now decrease as @xmath1 decreases , but @xmath199 decreases faster to make the central image less visible .
[ fig : wynsix ] shows us that the fraction of detectable images is a strong function of the threshold .
it enables us to predict the threshold required to find the missing central images . using the third panel of fig .
[ fig : wynsix ] as typical of the outer parts of cuspy core profiles , we see that the threshold has to be @xmath268 for the central image to be detectable in half the triplet systems .
even the most sensitive class radio map ( for b0218 + 357 ) probes only to a threshold of @xmath269 , so this provides an explanation of why the central images have so far remained missing despite deeper searches .
we have shown how contour integration can ease the evaluation of the magnifications of images .
this work extends the ideas presented in hunter & evans ( 2001 ) in two significant ways .
first , our earlier analysis was restricted to scale - free power - law potentials .
we have now generalized it to cover all elliptically stratified potentials .
second , we have obtained separate formulas for sums of the two direct images , for sums of the two inverted images , and for the magnification of the central image .
previously , we found only formulas for sums over all four images weighted with the signed magnifications .
our detailed applications are to power - law galaxies with cores .
we have shown that the caustics are then always simple closed curves , and have given conditions for each of the four different caustic configurations .
we have found an approximation for the magnitude of the central image which applies throughout the region inside the large outer radial caustic when the core is small and the slope index @xmath270 . for small cores and weaker cusps , that approximation is directly useful only in a smaller inner region , but shows the magnitude of the central image to grow on a short length scale .
we find that , for power - law lenses with small cores and an inner tangential caustic , the sums over separate pairs vary considerably with the image positions while the signed sums over all four images are generally remarkably uniform . that near - uniformity is a consequence of large cancellations between terms which vary with the position of the source .
hence uncertainties in the magnification can have major effects on estimates of those four image sums .
this lessens their usefulness for the modeling of lenses as proposed by witt & mao ( 2000 ) and as used also by us in our paper i. we have shown that similar large cancellations can arise with any elliptically stratified potential , and not just power - laws .
hence , they will occur also with elliptically stratified densities in the limit of low eccentricity , and perhaps more generally , though this needs to be verified . as an application
, we have examined the constraints implied by the missing central images of triplet and quintuplet systems .
there is only one convincing example of a gravitational lens system with a central image , namely the triplet of the ultraluminous quasar apm08279 + 5255 .
however , there are a total of @xmath246 doublets known .
although this is a heterogeneous sample , discovered by different observers using different techniques in different wavebands , nonethless the probability of detecting central images does seem to be very low .
we take a central image as observable if the magnification ratio of the faintest to the brightest image is @xmath271 ( although some of the lens systems studied with the highest sensitivity by class do go deeper ) .
a rough summary of the observations is that the probability of observing a central image for a triplet is @xmath272 .
the absence of central images is understandable if the mass distribution in the lensing galaxy population is nearly cusped , and the cusp is isothermal or stronger . for typical source and lens redshifts
, the size of the core radius @xmath3 must be @xmath8 pc .
the slope of the gravitational potential @xmath4 must be @xmath7 . most of the optical depth to strong lensing resides in the most massive galaxies , namely giant ellipticals . we know from high resolution imaging of nearby giant ellipticals that they typically have cuspy cores with @xmath9 or @xmath273 outside the break radius of a few hundred parsecs ( faber et al .
the break radius corresponds to only a few tens of millarcseconds at a typical lens redshift .
hence , strong lensing is primarily probing the steep outer part of the cuspy core profile .
this is much steeper than isothermal , as the surface density is falling typically like @xmath274 .
the cuspy cores by themselves can provide the explanation of the missing central images .
dark matter at large radii may alter the slope of the projected mass distribution in the inner few kpc .
however , provided the slope remains isothermal or steeper and the break radius remains small , then the central image remains unobservable .
the ratio of the faintest to the brightest image in cuspy core profiles is typically @xmath268 .
even the most sensitive radio maps available probe only to a threshold of @xmath269 , so this explains why
the central images have so far remained missing despite deeper searches . the sensitivity of the searches must be increased by a factor of @xmath246 to find them .
nwe thanks the royal society for financial support .
the work of ch is supported in part by nsf through grants dms-9704615 and dms-0104751 .
we thank an industrious referee for a helpful report .
alcock c. et al .
1997 , apj , 486 , 697 biggs a.d . ,
browne i.w.a .
, helbig p. , koopmans l.v.e .
, wilkinson p.n .
, perley r.a .
1999 , mnras , 304 , 349 blandford r. , kochanek c.s . , 1987 , apj , 321 , 658 bourassa r.r . , kantowski r. , 1975 , apj , 195 , 13 bourassa r.r . , kantowski r. , norton t.d . , 1973 ,
apj , 185 , 747 chen g.h . , hewitt j.n . , 1993 ,
aj , 106 , 1719 evans n.w . , 1993 , mnras , 260 , 191 evans n.w . , 1994 , mnras , 267 , 333 evans n.w . ,
de zeeuw p.t . , 1994 , mnras , 271 , 202 evans n.w . , wilkinson m.i . , 1998 , mnras , 296 , 800 faber s.m
. , et al . 1997 , aj , 114 , 1771 gerhard o.e . ,
kronawitter a. , saglia r.p .
, bender r. 2001 , apj , 121 , 1936 henrici p. , 1974 , applied and computational complex analysis , volume 1 ( john wiley : new york ) hunter c. , evans n.w .
, 2001 , apj , 554 , 1227 ( paper i ) ibata r.a . ,
lewis g.f .
1998 , apj , 500 , 575 ibata r.a . ,
lewis g.f .
, irwin m.j . , leh ' ar j. , totten e.j . 1999 ,
aj , 118 , 1922 kassiola a. , kovner i. , 1993 , apj , 417 , 450 keeton c.r . , mao s. , witt h.j . , 2000 , apj , 537 , 697 mao s. , witt h.j . ,
koopmans l.v.e .
, 2001 , mnras , 323 , 301 milne - thomson l.m . , 1960 ,
the calculus of finite differences ( st . martin s press : new york ) myers s.t . , 1999 ,
sci . , 96 , 4236 narayan r. , blandford r. , nityananda r. 1984 , nature , 310 , 112 pospieszalka a. , et al .
1999 , in `` gravitational lensing : recent progress and future goals '' , eds t.g .
brainerd , c.s .
kochanek , asp conf .
, in press rhie s.h . , 1997 , apj , 484 , 63 rusin d. , ma c.p
. , 2001 , apj , 549 , l33 rusin d. , tegmark m. , 2001 , apj , 553 , 709 schneider p. , ehlers j. , falco e.e .
, 1992 , gravitational lenses ( springer - verlag , new york ) turner e.l . 1980 ,
apj , 242 , l135 turner e.l .
, ostriker j.p . , gott j.r .
1984 , apj , 284 , 1 van der marel r.p .
, evans n.w . , rix h .- w . , white s.d.m .
, de zeeuw p.t . , 1994 , mnras , 271 , 99 wallington s. , narayan r. , 1993 , apj , 403 , 517 witt h.j . , 1990 , a&a , 236 , 311 witt h.j . , 1996 , apj , 472 , l1 witt h.j . , mao s. , 1997 , mnras , 291 , 211 witt h.j . , mao s. , 2000 , mnras , 311 , 689 xanthopoulos e. , et al . 1998 ,
mnras , 300 , 649 young p. , gunn j.e . , kristian j. , oke j.b . ,
westphal j.a . , apj , 241 , 507
in this appendix , we show that the power - law galaxies with cores in the presence of external shear give rise to either one , three or five images , depending on the position of the source and the extent of the core radius .
a point - like source is never lensed into more than five images .
the caustics are simple closed curves and there is exactly one point on a caustic in each radial direction from the center of the lens .
the models differ from elliptically stratified density distributions . for these ,
witt & mao ( 2000 ) and keeton , mao & witt ( 2000 ) showed that external shear can cause butterfly and swallowtail cusps to develop on the caustics and so sextuple imaging and higher does occur .
the caustics are curves in the source plane on which the imaging equation ( [ eq : newimaging ] ) has double roots .
the partial derivative of the imaging equation with respect to @xmath65 vanishes at a double root and hence we find caustics from the common solutions of the two equations @xmath275=-k_1(t ) , \\ \label{eq : caustictwo } bt^{b-1}[l(t)-a^2k_1(t)]+t^b[l^{\prime}(t)-a^2k_1^{\prime}(t ) ] = -k_1^{\prime}(t).\end{aligned}\ ] ] here , we have introduced the functions @xmath276[(p_0-t)\zetab -\qb \zeta].\ ] ] dividing the two sides of equations ( [ eq : causticone ] ) and ( [ eq : caustictwo ] ) gives a sextic polynomial in @xmath65 , whose coefficients involve @xmath32 . alternatively , we can solve for @xmath277 and @xmath278 and divide the results to obtain a non - polynomial equation , but one which is independent of @xmath117 and depends only on the angular argument @xmath279 of @xmath32 .
this equation is @xmath280 , \nonumber \\ & = & ( 1-a^2t^b)\left [ { -2p^3\!+\!6p^2|q|\cos\theta\!-\!6p|q|^2\!+\!2|q|^3\cos\theta \over ( p^2\!-\!|q|^2)(p^2\!+\!|q|^2\!-\!2p|q|\cos\theta)}\right].\end{aligned}\ ] ] as @xmath281 , it reduces to equation ( a2 ) of paper i. it contains the complex angle @xmath282 defined by @xmath283 rather than @xmath279 .
we find caustics by searching for roots for @xmath65 along each angular direction in the source plane .
once @xmath65 is known , @xmath117 follows from equation ( [ eq : causticone ] ) . in general , this search must be done numerically , but one can deduce from the graph of the term in square brackets ( which has vertical asymptotes at @xmath87 and @xmath71 and is plotted in figure 7 of paper i ) and the extra @xmath103 factor , that there will always be a root in @xmath80 for a point on the tangential caustic if @xmath284 , and another in @xmath81 for a point on the radial caustic if @xmath89 .
there are four special directions for which points on the caustics can be found explicitly .
they are directions in which @xmath285 is real , and we now consider them in turn .
when @xmath286 and @xmath287 , then @xmath86 at which @xmath288 is a triple root of equation ( [ eq : doubleroottwo ] ) .
it gives a cusp at @xmath289 on the tangential caustic provided @xmath284 .
the other solution of equation ( [ eq : doubleroottwo ] ) is the root of @xmath290 in @xmath81 for @xmath89 , which gives a point on the radial caustic .
the @xmath86 cusps on the tangential caustic are naked if their value of @xmath289 exceeds the value of @xmath117 for the fourth root for @xmath65 of equation ( [ eq : doubleroottwo ] ) .
that value is given by equation ( [ eq : causticone ] ) evaluated for that root .
the root can not be found explicitly when @xmath291 . hence the marginal cases shown in fig .
[ fig : christhree ] are found numerically by eliminating @xmath292 between the equations @xmath293 and solving for @xmath92 for given @xmath91 and @xmath294 .
the second of equations ( [ eq : nakedness ] ) comes from equating the two values for @xmath295 .
we can solve for @xmath65 explicitly in the coreless @xmath95 case , to give the condition @xmath296^{(b-2)/2},\ ] ] for naked cusps .
the other special directions are those for which @xmath297 and @xmath298
. then @xmath87 at which @xmath299 is a triple root of ( [ eq : doubleroottwo ] ) , and gives a cusp point on a caustic at @xmath300 provided @xmath89 .
the other solution of equation ( [ eq : doubleroottwo ] ) satisfies @xmath301 .
such a solution exists provided @xmath302 .
it gives a point on the radial or tangential caustic according to whether it is greater or less than @xmath70 .
the @xmath87 cusp is on the tangential caustic in the first case , and on the radial caustic in the second , giving condition ( [ eq : twolips ] ) as that for which the double lips configuration occurs .
double lips occur because the @xmath87 cusp always lies on whichever is the inner caustic in this special direction ; its @xmath117 value never exceeds that for the other root of equation ( [ eq : doubleroottwo ] ) .
the two @xmath65 and @xmath117 values are equal only when @xmath87 is a quadruple root of equation ( [ eq : doubleroottwo ] ) and then @xmath303 .
this latter case is the transitional one in which the two caustics share a cusp and coincide locally , and marks the stage at which the region for core triplets has shrunk to zero .
a simple analytical approximation for the radial caustic can be found for @xmath3 small and @xmath85 .
this caustic , which does not exist for @xmath95 , is large .
it is found by looking for large roots of equations ( [ eq : causticone ] ) and ( [ eq : caustictwo ] ) for @xmath65 , and approximating @xmath304 by @xmath305 and @xmath306 by @xmath307 . eliminating @xmath65
then gives the approximation @xmath308
^{(2-b)/b}[1+o(a^{2/b})].\ ] ] hence the radial caustic is approximately an ellipse elongated in the @xmath309-direction , as in figs . [
fig : christwo]b and d. the reason why those two ellipses for @xmath209 are not well described by equation ( [ eq : radcausticsmalla ] ) is that its limit of @xmath310 is not accurate in the marginal case of @xmath139 until @xmath311 .
a more accurate formula for the radial caustic for @xmath139 is @xmath312^{2/3}+o(a^{4/3}).\ ] ] we showed in paper i , by an analysis of the quartic obtained when @xmath95 so there is no complicating @xmath313 power in equation ( [ eq : doubleroottwo ] ) , that there is never more than one root in @xmath80 and one in @xmath314 .
the additional @xmath103 factor now present has only a small effect at finite @xmath65 when @xmath3 is small .
however , it does ensure that equation ( [ eq : doubleroottwo ] ) always has a root in @xmath81 when this interval exists , which is not the case when @xmath95 if @xmath201 .
if more roots are to occur at larger values of @xmath3 , there must be transitional cases at which the pair of equations ( [ eq : causticone ] ) and ( [ eq : caustictwo ] ) have a multiple root .
we show next , using the sextic derived from those equations by eliminating the @xmath313 terms , that this can not be , and hence that the cases listed in the previous paragraph are the only ones possible . to study the roots for @xmath65 in @xmath315 , we work with the variable @xmath316 , for which the sextic is : @xmath317\sigma^3 + 2(1+\cos\theta)[3(b - p_0 - 3)\sigma^2\\ & & + 2(b-3p_0 - 5)\sigma-4(1+p_0)]=0,\nonumber\end{aligned}\ ] ] where @xmath318 .
we let @xmath319 denote the coefficient of @xmath320 and apply descartes rule of signs ( henrici 1974 ) which shows that there is just one positive root for @xmath321 when there is a single sign change in the coefficient sequence @xmath322 .
we take @xmath323 because the equations simplify in the special cases for which @xmath324 as discussed earlier and look for possible sign changes in the coefficient sequence . for @xmath207 , @xmath325 , @xmath326 , and @xmath327
are positive and @xmath328 is negative .
also @xmath329 as in paper i , and @xmath330 if @xmath331 . for @xmath139 , @xmath325 , @xmath326 , and @xmath327 are positive , while @xmath126 , @xmath133 , and @xmath328 are negative . for @xmath267 , @xmath325 and @xmath326
are positive , @xmath126 , @xmath133 , and @xmath328 are negative , and @xmath332 . in each case , only one sign change can and does occur .
to show that there is just a single root in @xmath333 , we re - express the sextic in the variable @xmath334 which runs from 0 to @xmath335 .
it is then @xmath336 where @xmath337 , @xmath338 , @xmath339 , and @xmath340 are the coefficients of equation ( a4 ) of paper i. labelling this sextic as @xmath341 , one finds that @xmath325 , @xmath326 , and @xmath327 are always positive and @xmath328 always negative . for @xmath207 , @xmath329 and @xmath330 if @xmath331 . for @xmath85 , @xmath126 and @xmath133
are both negative . in either case
, there is exactly one sign change in the coefficient sequence , and therefore exactly one root for @xmath65 in @xmath80 .
in this appendix , we show how to evaluate the coefficients @xmath136 and @xmath137 of the powers @xmath342 introduced in section 3 . they can always be expressed as finite sums .
we consider first the scale - free case for which we give compact explicit expressions , and then the cored case for which we derive recursive relations .
we give explicit expressions for sums of coefficients @xmath145 needed for four - image sums in the scale - free case .
then we derive a compact expression for the @xmath145 sums valid for any elliptically stratified potential , and discuss its consequences . compact representations for the @xmath157 and @xmath146 terms can be derived for the coreless @xmath95 case by partial differentiation using leibniz s rule .
we find : @xmath343\nonumber\\ & = & \sum\limits^{\ell}_{m=0}{jb-1\choose m}{t_1^{bj-1-m}\over(\ell - m ) ! } { ( -1)^{\ell - m-1}(\ell+k - m)!\over k!(t_1-t_2)^{\ell+k+1-m}}.\end{aligned}\ ] ] this sum of @xmath344 terms can be written as the terminating hypergeometric series @xmath345 with the proviso that , when @xmath346 and the hypergeometric series becomes a geometric one , only the first @xmath344 terms of that series are to be used .
a similar expression for @xmath146 can be found from it by using the transformation relation in eq ( [ eq : ronederiv ] ) .
it is @xmath347 in obtaining separate formulae for the two pairs of bright images rather than for all four , the present work extends that of paper i for the coreless @xmath95 case .
we showed there that the coefficients for the sums of four images can all be expressed in terms of hypergeometric functions ; that is we showed that @xmath348 we also showed in appendix c of paper i that the infinite hypergeometric series ( [ eq : hypersum ] ) can be represented as the sum of the two finite components ( [ eq : ronehyper ] ) and ( [ eq : rtwohyper ] ) .
we did not appreciate the significance of the two separate components as two - image sums .
we did warn of the tendency of the separate @xmath157 and @xmath146 components to cancel for small @xmath349 , and that their sum could be computed more easily using the rapidly convergent infinite series ( [ eq : hypersum ] ) .
a simple instance of this is the @xmath350 case for which @xmath351 the sum here for @xmath146 consists of the first @xmath352 terms in the infinite binomial expansion of @xmath353 in powers of @xmath354 ; and hence the first @xmath352 terms in the expansion of @xmath355 .
another instance occurs when @xmath138 or @xmath139 .
then @xmath356 for @xmath357 with the result that the sum of the four signed magnifications are then independent of the position of the source ( witt & mao 2000 ; see also paper i ) .
these examples are simple instances of the large cancellations which can occur when contributions of the direct and the inverted image pairs , weighted with the signed magnifications , are combined .
the idea is to use leibniz s rule to evaluate @xmath358.\ ] ] the general derivative of the second factor is of the form @xmath359 = { t^{bj - m-1}_1\over(1-a^2t^b_1)^j}\sum\limits^m_{n=0 } \alpha_{m , n } \left[{a^2t^b_1\over(1-a^2t^b_1)}\right]^n,\ ] ] where the coefficients @xmath360 , @xmath361 , can be found iteratively from the relation @xmath362\alpha_{m-1,n } + b(j+n-1)\alpha_{m-1,n-1},\qquad \alpha_{0,0}=1.\ ] ] use of leibniz s rule then gives @xmath363^n.\nonumber\end{aligned}\ ] ] in an analogous manner , we find ; @xmath364^n.\nonumber\end{aligned}\ ] ] the only @xmath365 s for which it is simple to derive explicit expressions are @xmath366 the first set are the only ones that appear in the @xmath95 case when the double sums ( [ eq : roneasum ] ) and ( [ eq : rtwoasum ] ) reduce to the hypergeometric sums ( [ eq : ronehyper ] ) and ( [ eq : rtwohyper ] ) respectively .
we begin by rewriting equation ( [ eq : rtwodefn ] ) for @xmath146 as @xmath367.\ ] ] the choice @xmath368 is needed for equation ( [ eq : rtwodefn ] ) , but a similar equation with a different @xmath369 will arise for some other elliptically stratified potential @xmath370 with @xmath371 .
we again choose @xmath106 to be a contour which encloses @xmath86 but not @xmath65 or any singularity of @xmath369 .
then a residue calculation gives @xmath372 there is a similar relation @xmath373,\ ] ] for the @xmath157 coefficient .
when we evaluate its integral by residues , and combine the results , we obtain @xmath374 = -g[t_1,\ldots t_1,t_2,\ldots t_2].\ ] ] here @xmath375=\frac{g(t_1)-g(t_2)}{t_1-t_2},\ ] ] is a simple divided difference , while @xmath376 $ ] has @xmath70 repeated @xmath344 times and @xmath71 repeated @xmath352 times , and is a divided difference of order @xmath377 ( milne - thomson 1960 , chap . 1 ) .
the significance of formula ( [ eq : rgensum ] ) is as follows .
divided differences are well - behaved when @xmath378 becomes small .
more specifically , a mean value theorem shows that @xmath379 = -\frac{g^{(\ell+k+1)}(\bar{t})}{(\ell+k+1)!},\ ] ] where @xmath380 , at which the @xmath377th derivative of @xmath369 is evaluated , is some value of @xmath65 in the interval @xmath80 .
thus , when the axis ratio @xmath1 is close to 1 so that @xmath378 is small , while the function @xmath369 varies smoothly in @xmath80 as in section b.1 , then both @xmath157 and @xmath146 become very large when @xmath378 is small because both contain negative powers up to and including @xmath381 , while their sum varies little .
but this is not the case with the lensing potential of fig .
[ fig : christwo]a
. then @xmath3 is large and the negative @xmath382 powers cause the derivatives of @xmath369 to be large , so that @xmath145 is then also large . | a method is developed to evaluate the magnifications of the images of galaxies with lensing potentials stratified on similar concentric ellipses . in a quadruplet system , there are two even parity images , two odd parity images , together with a de - magnified and usually missing central image .
a simple contour integral is provided which enables the sums of the magnifications of the even parity or the odd parity images or the central image to be separately calculated without explicit solution of the lens equation .
we find that the sums for pairs of images generally vary considerably with the position of the source , while the signed sums of the two pairs can be remarkably uniform inside the tangential caustic in the absence of naked cusps . for a family of models
in which the lensing potential is a power - law of the elliptic radius , @xmath0 , the number of visible images is found as a function of flattening @xmath1 , external shear @xmath2 and core radius @xmath3 .
the magnification of the central image depends on the size of the core and the slope @xmath4 of the gravitational potential .
it grows strongly with the source offset if @xmath5 , but weakly if @xmath6 . for typical source and lens redshifts
, the missing central image leads to strong constraints ; the slope @xmath4 must be @xmath7 and the core radius @xmath3 must be @xmath8 pc .
the mass distribution in the lensing galaxy population must be nearly cusped , and the cusp must be isothermal or stronger .
this is in good accord with the cuspy cores seen in high resolution photometry of nearby , massive , early - type galaxies , which typically have @xmath9 ( or surface density falling like distance@xmath10 ) outside a break radius of a few hundred parsecs .
cuspy cores by themselves can provide the explanation of the missing central images .
dark matter at large radii may alter the slope of the projected density ; however , provided the slope remains isothermal or steeper and the break radius remains small , then the central image remains unobservable .
the sensitivity of the radio maps must be increased fifty - fold to find the central images in abundance . |
the idea that our world is a three brane embedded in a higher dimensional space - time with non - factorizable warped geometry has been much investigated since the appearance of papers @xcite .
the first solution for a warped metric was provided in @xcite . in this paper
it was suggested that gravitational interactions between particles on a brane in uncompactified five - dimensional space could have the correct four - dimensional newtonian behavior , provided that the bulk cosmological constant is fine - tuned to the brane tension .
as it was shown in @xcite the warped metrics could provide an alternative to compactification of extra dimension used for a possible explanation of the hierarchy problem @xcite .
brane solution described in @xcite has the form @xmath0 where the parameter @xmath1 is determined by the bulk cosmological and five - dimensional gravitational constants respectively .
an important issue in the study of braneworld models is the localization of standard model fields on the brane .
this problem for the brane ( [ grs ] ) was addressed in @xcite .
it was shown that the massless scalar field has a zero mode normalizable with the weight @xmath2 .
when this field is given mass the bound state becomes metastable against tunneling into an extra dimension @xcite .
fermion fields are not localized on the positive tension brane by gravitational interactions @xcite . in order to localize fermions
one can use the method of localization described in paper @xcite .
but the adding of bulk mass term makes the fermion bound state unstable as well trough the barrier penetration into an extra dimension @xcite .
based on this phenomenon the possibility of decaying cold dark matter into an extra space was considered in the framework of braneworld cosmology , see @xcite .
the comoving density of cold dark matter was taken to decay over time with a rate @xmath3 corresponding to the exponential decay of a metastable state .
but in quantum mechanics it is well established that an exponential decay can not last forever if the hamiltonian is bounded below and can not occur for small times if , besides that , the energy expectation value of the initial state is finite @xcite . from this point of view
we consider the dynamics of massive matter disappearance on the brane .
due to four - dimensional poincare invariance of the metric ( [ grs ] ) every fields in this background can be decomposed into four - dimensional plane waves @xmath4 where coordinate four momentum @xmath5 coincides with the physical momentum on the brane .
it is natural to assume that the three momentum @xmath6 does not change the picture of tunneling into an extra dimension . for the sake of simplicity we put @xmath7 in what follows . under this assumption
the equation @xmath8 gives @xmath9 where @xmath10 .
the continuum spectrum of eq.([eigeq ] ) starts from @xmath11 @xmath12~,\ ] ] where @xmath13 @xcite .
the normalization condition @xmath14 as well as the boundary condition @xmath15 can be satisfied by taking @xcite @xmath16 @xmath17 the general solution to eq.([timequat ] ) can be written as @xcite @xmath18 where @xmath19 to compute @xmath20 one needs to know the initial data @xmath21 . in what follows
we take the following condition @xmath22 corresponding to the free particle on the brane ( [ plwde ] ) with energy @xmath23 .
correspondingly the eq.([gensol ] ) takes the form @xmath24 \phi(0,~z')dz'~.\ ] ] at @xmath25 the particle is confined on the brane .
the probability of the particle to remain on the brain at instant @xmath26 is given by @xmath27 where @xmath28 denotes the brane localized initial state and the scalar product is understood with the measure @xmath2 @xmath29 from eq.([eveq ] ) one finds @xmath30~,\ ] ] @xmath31~,\ ] ] where @xmath32 in the above @xmath33 is real variable , but in the following it is understood that @xmath33 can be complex . the function @xmath34 has the simple poles determined by the equation @xmath35 the equation @xmath36 under assumption @xmath37 gives the pole in the fourth quadrant of a complex @xmath33 plane , @xmath38 , where @xmath39 since , when @xmath40 and @xmath41 , @xmath42 in the fourth quadrant , for evaluating of integrals entering eqs.([realpart ] , [ imaginpart ] ) one can deform the integration counter as it is shown in fig.1 .
in this way one finds @xmath43 from this equation one sees that the second term gives an imaginary contribution and correspondingly @xmath44 is determined by the residue term only . in the same way one finds that @xmath45 is also determined only by the residue term .
following to the quantum mechanical formalism , and by taking into account that @xmath46 and the localization width of the scalar is @xmath47 , one can simply take @xmath48 thus , the probability of finding the particle on the brane equals one when @xmath25 . by evaluating the residue term and using eq.([initstate ] ) one finds @xmath49
the minimal representation of spinors in five dimensions can be chosen to be four dimensional .
the five - dimensional minkowskian gamma matrices can be chosen as follows @xmath50 where @xmath51 with @xmath52 the three pauli matrices .
the dirac equation in the background @xmath53 for the fermion coupled with the domain wall @xmath54 reads @xmath55\psi=0~.\ ] ] solutions to this equation are linear combination of wave functions of the form @xmath56 we do not expect the three momentum @xmath6 to affect the overall picture of particle tunneling into an extra space . therefore , as in the case of scalar field , we assume @xmath57 under this assumption the eq.([fermeq ] ) takes the form @xmath58 where @xmath59~.\ ] ] so that the time evolution of initially brane localized fermion @xmath60 is given by @xmath61 where the scalar product is understood with the measure @xmath62 and @xmath63 in terms of the right- and left - handed components @xmath64 the eq.([fermeigeneq ] ) takes the form @xmath65 @xmath66 where @xmath67 . after eliminating @xmath68 from eqs.([chiralll ] , [ chirallr ] )
one obtains a second order equation for @xmath69 @xmath70\chi_l=0~.\end{aligned}\ ] ] for the domain wall profile in the thin wall limit one can take @xmath71 to the right ( @xmath72 ) and left ( @xmath73 ) of the brane one gets @xmath74\chi_l=0~,\nonumber\\ \left[\partial_z^2 + 5k\partial_z - gkv+6k^2-g^2v^2+e^2e^{-2kz}\right]\chi_l=0~.\end{aligned}\ ] ] the solution of eq.([leftrightz ] ) reads @xmath75~,\ ] ] where @xmath76 . from eqs.([chiralll],[chirallr ] ) one finds the following boundary condition @xmath77 where the prime stands for the derivative with respect to @xmath78 . to satisfy this boundary condition as well as the normalization condition @xmath79 the coefficients @xmath80 should have the form @xcite @xmath81@xmath82 the solution to eq.([leftrightz ] ) for @xmath83 @xmath84
is localized on the brane as long as @xmath85 with the localization width @xmath86 @xcite . under assumption that the initial state is given by the left - handed particle localized on the brane one finds @xmath87
the function @xmath88 may have the poles @xmath89 in the fourth quadrant of complex @xmath33 plane determined by the zeros of @xmath90 indicating the presence of resonances in the spectrum of kk modes .
for evaluating the integral ( [ leamlitu ] ) one can deform the integration contour as it is shown in fig.1 . in this way one gets
@xmath91 where @xmath92 so that the probability to find the particle on the brane at time @xmath26 is given by @xmath93 one can easily check that @xmath94 namely , by using eqs.([fereqimzer],[ferhamimzer ] ) and taking into account that @xmath95 is an even function of @xmath96 one sees that in this expression the odd function is integrated over a symmetric region .
so that the evolution of the decay process requires some time to reach the regime of exponential decay . for large values of time @xmath97 the second term in eq.([fermintdefor ] ) becomes dominant for which in this limit only modes with @xmath98 are relevant .
therefore , the non - escape probability behaves asymptotically as @xmath99 thus , after a long time the decay proceeds with respect to the power law .
as it is shown in @xcite one can choose the index @xmath100 in such a way ( @xmath101 ) the hankel function @xmath102 to have exactly one zero in the fourth quadrant of complex @xmath103 plane .
this is the only zero of the product @xmath104 located in the fourth quadrant since for real values of @xmath100 , @xmath105 . from the trajectories of zeros presented in @xcite one
sees that if @xmath100 is close to @xmath106 ( @xmath107 ) the function @xmath108 has the zero @xmath38 in the fourth quadrant such that @xmath109 and @xmath110 .
( [ zeroenfersol ] ) for the initial state of the low lying kk resonances one can simply take @xmath111 so that at @xmath25 the particle is known to be on the brane , @xmath112 , with probability one . using eq.([ferinitstate ] ) and the previous consideration one can perform the concrete calculations for different low lying kk resonances .
from the present consideration one sees that if the transverse equation contains the second order time derivative the corresponding decay law has the exponential form as long as @xmath113 .
so that under this assumption the decay of massive initially brane localized metastable modes for scalar , vector and gravitational fields follows to the exponential law .
it should be stressed that the choice @xmath22 which is essential to this result is not unique .
however we find this condition to be natural as it corresponds to the free particle on the brane ( [ plwde ] ) with energy @xmath23 corresponding to the metastable state .
in contrast , the decay of massive quasilocalized fermion modes does not follow completely to the exponential law . as in the standard quantum mechanical case mentioned in the introduction , initially the decay is slower than exponential , then comes the exponential region and after a long time it obeys a power law .
the present approach allows one to evaluate the preexponential term as well as the whole dynamics of the decay and can be important to evaluate dynamics for any process based on this phenomenon , for instance to examine more precisely the cosmological constraints on the paradigm of decaying dark matter in the braneworld model .
n. arkani - hamed , s. dimopoulos and g. dvali , phys . lett . * b429 * ( 1998 ) 263 , hep - ph/9803315 ; n. arkani - hamed , s. dimopoulos and g. dvali , phys . rev .
* d59 * ( 1999 ) 086004 , hep - ph/9807344 ; i. antoniadis , n. arkani - hamed , s. dimopoulos and g. dvali , phys . lett .
* b436 * ( 1998 ) 257 , hep - ph/9804398 . | time evolution of the probability of quasilocalized massive modes decay trough the tunneling into an extra dimension is analyzed in the randall - sundrum braneworld model . |
since the first exact solution of einstein equation was found out , studying properties of black holes has become an important part of gravitational physics .
black hole physics causes deep , unsuspected connections among classical general relativity , quantum physics and statistical mechanics .
however , the classical definition of a black hole @xcite , which formalizes the notion of spacetime region from which nothing can ever escape , is too global : it requires knowledge of the entire future of the spacetime , and can not satisfy the requirement for practical research @xcite , so alternate notions of black hole @xcite were devised which possess a quasi - local description formalism . in recent years , a new , quasi - local framework was introduced by ashtekar and his collaborators to analyze different facets of black holes in a unified way @xcite . compared with the event horizon
, this framework does nt need the knowledge of overall spacetime , and only involves quasi - local conditions , so it accords with the practical physical process . in this framework
, black holes in equilibrium are described by ( weakly ) isolated horizons ( wih ) .
this paradigm leads to significant generalization of several results in black hole physics and obtains considerable success . in 1978
, damour found that instead of the conventional view of a black hole as simply an empty region in spacetime , the black hole horizon could be viewed as a physical membrane endowed with specified mechanical and electromagnetic properties ; for example , the membrane of kerr black hole behaves as a metallic shell whose surface resistivity is @xmath1 @xcite .
such idea was investigated by other researchers later and was called membrane paradigm of black hole " @xcite .
this paradigm was reconsidered in the framework of ads / cft correspondence more than twenty years later .
there are similar properties between the fluid / gravity duality in the ads / cft framework and black - hole membrane @xcite and this kind of universality was interpreted as the wilson renormalization group flow in the ads / cft framework @xcite .
more developments in this area can be found in refs .
this paper is concerned with the electrical properties of isolated horizons .
the electrical laws of kerr black hole @xcite are only related to the quantities on the horizon , so it is natural to think that the laws are suitable for isolated horizon . hawking radiation of weakly isolated horizon
was investigated recently @xcite , which confirms one aspect of thermodynamic nature of isolated horizon .
considering the importance of carnot cycle near a black hole to black hole thermodynamics @xcite , we design a reversible carnot cycle outside a non - rotating isolated horizon to give a further confirmation of the thermodynamic nature of isolated horizon . the organization of this paper is as follows . in section @xmath2 ,
we briefly review the definition and near - horizon geometry of ( weakly ) isolated horizon . in section @xmath3
, we establish the electrical laws of an isolated horizon by following damour s method . in section @xmath4
, we investigate the properties near a non - rotating isolated horizon under the first - order approximation of @xmath0 , and construct a reversible carnot cycle near the horizon to confirm the thermodynamical nature of isolated horizon .
finally , we make some discussions and conclusions .
in this section we briefly review the definition and geometric properties of ( weakly ) isolated horizon . according to the works by ashtekar and his collaborators @xcite , weakly isolated horizon ( wih ) is defined by + * definition * let @xmath5 be a spacetime .
@xmath6 is a 3-dim null hyper - surface in @xmath7 and @xmath8 is the tangent vector field of the generator of @xmath6 .
@xmath6 is said to be a * weakly isolated horizon * ( wih ) , if + 1)@xmath6 has the topology of @xmath9 ; + 2)the expansion of the null generator of @xmath6 is zero , i.e. @xmath10 on @xmath6 ; + 3)@xmath11 is future causal for any future causal vector @xmath12 and einstein equation holds in the neighborhood of @xmath6 ; + 4)@xmath13l^{b}=0 $ ] on @xmath6 , where @xmath14 is the induced covariant derivative on @xmath6 .
a weakly isolated horizon is said to constitute an * isolated horizon * ( ih ) if @xmath13=0 $ ] on @xmath6 @xcite . from the definition ,
isolated horizon is the special case of weakly isolated horizon , so ih has all the properties of wih .
it is convenient to introduce bondi - like coordinates @xmath15 in the neighborhood of the horizon @xmath6 in the following way @xcite .
first , denote the tangent vector of null generator of @xmath6 as @xmath8 and another real null vector field as @xmath16 .
the foliation of @xmath6 gives us the natural coordinates @xmath17 .
lie dragging @xmath18 along each generator of @xmath6 together with the parameter @xmath19 of @xmath8 forms the coordinates @xmath20 on @xmath6 .
finally , choose the affine parameter @xmath0 of @xmath16 as the fourth coordinate , then we obtain the bondi - like coordinates @xmath21 near the horizon . with the bondi - like coordinate system in hand
, we construct the null tetrad as @xmath22 where @xmath23 on @xmath6 ( following the notation in ref .
@xcite , equalities restricted to @xmath6 are denoted by @xmath24 " ) , and @xmath25 .
note that @xmath16 and @xmath8 are future directed .
we take the spacetime metric @xmath26 to have a signature @xmath27 , so the metric can be expressed as @xmath28 the matrix form of the metric is @xmath29 and , on the horizon , the metric reduces to @xmath30 where @xmath31 means the value of @xmath32 on the horizon .
the dual basis of the tetrad ( [ biaojia ] ) can be calculated as @xmath33 the commutators@xmath34 $ ] and @xmath35 $ ] tell us that @xmath36 in the newman - penrose formalism , we can require the following gauge conditions : @xmath37 which mean that the tetrad vectors ( [ biaojia ] ) are parallelly transported along @xmath38 in spacetime .
the fourth requirement in the definition of wih implies that there exists a one form @xmath39 on @xmath6 such that @xmath40 and @xmath41 . in terms of the newman - penrose formalism
, @xmath39 can be expressed as @xmath42 where @xmath43 is constant on @xmath6 @xcite .
the definition of wih also implies @xmath44 based on ref .
@xcite , not any choice of time direction can give a hamiltonian evolution , and only some suitably chosen time direction can lead to a well - defined horizon mass . in ref .
@xcite , a. ashtekar and b. krishnan gave a canonical way to choose the time direction @xmath45 for a wih , and the restriction of @xmath45 to @xmath6 should be a linear combination of the null normal @xmath46 and the axisymmetric vector @xmath47 , @xmath48 where @xmath49 and @xmath50 are constant on the horizon . compared with the schwarzschild case , the parameter of @xmath45 takes the place of the killing time . with the canonical time direction @xmath45 , ref .
@xcite established the zeroth and the first law of wih . by definition ,
the surface gravity of @xmath6 is @xmath51 , which is constant on @xmath6 , so the zeroth law of black hole mechanics is valid for wih .
the first law is expressed as @xmath52 where @xmath53 is the horizon mass , @xmath54 is the area of the cross section of wih , @xmath50 is the angular velocity of the horizon , and @xmath55 is the angular momentum .
the first law of wih is the generalization of the first law of stationary black hole . since hawking radiation of wih
was investigated in refs .
@xcite , the laws of wih mechanics were upgraded to the laws of wih thermodynamics .
in this section , we investigate the electrical laws of an isolated horizon following damour s method @xcite .
let us choose a set of real bondi - like coordinates @xmath15 . according to eq .
( [ 1 ] ) , and the fact that the topological structure of an isolated horizons is @xmath9 , the metric on the horizon can be also expressed as @xmath56 the physical observer should be the canonical time @xmath57 . according to the relationship : @xmath58 we introduce a new set of coordinates @xmath59 .
we have substituted @xmath60 for @xmath61 for simplicity .
the relationship between bondi - like coordinates @xmath15 and the new ones on the horizon is @xmath62 so we have @xmath63 the metric ( [ rn3 ] ) in the new coordinates can be expressed as @xmath64 then the intrinsic geometry of cross section @xmath65const and @xmath66 is @xmath67 the basis of the cross section is @xmath68 and the corresponding dual basis is @xmath69 given an electromagnetic test field @xmath70 regular on the horizon @xmath6 we define the tangential electric field and the normal magnetic induction by restricting the form @xmath71 to the horizon @xmath66 .
namely , @xmath72\wedge dt+f_{\widetilde{\theta}\widetilde{\varphi}}d\widetilde{\theta}\wedge d\widetilde{\varphi}\nonumber\\ & = & ( e_{(\widetilde{\theta})}\omega^{(\widetilde{\theta})}+e_{(\widetilde{\varphi})}\omega^{(\widetilde{\varphi})})\wedge dt+b_{\perp}\omega^{(\widetilde{\theta})}\wedge\omega^{(\widetilde{\varphi})},\end{aligned}\ ] ] where @xmath73 by following damour s method @xcite , we can endow charges and currents on the horizon to keep the conservation of four - current outside the black hole since we do not want to consider what happens inside the black hole .
there exists a four - current @xmath74 which is defined and conserved all over spacetime , @xmath75 .
we find a complementary current @xmath76 with support on @xmath77 such that @xmath78 is conserved , where @xmath79 is the heaviside function . by replacing @xmath80 with @xmath81 in the equation @xmath82
, we get the conserved current @xmath78 where @xmath83 .
next we calculate the dirac distribution @xmath84 on the horizon . for the spacetime ,
the volume element is @xmath85 and the volume element of the cross section is @xmath86 from the equation , @xmath87 we easily find @xmath88 now we can write the complementary current @xmath76 , with support on the horizon , as @xmath89 with @xmath90 which is the surface four - current density and can be decomposed into a surface charge density @xmath91 and the geometrical components of a surface current density @xmath92 , @xmath93 it s not hard to find the following two relationships , @xmath94 @xmath95 where @xmath96 can be interpreted as the velocity of the horizon . from eqs .
( [ bj1 ] , [ bj2 ] ) , we have @xmath97.\end{aligned}\ ] ] in vector form , @xmath98 which is the ohm s law of isolated horizon , and the surface resistivity is @xmath99 . for kerr black holes , @xmath100
, we have @xmath101 , so the surface resistivity is @xmath1 , which returns to damour s result @xcite .
it is useful to introduce the notions of the surface conduction current @xmath102(the total current @xmath92 minus the convection current @xmath103 ) and thedragged - along " electric field @xmath104 , @xmath105 so the ohm s law ( [ 71 ] ) can be rewritten as @xmath106 next we use the surface conduction current to express the joule s law .
we can define the heat@xmath107 dissipated in the hole as @xcite @xmath108 where @xmath60 is the angular velocity and @xmath109 and @xmath110 are the increases in , respectively , area , mass , and angular momentum of ih .
the total energy flux into the hole is given by an integral on the horizon @xcite , @xmath111 where @xmath112 is the test energy - momentum tensor at the horizon .
the angular momentum flux is @xmath113 so we get the heat production as @xmath114 in the case of an electromagnetic field we have on the horizon , @xmath115 and from eqs .
( [ bj1 ] , [ bj2 ] ) , we have @xmath116 by putting eqs .
( [ bj3 ] ) into eq .
( [ bj4 ] ) , we obtain @xmath117\nonumber\\ & = & \frac{1}{b_{t}}\big[k^{(\widetilde{\theta})}(e_{(\widetilde{\theta})}+vb_{\bot})+e_{(\widetilde{\varphi})}(k^{(\widetilde{\varphi})}-\sigma v)\big]\nonumber\\ & = & \frac{1}{b_{t}}(\overrightarrow{e}+\overrightarrow{v}\times\overrightarrow{b}_{\bot})\cdot(\overrightarrow{k}-\sigma\overrightarrow{v})\nonumber\\ & = & \frac{1}{b_{t}}\overrightarrow{e}^{*}\cdot \overrightarrow{c},\end{aligned}\ ] ] where we have used eqs .
( [ rn4 ] ) in the last equality .
hence we get the joule s law , @xmath118 this is the joule s law of isolated horizon , and the surface resistivity is also @xmath119 . from the two laws we see that ih can be thought as a metal shell whose surface resistivity is @xmath120 .
we do not use all the conditions of isolated horizon , so the results are very general .
in this section , we firstly investigate some properties near a non - rotating ih , and find that under the first - order approximation of @xmath0 , @xmath121 is a killing vector and there exists a hamiltonian conjugate to it , so @xmath122 is a physical observer .
then we construct a reversible carnot cycle with the isolated horizon as a cold reservoir to confirm the thermodynamic nature of ih . for a non - rotating isolated horizon ,
the angular momentum is zero , that is , @xmath123 .
according to the definition of angular momentum of ih and eq .
( [ bj6 ] ) , we find that @xmath124 so we have from the eqs .
( [ rn5 ] ) , @xmath125 the tetrad condition @xmath126 @xcite means that @xmath127 for ih , @xmath13\widehat{=}0 $ ] , so we have @xmath13n_b\widehat{=}0 $ ] . considering @xmath128 , we obtain that @xmath129 from eqs .
( b19 , b20 ) in ref .
@xcite , @xmath130 where @xmath131 , we find that @xmath132 so from the eq .
( [ rn7 ] ) , we have @xmath133 these two conditions above are given by isolated horizon .
so we find that @xmath134 are not related to @xmath19 on the horizon , that is to say , the taylor series of @xmath135 to the first - order approximation of @xmath0 are independent of @xmath19 . from eqs .
( 77 , 78 , 79 ) in ref .
@xcite and eq .
( [ rn6 ] ) the metric and its inverse to the first - order approximation of @xmath0 are not related to @xmath19 , so @xmath121 can be thought as a killing vector approximatively .
based on ref .
@xcite , the necessary and sufficient condition for the existence of a hamiltonian conjugate to @xmath136 on slice @xmath137 is @xmath138 where @xmath139 is the presymplectic current 3-form . in general relativity ,
the presymplectic current 3-form is @xmath140 where @xmath141,\end{aligned}\ ] ] with @xmath142 in our case it is straightforward to find that @xmath143 in the last equality , we use @xmath144 .
so there exists a hamiltonian conjugate to @xmath121 near the horizon , and @xmath121 can be thought as a physical observer .
next we will compute the energy as measured at infinity of a particle near the horizon .
the taylor series of the coefficients to the first - order approximation of @xmath0 are @xmath145where @xmath146 is the first - order derivative of @xmath0 .
putting expressions ( [ hr1 ] ) into the null tetrad ( [ biaojia ] ) , we get @xmath147\frac{\partial}{\partial r},\nonumber\\ n^{a}&=&-\frac{\partial}{\partial r},\nonumber\\ m^{a}&=&o(r^{2})\frac{\partial}{\partial r}+\big[\,\xi_{3}(0)+\xi_{3}^{(1)}(0)r+o(r^{2})\big]\frac{\partial}{\partial \varsigma}\nonumber\\ & & + \big[\,\xi_{4}(0)+\xi_{4}^{(1)}(0)r+o(r^{2})]\frac{\partial}{\partial \overline{\varsigma}},\nonumber\\ \overline{m}^{a}&=&o(r^{2})\frac{\partial}{\partial r}+\big[\,\overline{\xi_{3}}(0)+\overline{\xi_{3}}^{(1)}(0)r+o(r^{2})\big]\frac{\partial}{\overline{\varsigma}}\nonumber\\ & & + \big[\,\overline{\xi_{4}}(0)+\overline{\xi_{4}}^{(1)}(0)r+o(r^{2})\big]\frac{\partial}{\partial \varsigma}.\label{jiehe1}\end{aligned}\ ] ] introduce new coordinates in the neighborhood of ih as @xcite @xmath148 where @xmath149 and @xmath150 .
it is easy to find that @xmath151 we use eqs .
( [ 31 ] ) to express the null tetrad ( [ jiehe1 ] ) in the new coordinates as @xmath152\frac{\partial}{\partial r},\nonumber\\ n^{a}&=&\frac{1}{f}\frac{\partial}{\partial t}-\frac{\partial}{\partial r},\nonumber\\ m^{a}&=&o(r^{2})\big[-\frac{1}{f}\frac{\partial}{\partial t}+\frac{\partial}{\partial r}\big]+\big[\xi_{3}(0)+\xi_{3}^{(1)}(0)r+o(r^{2})\big]\frac{\partial}{\partial \varsigma}\nonumber\\ & & + \big[\xi_{4}(0)+\xi_{4}^{(1)}(0)r+o(r^{2})\big]\frac{\partial}{\partial \overline{\varsigma}},\nonumber\\ \overline{m}^{a}&=&o(r^{2})\big[-\frac{1}{f}\frac{\partial}{\partial t}+\frac{\partial}{\partial r}\big]+\big[\,\overline{\xi_{3}}(0)+\overline{\xi_{3}}^{(1)}(0)r+o(r^{2})\big]\frac{\partial}{\partial \overline{\varsigma}}\nonumber\\ & & + \big[\ , \overline{\xi_{4}}(0)+\overline{\xi_{4}}^{(1)}(0)r+o(r^{2})\big]\frac{\partial}{\partial \varsigma}.\end{aligned}\ ] ] from the expressions above we get easily @xmath153 putting eqs .
( [ hr2 ] ) into eqs .
( [ badir1 ] ) and ignoring the higher order terms , we obtain @xmath154 we will use these results later . the hamilton - jacobi equation in curved spacetime is @xmath155 where @xmath156 is the rest mass , and @xmath157 is the hamilton principal function . because @xmath158 is a killing vector , we can separate @xmath157 as @xmath159 the components of generalized momentum @xmath160 are @xmath161 where @xmath162 is the energy of the particle .
we consider a particle at rest outside a non - rotating isolated horizon , so for this particle , @xmath163 and @xmath164 are constant , which means @xmath165 then eq . ( [ 2 ] ) becomes @xmath166 and it is easy to find that @xmath167 this is the energy as measured at infinity of a particle at rest outside the horizon . when a particle is located at @xmath168 which is the radial coordinate distance away from the horizon , we have @xmath169 putting this expression into eq .
( [ 10 ] ) , we obtain @xmath170 now let us change the coordinate distance @xmath171 to the proper distance @xmath172 .
the general express is @xmath173 we consider the radial distance , so @xmath174 that is , @xmath175^{\frac{1}{2}}dr.\end{aligned}\ ] ] so the proper distance @xmath172 corresponding to coordinate distance @xmath171 is @xmath176^{\frac{1}{2}}dr=\int_{0}^{\delta}\left[\frac{1}{2(\varepsilon+\overline{\varepsilon})r}\right]^{\frac{1}{2}}dr\nonumber\\ & = & \frac{1}{[2(\varepsilon+\overline{\varepsilon})]^{\frac{1}{2}}}\int_{0}^{\delta}\left[\frac{1}{r}\right]^{\frac{1}{2}}dr=\frac{1}{[2(\varepsilon+\overline{\varepsilon})]^{\frac{1}{2}}}\int_{0}^{\delta}r^{-\frac{1}{2}}dr\nonumber\\ & = & \frac{\sqrt{2}}{(\varepsilon+\overline{\varepsilon})^{\frac{1}{2}}}\delta^{\frac{1}{2}}.\label{12}\end{aligned}\ ] ] putting the result above into eq .
( [ zhs1 ] ) , we have @xmath177 the surface gravity of ih is @xmath178 , so the energy can be rewritten as @xmath179 this is the energy as measured at infinity of a particle at rest outside the horizon with the radial proper distance @xmath172 away from the horizon
. the red - shift factor is defined by @xmath180 and we have @xmath181 so the energy as measured at infinity of the particle can be expressed as @xmath182 this is the important result we will use in the following discussion . hawking radiation of isolated horizon was confirmed recently @xcite , however , can it really be regarded as a thermodynamical system ?
we will analyze its thermodynamical property in another way .
following ref .
@xcite , we design a reversible carnot cycle with a non - rotating isolated horizon being the cold reservoir .
the thermodynamic nature of isolated horizon is confirmed further in the framework of carnot cycle .
we divide the total process into the following four steps ( see fig .
1 ) : * step ( a ) * a hot reservoir near an isolated horizon is filled with thermal radiation of temperature @xmath183 . like ref .
@xcite we require the chemical potential of the thermal radiation vanish .
the isolated horizon is used as the cold reservoir , so we need @xmath184 ( @xmath185 is the temperature of ih ) .
the initial state of the working substance is an empty box and the mass of the empty box is negligible . for designing a reversible process ,
we replace one side of the box with a piston in our model .
when we pull the piston slowly down to the other side of the box such that the box finally is full of thermal radiation of temperature @xmath183 , there is some work done by the thermal radiation ( see fig .
1 ) . after the process
above , we have a box of substance with energy @xmath186 .
since the process step ( a ) is quasi - static and isothermal , we can use the first law of thermodynamics in the following form , @xmath187 where @xmath188 is the heat flowing into the box , @xmath189 is the entropy of the matter in the box , and @xmath190 is the work done by the thermal radiation
. please note that all the quantities are locally measured , so the corresponding quantities measured at infinity are @xmath191 where @xmath192 is the red - shift factor at the hot reservoir .
+ * step ( b ) * adiabatic process .
this process actually consists of two sub - steps . * * sub - step ( b1)**the box is lowered slowly down to some point near the black hole with the red - shift factor @xmath193 . in this step
, the substance inside does not change , so all locally measured quantities , such as energy , entropy , temperature , etc , remain unchanged .
however , the energy measured by the agent holding the string at infinity becomes @xmath194 , which means the work done in the lowering process is @xmath195 please note that this is not the thermodynamic work done by the substance in the box since the volume of the substance is unchanged , and the work is done by gravity in a pure mechanical process .
an analog of the adiabatic process in an ordinary carnot cycle is given in sub - step ( b2 ) below .
* sub - step ( b2 ) * this sub - step plays an essential role in our model . in a typical carnot cycle
, the working substance should experience an adiabatic expansion to cool down to temperature @xmath196 , the temperature of the cold reservoir .
however , the adiabatic lowering process sub - step ( b1 ) does not change the temperature of the substance , so it is necessary to add another adiabatic process to cool down the substance from @xmath183 to @xmath196 .
but the temperature at any point outside the black hole is not well - defined since the box is not in direct contact with the isolated horizon .
we can use a plausible criterion , i.e. the conservation of the total entropy , to specify @xmath196 as follows .
suppose that the box cools down to @xmath196 via an adiabatic expansion .
the entropy of box remains @xmath197 ( see eq . ( [ 26 ] ) ) .
then the box gives up all its energy to the black hole which is shown in step ( c ) below .
the energy that the black hole finally absorbs is @xmath198 ( see eq .
( [ 25 ] ) ) .
hence the change in the total entropy is @xmath199 because for a reversible adiabatic process , the total entropy should not change , that is , @xmath200 , we find immediately from eq .
( [ by1 ] ) , @xmath201 different from sub - step ( b1 ) , sub - step ( b2 ) changes the parameter set of the box from @xmath202 to @xmath203 .
the entropy @xmath189 keeps unchanged and the adiabatic expansion takes place in the same location , so the work measured from infinity is given by @xmath204 * step ( c ) * release the substances from the box into the black hole .
we use the piston in the opposite way to push out the thermal radiation in a quasi - static and isothermal manner .
we have @xmath205 where prime is used to label locally measured quantity .
the corresponding quantities measured at infinity are @xmath206 note that @xmath207 is the energy released from the box seen by a local observer while the energy absorbed by the black hole is @xmath208 , instead of @xmath207 .
* step ( d ) * lift the empty box back and return to its initial state .
nothing happens in this process because the mass of the empty box is negligible .
now let us compute the thermal efficiency for the complete reversible carnot cycle .
it is easy to check that the total work @xmath209 satisfies the familiar relation in a typical carnot cycle , @xmath210 hence the efficiency is @xmath211 this is the desired efficiency for a reversible carnot engine operating between two heat sources with temperatures @xmath183 and @xmath212 respectively .
therefore the thermodynamic nature of isolated horizon is confirmed in the framework of carnot cycle and the efficiency of this cycle confirms that isolated horizon behaves as a thermodynamic object with hawking temperature @xmath185 .
in this paper , we investigate the electrical and thermodynamical properties of isolated horizon . by following damour s method
, we establish the ohm s law and joule s law of an isolated horizon , so we generalize damour s results . from the calculation , we find that the results are very general , since we do not use all the conditions of isolated horizon .
we investigate the geometry in the vicinity of a non - rotating isolated horizon , and find that under the first - order approximation of @xmath0 , @xmath121 is a killing vector and there exists a hamiltonian conjugate to it , so @xmath121 is a physical observer .
we calculate the energy as measured at infinity of a particle at rest outside the horizon and construct a reversible carnot cycle with an isolated horizon as a cold reservoir , which gives a further confirmation of the thermodynamic nature of isolated horizon .
99 s. w. hawking and g. f. r. ellis , the large scale stucture of spacetime , cambridge university press , 1973 .
r. m. wald , general relativity , university of chicago press , chicago , 1984 .
d. pejerski and e. newman , j. math . phys . *
s. a. hayward , class .
. grav . * 10*(1993)l137 .
a. ashtekar , c. beetle , s. fairhurst , class .
. grav . * 16*(1999)l1-l7 .
a. ashtekar , et al .
lett , * 85*(2000)3564 .
a. ashtekar , c. beetle , j. lewandowski , class .
quant . grav . *
a. ashtekar , b. krishnan , living rev . rel . *
t. damour , phys .
d * 18*(1978)3598 .
k. thorne , r. price , and d. macdonald , black holes : the membrane paradigm , yale university press ( 1986 ) ; r. h. price and k. s. thorne , phys .
d * 33*(1986)915 .
m. k. parikh , f. wilczek , phys .
d * 58*(1998)064011 .
g. policastro , d. t. son and a. o. starinets , phys .
lett . * 87*(2001 ) 081601 ; g. policastro , d. t. son and a. o. starinets , jhep * 0209*(2002)043 . n. iqbal and h. liu , phys .
d * 79*(2009)025023 ; i. bredberg , c. keeler , v. lysov and a. strominger , jhep * 1103*(2011)141 .
cai , l. li and y .- l .
zhang , jhep * 1107*(2011)027 ; c. niu , y. tian , x .-
n . wu and y. ling , phys . lett .
b * 711*(2012)411 ; v. lysov and a. strominger , phd dissertation , from petrov - einstein to navier - stokes " , harvard university , 2014 .
xiaoning wu , sijie gao , phys .
d * 75*(2007)044027 .
xiaoning wu , chao - guang huang , jia - rui sun , phys .
d * 77*(2008)124023 .
j. d. bekenstein , phys .
d * 7*(1973)2333 .
j. d. bekenstein , phys .
d * 23*(1981)287 . w. g. unruh and r. m. wald , phys .
d * 25*(1982)942 .
h. friedrich , proc .
a * 378*(1981)169 - 184 , 401 - 421 .
badri krishnan , class .
quantum grav . * 29*(2012)205006 .
s. w. hawking , in black holes , edited by b. dewitt and c. dewitt(gordon and breach , new york , 1973 ) .
r. m. wald , a. zoupas , phys .
d * 61*(2000)084027 .
deng xi - hao , gao si - jie , chinese physics b * 18*(2009)927 . | the electrical laws and carnot cycle of isolated horizon ( ih ) are investigated in this paper .
we establish the ohm s law and joule s law of an isolated horizon , and find that the conceptual picture of black holes ( membrane paradigm ) can also apply to this kind of quasi - local black holes .
we also investigate the geometrical properties near a non - rotating ih , and find that under the first - order approximation of @xmath0 , there exist a killing vector and a hamiltonian conjugate to it , so this vector is a physical observer .
we calculate the energy as measured at infinity of a particle at rest outside a non - rotating ih , and use this result to construct a reversible carnot cycle with the isolated horizon as a cold reservoir , which confirms the thermodynamic nature of isolated horizon .
* keywords * : isolated horizon , membrane paradigm , electrical laws , black hole thermodynamics , reversible carnot cycle |
a graph is defined as a set of nodes ( vertices ) and a set of links between nodes ( edges ) @xcite . by graph evolution or growth
we mean subsequent attaching of new nodes with @xmath2 edges to previously existing nodes @xcite .
such growing graphs may reflect some features of real evolving networks , e.g. a network of collaborators , a network of citations of scientific papers , some biological networks ( food chains or sexual relations ) or internet and world - wide - web pages with links between them @xcite .
the distance between nodes is the shortest number of edges which leads from one node to the other .
the node - node distance ( nnd ) distribution depends on how subsequent nodes are attached .
if each node is connected with _ only one _ of preexisting nodes ( @xmath3 ) _ a tree _ appears .
when a new node is attached to several different nodes with @xmath4 edges , the growing structure is called _ a simple graph_. we may choose nodes to which new nodes are attached in preferential way or randomly . in the latter case we deal with _ exponential _ networks .
if the probability of choosing a node is proportional to its degree ( e.g. to the number of its nearest neighbors ) the growing structure is called _ scale - free _ or albert barabsi networks @xcite . in this paper ,
the numerical algorithm for the network growth basing on _ distance matrix _ evolution is presented both , for exponential and scale - free networks ( @xmath5 ) @xcite . the nnd distribution and its characteristics are calculated . for the exponential trees
the iterative formulas for @xmath6-th ordinary moments are derived .
a graph with edges of unit length may be fully characterized by its distance matrix @xmath7 , an element @xmath8 of which is equal to the shortest path between nodes @xmath9 and @xmath10 .
this matrix representation is also particularly useful when computer simulations for graph evolution are applied . attaching subsequent node with one edge ( @xmath3 ) to previously existing network of @xmath11 nodes
corresponds to adding a new @xmath12-th row and a new column to @xmath13 large distance matrix @xmath7 .
the distance from newly added @xmath12-th node to all others via selected labeled as @xmath14 node is larger by one than the distance between @xmath14-th and all others .
thus , new @xmath12-th row / column is a simple copy of @xmath14-th row / column but with all of its elements incremented @xcite : @xmath15 similarly , when new node is attached to the network with two edges ( @xmath16 ) to two different nodes which are labeled as @xmath14 and @xmath17 but the distance from all other nodes @xmath9 to the newly added @xmath12-th is one plus the smaller distance between @xmath14@xmath9 or @xmath17@xmath9 nodes pairs @xcite : @xmath18 in the case mentioned above of the growth of the simple graphs also the reevaluation of distances between nodes @xmath9 and @xmath10 must be done to check if adding a new node provides the shortcut @xcite : @xmath19 in both cases diagonal elements of new row / column are zero @xcite : @xmath20 selecting rows / columns ( nodes to which we attempt to add a new node ) may be random or preferential . in the latter case an additional evolving vector is introduced , which contains the node labels .
these labels occur as vector elements with a probability proportional to the degree of the node
. random selection of elements of such a vector correspond to albert barabsi construction rule .
the procedure is known as the kertsz algorithm @xcite .
let us define @xmath6-th moments of the nnd distribution for all distances @xmath21 = \dfrac{1}{n^2}\sum_{i=1}^n\sum_{j=1}^n [ s^n(i , j ) ] , \label{eq_ave}\ ] ] and only for non - zero distances @xmath22 = \dfrac{1}{n(n-1)}\sum_{i=1}^n\sum_{\substack{j=1\\j\ne i}}^n [ s^n(i , j ) ] , \label{eq_mom}\ ] ] where @xmath23 , @xmath24 and @xmath25 $ ] denote an average over @xmath26 matrix elements , an average over @xmath27 non - diagonal matrix elements , and an average over @xmath28 independent realizations of the evolution process ( matrices ) , respectively .
moments and for @xmath29 are sometimes called _ the network diameter_. both double sums in r.h.s . of eqs . and are equal , due to obvious fact , that @xmath30 .
that allows to derive simple dependence between averages @xmath23 and @xmath24 : @xmath31 for the exponential trees basing on @xmath30 and distance matrix symmetry @xmath32
we are able to construct iterative equations for @xmath33 as dependent on @xmath34 ( @xmath35 ) @xmath36= n^2\ell^n_n+2\sum_{i=1}^n \big(1+[s(i , q)]\big)^n , \label{eq_main}\ ] ] where @xmath17 is the number of the randomly selected row / column of the distance matrix @xmath7 . combining eq . with eq .
gives the desired iterative formula : @xmath37
for the trees the mean of the nnd @xmath38 and its dispersion @xmath39 grow logarithmically with @xmath11 ( see tabs .
[ tab1 ] , [ tab2 ] ) @xcite . for the graphs
only the first cumulant ( the average of the nnd @xmath38 ) grows logarithmically ( see tab .
[ tab2 ] ) @xcite .
such a slow increase of @xmath38 with number of network nodes is known as the small - world effect @xcite .
.the mean distance @xmath40 for different evolving networks . [ cols=">,^,^,^,^",options="header " , ] the histogram of nnd is presented in fig .
[ fig_his ] . as we expected the nnd for the graphs
are more condensed than the nnd for the trees , as well as , the scale - free graphs ( trees ) are more condensed than the exponential graphs ( trees ) . knowing the moments @xmath41 the averages of the @xmath6-th powers of the non - diagonal distance matrix elements
allows to build all statistical parameters which characterize the nnd distribution , e.g. the average distance @xmath42 , the distance dispersion @xmath43 , its skewness @xmath44 and kurtosis @xmath45 the values of such characteristics of the nnd for the exponential trees obtained via eq .
are presented in figs .
[ fig_mom ] and [ fig_sta ] . for trees ,
the distributions are similar to the poisson distribution ( see fig .
[ fig_his ] ) .
however , even for large @xmath11 the skewness and kurtosis do not vanish as one may expect for the normal distribution @xcite .
k.m.s participation in the 37-th polish physicists meeting in gdask was financed by krakw branch of the polish physical society ( ptf ) .
the numerical calculations were carried out in ack cyfronet
the machine time on sgi 2800 is financed by the state committee for scientific research ( kbn ) in poland with grant no .
kbn / sgi2800/agh/018/2003
. 99 j.clark , d.a.holton , _ a first look at graph theory _ ,
world scientific , 1991 .
r.j.wilson , _ introduction to graph theory _ , longman scientific and technical , new york , 1987 .
c.berge , _ graphs _ , north holland , 1985 .
f.harary , _ graph theory _
, addison - wesley , 1969 .
s.n.dorogovtsev , j.f.f.mendes , adv .
* 51 * ( 2002 ) 1079 .
s.n.dorogovtsev , j.f.f.mendes , in _ from the genome to the internet _ ,
s.bornholdt and h.g.schuster , viley - vch , berlin , 2002 .
s.n.dorogovtsev , j.f.f.mendes , a.n.samukhin , nucl . phys . *
b653 * ( 2003 ) 307 .
m.e.j.newman , siam review * 45 * ( 2003 ) 167 .
r.albert , a .- l.barabsi , rev .
* 286 * ( 2002 ) 47 .
l.barabsi , r.albert , science * 286 * ( 1999 ) 509 .
k.malarz , j.czaplicki , b.kawecka-magiera , k.kuakowski , int . j. mod
c14 * ( 2003 ) 1201 .
k.malarz , k.kuakowski , cond - mat/0304693 .
d.stauffer , private communication .
s.milgram , psychol . today * 2 * ( 1967 ) 60 .
g.szab , m.alava , j.kertsz , phys
* e66 * ( 2002 ) 026101 . | we present the simulation of the time evolution of the distance matrix .
the result is the node - node distance distribution for various kinds of networks . for the exponential trees ,
analytical formulas are derived for the moments of the distance distribution .
* node - node distance distribution for growing networks * + k.malarz@xmath0 , j.karpiska , a.kardas , and k.kuakowski@xmath1 _ faculty of physics and nuclear techniques + agh university of science and technology + al .
mickiewicza 30 , pl-30059 krakw , poland + _ e - mail : @[email protected] , @[email protected] |
for standard background from the _ geometry of numbers _ and _ convex geometry _ we refer to @xcite .
the cardinality of a set @xmath10 is denoted by @xmath11 . throughout the paper
we fix a @xmath1-dimensional vector space @xmath12 over @xmath13 with @xmath14 and a lattice @xmath15 of rank @xmath1 in @xmath12 . in most of the cases it will be convenient to use coordinate - free notation
. however , when we use analytic expressions with matrices or the notion of volume , we fix coordinates . in such cases @xmath12
is identified with @xmath3 , the set of columns of @xmath1 real numbers .
if @xmath16 is a lattice of rank @xmath7 , then the _ determinant _ of @xmath16 ( denoted by @xmath17 ) is the @xmath9-dimensional volume of any dirichlet cell of @xmath16 .
a subset @xmath18 of @xmath12 is said to be a _ lattice polytope _ ( with respect to the lattice @xmath15 ) if @xmath18 is the convex hull of finitely many points of @xmath15 .
let @xmath18 be an @xmath9-dimensional lattice polytope in @xmath12 , where @xmath7 .
then by @xmath19 we denote the @xmath9-dimensional _ volume _ of @xmath18 .
we also use the @xmath9-dimensional _ normalized volume _ of @xmath18 defined by @xmath20 where @xmath21 and @xmath10 is the linear hull of vectors @xmath22 with @xmath23 .
we remark that @xmath24 does not depend on the choice of coordinates in @xmath12 .
let @xmath0 denote the family of all @xmath1-dimensional lattice simplices in @xmath12 with precisely one interior lattice point .
it was shown by hensley @xcite that the volume of the elements of @xmath0 is bounded by a constant depending only on @xmath1 . continuing
the research initiated in @xcite , we present bounds on the size of lattice simplices with precisely one interior lattice point .
note that the family @xmath0 appears naturally in _ algebraic geometry _ ; see @xcite .
furthermore , as seen from @xcite , @xmath0 can be used in the study of the class @xmath25 of inclusion - maximal @xmath1-dimensional lattice polytopes without interior lattice points .
the elements of @xmath25 are important for the _ cutting - plane theory _ in integer and mixed integer optimization ( see @xcite and the references therein ) .
we prove the following theorem , which is our main tool .
[ bary : ineq : alt ] let @xmath5 .
let @xmath26 with @xmath27 be the barycentric coordinates of the unique interior lattice point of @xmath2 . then
, for every partition @xmath28 of @xmath29 , one has @xmath30 in theorem [ bary : ineq : alt ] without loss of generality one can assume @xmath31 . then , using , it is possible to determine lower bounds for @xmath26 s in terms of @xmath9 and @xmath1 . as a consequence
we obtain the following result on the maximum size of simplices from @xmath0 .
[ balance : of : simplices ] let @xmath14 . then the following statements hold .
a. [ main : upper : bounds ] for every @xmath5 there exist faces @xmath32 of @xmath2 such that , for every @xmath7 , @xmath8 is @xmath9-dimensional and satisfies @xmath33 b. there exists @xmath34 and faces @xmath35 of @xmath36
such that , for every @xmath7 , @xmath8 is @xmath9-dimensional and satisfies @xmath37 loosely speaking , theorem [ balance : of : simplices ] asserts that the largest simplices from @xmath0 are extremely anisotropic .
part i contains the main message of theorem [ balance : of : simplices ] .
part ii is given as a complement estimating the quality of the upper bounds from part i. for @xmath38 let @xmath39 by theorem [ balance : of : simplices ] , we obtain @xmath40 .
this improves the best available bound @xmath41 , which can be found in ( * ? ? ?
* ( 10 ) ) .
let us give a short overview of results related to theorem [ balance : of : simplices ] .
the study of upper bounds on the size of lattice polytopes with a fixed positive number of interior lattice points was initiated in @xcite and continued in @xcite and @xcite .
the sources @xcite and @xcite present qualitative arguments which show @xmath42 for every @xmath14 ( see also @xcite and @xcite ) .
analoga of the results from @xcite for lattice polytopes without interior lattice points were given in @xcite .
the sources @xcite and @xcite provide classification of @xmath1-dimensional lattice polytopes with precisely one interior lattice point for the cases @xmath43 and @xmath44 , respectively .
inequalities involving the area of lattice polygons can be found in @xcite .
the abbreviations @xmath45 and @xmath46 stand for the affine hull , convex hull and interior , respectively .
the origin of @xmath12 is denoted by @xmath47 . by @xmath48
we denote the standard basis of @xmath3 .
the @xmath49-norm is denoted by @xmath50 .
we shall use the following three well - known results from the geometry of numbers .
[ mink:1st ] let @xmath51 be an open convex set in @xmath12 with @xmath52 and @xmath53 .
then @xmath51 contains a nonzero element of @xmath15 .
[ mink : cor ] let @xmath54 be a real @xmath55 matrix with @xmath56
. then there exists @xmath57 satisfying @xmath58 . [ blichfeldt ]
let @xmath59 be a compact convex set in @xmath12 containing @xmath60 points of @xmath15 that linearly span @xmath12
. then @xmath61 .
note that theorem [ mink : cor ] is a straightforward consequence of theorem [ mink:1st ] .
let @xmath2 be a @xmath1-dimensional simplex in @xmath12 and let @xmath62 be the vertices of @xmath2 .
then every point @xmath63 can be uniquely given by the values @xmath64 satisfying @xmath65 the values @xmath66 are said to be the _ barycentric coordinates _ of @xmath67 with respect to the simplex @xmath2 .
the barycentric coordinates have the following geometric interpretation .
let @xmath68 be the affine hull of the facet of @xmath2 opposite to @xmath69 .
then @xmath70 is the _ signed weighted distance from @xmath67 to @xmath68 _ : @xmath71 is the ratio of the distance of @xmath67 to @xmath68 to the distance of @xmath69 to @xmath68 , and @xmath72 if and only if @xmath67 and @xmath69 are in the same open halfspace defined by @xmath68 .
in particular , we have : * @xmath67 lies in the interior of @xmath2 if and only if @xmath73 for all @xmath27 . * for @xmath27 , @xmath74 if and only if @xmath75 .
the barycentric coordinates @xmath66 of @xmath67 can be found from @xmath76 ( we recall that in matrix expressions we assume @xmath77 . ) since @xmath62 , are affinely independent , the @xmath78-matrix in is invertible .
thus , we can express @xmath70 s by @xmath79 it follows that for every @xmath27 we have @xmath80 , where @xmath81 is an affine - linear function in @xmath67 .
( a real - valued function @xmath82 on @xmath12 is called _ affine - linear _ if @xmath83 for all @xmath84 . )
the functions @xmath85 are uniquely determined by @xmath86 where @xmath87 denotes the _ kronecker delta_. this follows by applying @xmath88 to and then using the affine - linearity .
we shall use pairs @xmath28 of disjoint subsets of @xmath29 with @xmath89 .
such a pair @xmath28 is called a _ partition _ if both @xmath90 and @xmath91 are nonempty .
every face of @xmath2 can be represented by @xmath92 for an appropriate pair @xmath28 as above .
clearly , partitions @xmath28 correspond to proper faces of @xmath2 .
let @xmath5 and let @xmath93 be the barycentric coordinates of the interior lattice point of @xmath2 .
our aim is to provide positive lower bounds for @xmath26 s in terms of @xmath1 .
let us sketch how such bounds can be obtained ( the idea is due to hensley ( * ? ? ?
* the proof of lemma 2.3 ) ) .
let @xmath94 be the interior lattice point of @xmath2 .
one can show by contradiction that @xmath94 can not be too close to the boundary of @xmath2 ( that is , the minimal @xmath26 can not be too small ) .
if @xmath94 is close to a face @xmath95 of @xmath2 , then , with the help of minkowski s first fundamental theorem , one can determine a point @xmath96 in @xmath97 whose barycentric coordinates are rational numbers of the form @xmath98 with @xmath99 and @xmath100 and such that @xmath96 is close to @xmath94 .
it then turns out that @xmath101 is another integral point in the interior of @xmath2 , yielding a contradiction to @xmath5 .
see also fig .
[ jumping - idea ] for an illustration .
let @xmath94 be the interior lattice point of @xmath2 .
the point @xmath94 can be given by @xmath104 , where @xmath62 are appropriately indexed vertices of @xmath2 .
without loss of generality let @xmath105 with @xmath106 .
consider real unknowns @xmath107 and @xmath108 and the vector @xmath109 if @xmath110 then @xmath111 is an affine combination ( with rational coefficients ) of @xmath112 and @xmath113 is an affine combination of @xmath94 and @xmath96 belonging to @xmath15 . below
we determine a sufficient condition for the existence of values @xmath114 as above for which the point @xmath102 does not coincide with @xmath94 and lies in the interior of @xmath2 . if holds , then @xmath115 with @xmath116 and @xmath117 with @xmath118 are the barycentric coordinates of @xmath102 .
thus ( taking into account the geometric meaning of the barycentric coordinates ) we obtain that , assuming , the following conditions are equivalent : for theorem [ mink : cor ] can be applied , but first we investigate the relation of to . we claim that if is solvable with respect to @xmath123 , then also is solvable with respect to @xmath123 .
assume that conditions are fulfilled .
then @xmath108 is not zero .
in fact , if @xmath126 , then together with and @xmath127 implies @xmath128 for every @xmath121 .
this yields a contradiction to .
hence @xmath129 .
the system is invariant under the change of @xmath123 to @xmath130 .
thus , we can assume @xmath131 .
condition together with implies .
the implication from to is trivial .
this verifies the claim .
we compute the determinant of @xmath134 by transforming the matrix into an upper triangular form : all @xmath137 s in the last row can be turned to @xmath138 by adding an appropriate linear combination of the first @xmath135 rows of @xmath134 .
thus @xmath139 hence @xmath140 and since @xmath141 , we even have @xmath142 .
we show @xmath143 by contradiction .
assume that @xmath144 .
then , by theorem [ mink : cor ] , is solvable .
the system is a matrix form of . above we showed that solvability of implies solvability of . thus , there exists @xmath123 satisfying . with such @xmath123 the lattice point @xmath102 given by does not coincide with @xmath94 and lies in the interior of @xmath2 .
this yields a contradiction to @xmath5 .
thus , @xmath143 and the assertion follows from .
theorem [ bary : ineq : alt ] presents exponentially many inequalities for the barycentric coordinates of the interior lattice point : one for each @xmath90 with @xmath145 .
it turns out that most of these inequalities are redundant .
in fact , assume that for @xmath146 and @xmath147 one has @xmath148 .
then we can modify @xmath28 by moving @xmath9 from @xmath90 to @xmath91 and moving @xmath149 from @xmath91 to @xmath9 . by this change
we lower the left hand side and raise the right hand side of making the inequality tighter .
we thus see that the inequalities for the partitions @xmath28 satisfying @xmath150 for every @xmath146 and every @xmath147 imply the inequalities for all the remaining partitions .
such strongest inequalities can be described as follows .
after an appropriate reindexing , we can assume that the barycentric coordinates @xmath93 are in an ordered sequence , say @xmath31 .
then , taking into account the redundancy , only the following @xmath1 inequalities remain : below we use to derive lower bounds for the barycentric coordinates @xmath93 .
first we observe that the inequalities @xmath153 together with @xmath154 yield @xmath155 .
inequality can be relaxed to @xmath156 on the left hand side of we could use a smaller factor instead of @xmath157 , but with the factor @xmath157 the lower bounds for @xmath158 s that we give below can be expressed by simpler formulas . by each @xmath159
is bounded in terms of its ` predecessors ' @xmath160 .
thus , consecutively applying we can bound every @xmath158 in terms of @xmath149 and @xmath161 . taking into account @xmath155
, we then see that each @xmath158 can be bounded from below in terms of @xmath149 and @xmath1 . in this way
we arrive at part i of the following a. for every @xmath5 the ( ordered ) barycentric coordinates @xmath31 of the unique interior lattice point of @xmath2 satisfy @xmath162 for every @xmath27 .
b. [ beta : for : zpw ] there exists @xmath34 such that the barycentric coordinates @xmath31 of the unique interior lattice point of @xmath36 satisfy @xmath163 for every @xmath27 .
we use induction on @xmath9 . for @xmath164 , is fulfilled .
assume the inequalities hold for all @xmath165 with some @xmath152 .
using and the inductive assumption we obtain @xmath166 thus , has been verified for @xmath167 .
[ zpw - example ] let @xmath77 and @xmath168 .
below we introduce the simplex @xmath34 given in @xcite .
the simplex @xmath36 is a slight modification of the simplex constructed by zaks , perles and wills @xcite . we first introduce a recursive sequence @xmath169 by the following relations for @xmath172 will be useful : @xmath173 relations , and can be found in @xcite .
we define the simplex @xmath174 hensley @xcite noticed that @xmath175 is the unique interior lattice point of @xmath36 .
thus , @xmath34 .
the bounds show that @xmath36 is large for large @xmath1 .
consider @xmath36 as in example [ zpw - example ] . from we
see that @xmath176 and @xmath177 are the barycentric coordinates of the interior lattice point of @xmath36 .
the assertion follows from the lower bounds for @xmath172 given in . for @xmath5 let @xmath178 be the smallest barycentric coordinate of the interior lattice point of @xmath2 . by theorem [ beta : bounds].i .
we show that @xmath179 from the arguments given in the remainder of the paper it will be seen that the finiteness of @xmath180 is equivalent to the inequality @xmath181
. pikhurko ( * ? ? ?
* section 7 ) indicates that @xmath181 follows from the arguments of lawrence ( * ? ? ?
* lemma 5 ) .
however , @xcite does not yield any explicit positive lower bound for @xmath182 for a general @xmath1 .
the bound is tight for @xmath164 .
this can be seen by considering the simplices @xmath183 and @xmath184 ( with respect to @xmath168 ) whose unique interior lattice point is the center of mass .
for large @xmath1 s ( * ? ? ? * lemma 2.2 ) provides the better bound @xmath185 than our bound @xmath186 contained in . from this
we see that the bound seems to become less tight as @xmath9 moves from @xmath138 to @xmath1 .
still it should be mentioned that our lower bound for @xmath187 is better than the bound from ( * ? ? ?
* lemma 2.2 ) for @xmath188 .
the question on the determination of @xmath182 given by was raised by hensley @xcite , who conjectured that the infimum in is attained for the simplex @xmath36 from example [ zpw - example ] .
this conjecture is true for simplices @xmath5 of the form @xmath189 with @xmath190 , as follows from @xcite and @xcite . in the following lemma
we discuss affine sections of @xmath2 passing through a fixed interior point of @xmath2 and parallel to faces of @xmath2 , for an illustration see fig .
[ simplex - sections ] .
we shall use the affine - linear functions defined by .
[ volumes : of : sections ] let @xmath2 be a @xmath1-dimensional simplex in @xmath12 with vertices @xmath62 .
let @xmath94 be a point in @xmath124 with barycentric coordinates @xmath93 .
let @xmath191 be disjoint sets with @xmath89 and @xmath192 and let @xmath193 then the following holds : a. the set @xmath194 is a simplex of dimension @xmath195 , and the points @xmath196 with @xmath147 are the vertices of @xmath194 .
b. if @xmath197 , then the values @xmath198 with @xmath147 are the barycentric coordinates of @xmath67 with respect to @xmath194 . c. one has @xmath199 , where @xmath200 is defined by .
thus @xmath67 is a convex combination of the points @xmath196 with @xmath147 .
these @xmath203 points are affinely independent since they are obtained by applying the nonsingular affine transformation @xmath204 ( @xmath205 ) to the affinely independent points @xmath206 with @xmath147 .
this yields parts i and ii .
part iii follows in view of the equality @xmath207 .
[ vol : bary : lem ] let @xmath2 be a @xmath1-dimensional simplex in @xmath12 such that the interior of @xmath2 contains precisely one point of @xmath15 .
let @xmath208 be the barycentric coordinates of the unique interior lattice point of @xmath2 and let @xmath209 be a subset of @xmath29 of cardinality @xmath1 .
then @xmath210 let @xmath94 be the lattice point in the interior of @xmath2 .
we have @xmath211 , where @xmath69 s are appropriately indexed vertices of @xmath2 . without loss of generality
assume @xmath212 .
consider affine - linear functions @xmath85 defined by .
we have @xmath213 for @xmath27 .
the simplex @xmath214 is a reflection of @xmath2 in @xmath94 ; @xmath215 can be described as @xmath216 where @xmath217 s are defined by @xmath218 for @xmath219 and @xmath27 . by affine - linearity of @xmath220 , for every @xmath221 we have @xmath222
consequently , for every @xmath63 , @xmath223 or @xmath224 .
the set @xmath225 is a parallelotope with center at @xmath94 .
let us show @xmath226 ( see also fig .
[ box - from - simplex ] ) : since @xmath94 is the only interior lattice point of both @xmath2 and @xmath215 , we deduce that @xmath94 is the only interior lattice point of @xmath18 .
due to ` affine invariance ' of the ratio @xmath229 for computing @xmath229 we can assume that @xmath77 and @xmath230 . with this assumption we get
@xmath231 and @xmath18 is precisely the set of vectors @xmath232 for which @xmath233 .
then @xmath234 , @xmath235 and we obtain @xmath236 . by minkowski s fundamental theorem , @xmath237 .
hence @xmath238 [ face : vol : versus : beta ] let @xmath5 and @xmath93 be the barycentric coordinates of the interior lattice point of @xmath2 .
let @xmath90 and @xmath209 be disjoint sets such that @xmath239 is a @xmath1-element subset of @xmath29
. then @xmath240 let @xmath241 .
by lemma [ volumes : of : sections].iii we have @xmath242 with @xmath194 and @xmath243 defined by and , respectively . by lemma [
volumes : of : sections].ii , the interior lattice point of @xmath2 has barycentric coordinates @xmath244 with @xmath147 with respect to the simplex @xmath194 .
thus , applying lemma [ vol : bary : lem ] ( for dimension @xmath195 ) to the @xmath245-dimensional simplex @xmath194 we get @xmath246 first we show part i. let @xmath247 be the barycentric coordinates of the unique interior lattice point of @xmath2 . for @xmath7
we define @xmath8 by @xmath248 , where @xmath249 . by theorems [ beta : bounds ] and
[ face : vol : versus : beta ] we obtain @xmath250 this yields .
inequality follows directly from and blichfeldt s theorem .
j. lawrence , _ finite unions of closed subgroups of the @xmath257-dimensional torus _ , applied geometry and discrete mathematics , dimacs ser .
discrete math .
, vol . 4 , amer .
soc . , providence , ri , 1991 , pp . | let @xmath0 be the set of all @xmath1-dimensional simplices @xmath2 in @xmath3 with integer vertices and a single integer point in the interior of @xmath2 .
it follows from a result of hensley that @xmath0 is finite up to affine transformations that preserve @xmath4 .
it is known that , when @xmath1 grows , the maximum volume of the simplices @xmath5 becomes extremely large .
we improve and refine bounds on the size of @xmath5 ( where by the size we mean the volume or the number of lattice points ) .
it is shown that each @xmath5 can be decomposed into an ascending chain of faces @xmath6 such that , for every @xmath7 , @xmath8 is @xmath9-dimensional and the size of @xmath8 is bounded from above in terms of @xmath9 and @xmath1 .
the bound on the size of @xmath8 is double exponential in @xmath9 .
the presented upper bounds are asymptotically tight on the log - log scale .
primary 52b20 ; secondary 14j45 , 90c11 barycentric coordinates ; inequality ; lattice - point enumerator ; lattice polytope ; minkowski s fundamental theorem ; simplex ; volume |
laser wakefield acceleration ( lwfa ) @xcite offers the potential to construct compact accelerators that have numerous potential applications , including the building blocks for a next generation linear collider and the electron beam source for ultra - compact xfels .
it has thus attracted extensive interest , and the last decade has seen an explosion of experimental results .
fully nonlinear particle - in - cell ( pic ) simulations have been instrumental in this progress as an aid in designing new experiments , in interpreting experimental results , and in testing new ideas .
furthermore , developing predictive theoretical models is challenging due to the strong nonlinear effects that are present in the blowout regime of lwfa @xcite ; therefore numerical simulations are also critical in exploring the physics of lwfa .
particle - in - cell simulations have been extensively applied in lwfa research because the pic algorithm follows the self - consistent interactions of particles through the electromagnetic ( em ) fields directly calculated from the full set of maxwell equations . when modeling lwfa using the pic algorithm the laser wavelength needs to be resolved which is usually on the scale of 1 @xmath3 ; meanwhile , the length of the plasma column that the laser propagates through can be on the scale of @xmath4 to @xmath5 @xmath3 . as a result , of this disparity in cell size and propagation distance , full three - dimensional ( 3d ) pic simulations of lwfa
can be very cpu - time consuming .
to capture the key physics while reducing the computation time , reduced models are continually being proposed .
these include models that combine the ponderomotive guiding center with full pic for the wake @xcite or with the quasi - static approach @xcite . however , these models can not as yet model full pump depletion lengths , and the quasi - static approach can not model self - injection .
recently , two methods have been proposed that can speed up the lwfa simulation without losing key physics in the modeling of lwfa .
one method is the lorentz boosted frame technique @xcite . in this method
the lwfa simulations are performed in an optimized lorentz boosted frame with velocity @xmath6 , in which the length of the plasma column is lorentz contracted , while the laser wavelength is lorentz expanded . assuming the reflection of the laser light is not important in the lab frame , then in a properly chosen lorentz transformed frame the time and space scales to be resolved in a numerical simulation are minimized , and savings of factors of @xmath7 can be achieved .
another method that has been recently proposed is to expand the fields in to azimuthal harmonics and to truncate the expansion @xcite .
this can reduce the computational costs of modeling 3d problem with low azimuthal asymmetry to that on the order of 2d @xmath2 simulations .
it was pointed out in ref . @xcite that it would be intriguing to combine these two methods in order to combine the speedups provided by each . similarly to full pic simulations in the cartesian geometry , it was found that in the quasi-3d geometry one of the obstacles to performing lorentz boosted frame simulations is the multi - dimensional numerical cerenkov instability ( nci ) @xcite that inevitably arises due to the unphysical coupling between langmuir modes ( main and aliasing ) and em modes of the relativistic drifting plasma in the simulations .
the coupling arises in the lorentz boosted frame between modes which are purely longitudinal ( langmuir modes ) and purely transverse ( em modes ) in the lab frame .
the coupling occurs at specific resonances @xmath8 where @xmath9 and @xmath10 are the time and space aliases and @xmath11 and @xmath12 are the time step and grid size respectively .
while the multi - dimensional nci theory in cartesian coordinates has been well studied ( see e.g. @xcite ) , there are currently no analytical expressions for the numerical dispersion relation of relativistic plasma drifting in the quasi-3d geometry .
however , osiris @xcite simulations have shown that its behavior for the quasi-3d @xmath2 geometry is very similar to that in cartesian geometry .
it was therefore recently proposed and demonstrated that a hybrid yee - fft solver could be used to suppress the nci in the cartesian and quasi-3d geometries @xcite . in the regular yee ( a finite difference ) solver in a quasi-3d geometry @xcite ,
maxwell equations are solved in @xmath13 space for each azimuthal mode @xmath14 . in the hybrid yee - fft solver
, we perform a fourier ( discrete ) transform in the drifting direction of the plasma ( denote as @xmath15 ) , and solve maxwell equations in @xmath16 space for each mode @xmath14 ; meanwhile , in the @xmath17 direction the derivatives are represented as second order finite difference operators on a yee grid .
the current is corrected to maintain the correctness of gauss law . when maxwell s equations are solved in this way , the corresponding nci modes can be systematically eliminated by applying the same strategies used for a multi - dimensional spectral maxwell solver @xcite .
the fastest growing modes of the nci at @xmath18 can be conveniently suppressed by applying a low - pass filter in the current , the highly localized @xmath19 nci modes can be moved away from physical modes by reducing the time step , and can be completely eliminated by modifying the em dispersion at the @xmath16 range where the @xmath19 nci modes are located .
furthermore , higher order spatial aliasing nci modes can be suppressed by applying higher order particle shapes .
we present osiris simulation results which show that lorentz boosted simulations of lwfa can be performed in this geometry with no evidence of nci .
it is worth noting that recently a pic algorithm based on a fully spectral solver in quasi-3d geometry has been proposed by lehe et .
. this scheme was demonstrated with a single - node algorithm .
in addition , according to how the lab frame information is located in the @xmath20 space , we show that the computation loads can be further reduced by applying a moving window in the boosted frame simulation .
in the boosted frame the window follows the plasma as opposed to the laser , which is the case when using a moving window in the lab frame .
the remainder of this paper is organized as follows : in section [ sect : hybridsolver ] we briefly discuss the hybrid yee - fft solver in quasi-3d geometry , and the corresponding nci mitigation strategies . in section [
sect : simsetup ] , we discuss the simulation setups for modeling lwfa in the lorentz boosted frame .
we discuss the distribution of the data needed for the reconstruction of lab frame information with an emphasis on showing that using a moving window in the direction of the plasma drift can further reduce the computation load .
we then show sample quasi-3d simulations of lwfa in the lorentz boosted frame in section [ sect : sample ] , and compare the results with the corresponding lab frame data .
good agreement is obtained demonstrating that the lorentz boosted frame technique together with the quasi-3d algorithm and a moving window can be combined to achieve unprecedented speed up .
the results are summarized in section [ sect : summary ] .
a key issue that needs to be addressed when performing lwfa simulations in a lorentz boosted frame is the existence of a violent numerical instability , called the numerical cerenkov instability ( nci ) .
the nci arises when a plasma drifts relativistically on the grid .
there has been much recent progress in identifying the nci as the source of the instability , in deriving the numerical dispersion relations and determining growth rates , and in identifying mitigation strategies @xcite . in ref .
@xcite a hybrid yee - fft solver was proposed for the elimination of the nci in cartesian geometry . in this solver ,
maxwell equations are fourier transformed in the drifting direction of the plasma ( denoted as the @xmath15 direction ) .
the fields are solved in the corresponding @xmath21 space , where conventional second order finite difference operators on a yee mesh are used in @xmath22 .
when maxwell equations are solved in this way , the corresponding em dispersion of the solver leads to nci patterns that are very similar to those from a fully spectral maxwell solver in which maxwell equations are solved in multi - dimensional @xmath23-space .
therefore one can systematically eliminate the nci using approximately the same strategies developed for a fully spectral solver .
importantly , the hybrid yee - fft solver works for both cartesian geometry @xmath24 , and quasi-3d geometry @xmath25 .
when the field solver is modified from a standard yee solver to a hybrid yee - fft solver , essentially the em dispersion in the @xmath15 direction is modified from second - order accuracy ( derived from its finite difference form ) into a greater than @xmath26-th order accuracy . however , in osiris ( and most of the modern pic codes ) the @xmath27 and @xmath28 fields are advanced using faraday s law and ampere s law , while gauss s law is satisfied by applying a charge conserving current deposition scheme @xcite . if the continuity equation is rigorously satisfied at each time step then by taking the finite difference version of ampere s law , gauss law is satisfied if it is true at @xmath29 .
however , the rigorous charge conserving current deposit is known only for second order finite difference operators .
therefore , when we use a fft for the differential operator along @xmath15 direction in faraday s and ampere s law , we need to modify the current appropriately so the continuity equation is still true for the modified differential operator . to accomplish this ,
we fourier transform the current for each azimuthal mode , @xmath30 , along @xmath15-direction obtained from the current deposition scheme described in @xcite , and then apply a correction of the form , @xmath31 before we use it in the solver .
this correction is applied to each azimuthal mode @xmath14 in the simulation to ensure that gauss s law is satisfied for each @xmath14 mode .
we then fourier transform @xmath27 and @xmath28 along @xmath15-direction , and solve faraday s law and ampere s law for each azimuthal mode @xmath14 , and each fourier mode @xmath16 , using the corrected current as the source term , @xmath32 note that @xmath33 and @xmath34 adopt the conventional finite difference form as in the yee solver .
the code is gridless in @xmath35 so @xmath36 is replaced with @xmath37 .
we have found previously that the nci pattern for the quasi-3d hybrid yee - fft solver is similar to its counterpart in the cartesian 3d geometry @xcite . as a result , we can apply approximately the same mitigation strategies used for the fully spectral solver in cartesian geometry to systematically eliminate the nci modes for this solver @xcite .
we first eliminate the fastest growing @xmath38 modes ( @xmath10 is the spatial aliasing in @xmath15 direction ) by applying a low - pass filter in the current .
this can be easily accomplished since the current density is already in @xmath16 space after the fourier transform .
the second fastest growing nci modes @xmath19 can be eliminated by reducing the simulation time step .
using a reduced time step not only reduces the @xmath19 nci growth rate but also moves their locations away from the modes where important physics is occurring .
note that this can not be done using a pure finite difference solver since numerical dispersion errors for light waves get worse as the time step is reduced if the cell sizes remain fixed .
the @xmath19 nci modes can also be eliminated by slightly modifying the @xmath16 operator to create a small bump in the dispersion relation to precisely avoid intersections between the main em modes and main langmuir modes at that highly localized region in @xmath16 . as for higher order nci modes , their growth rates can be reduced by applying higher order particle shapes . applying the strategies described above , we have found that we can systematically eliminate the nci modes in the quasi-3d geometry .
fourier transforming the current into @xmath16 space is not only important for the efficient filtering of the nci modes , but is also required to accurately correct ( compensate ) the current in @xmath16 space to exactly match the modified maxwell solver .
it is worth noting that it is now a common practice to modify either the maxwell solver or the field interpolation to change the em dispersion relation in order to obtain a more desirable dispersion relation @xcite . within these schemes ,
gauss law is satisfied by either directly solving it ( as is the case in upic @xcite ) , or by using a current that satisfies the continuity equation through a correction ( compensation ) to match the current deposition scheme with the maxwell solver ( as is the case in here and in @xcite ) . in the approaches which aim to eliminate the nci @xcite ( rather than reducing its growth rate for any time step size to that for an optimum time step @xcite ) , the use of an fft is inevitable ( at least in one direction ) since it is very challenging to design a current deposition scheme in real space to match the em dispersion of a modified maxwell solver for an arbitrary order finite difference scheme .
the setup of a quasi-3d lwfa simulation in a lorentz boosted frame is almost identical to its counterpart in cartesian 2d/3d geometry . in a boosted frame with lorentz factor @xmath39 that moves in
the propagation direction of the laser , the laser pulse is colliding with a counter - propagating relativistically drifting plasma @xcite . due to the lorentz transform , the plasma density increases by @xmath39 while the total plasma column length contracts by @xmath39 . the laser wavelength and pulse length stretch by @xmath40 , while its rayleigh length contracts by @xmath39 . to avoid initializing a laser with very wide transverse size due to the contracted rayleigh length and stretched pulse length ,
a moving antenna is placed at the edge of the plasma boundary to inject a laser pulse into the plasma @xcite . in lwfa simulations in the lab frame (
i.e. , a stationary plasma ) the use of a moving window @xcite , which only follows the physical domain near the laser , significantly reduces the computational load .
the moving window essentially drops plasma sufficiently far behind the laser and adds fresh plasma in front of the laser .
this is illustrated in fig .
[ fig : x1trange ] ( a ) where we plot the range of space time data from a lab frame simulation .
the solid box shows the total space time area while the dashed box shows the reduced area through the use of a moving window .
the moving window has a length @xmath41 $ ] , and the simulation duration is @xmath42 $ ] .
we also show the simulation data that is dumped as colored lines .
the data is dumped every @xmath43 $ ] .
the red ends of the data lines indicate the starting end of the moving window , while the blue ends indicate the rear end .
connecting the red ends of the data lines , we obtain the @xmath44 relation for the head of the moving windows , @xmath45 ( the speed of light @xmath46 is normalized to 1 ) .
the data obtained in the lab frame ( assuming the code dumps data at a constant time interval ) rotates in space - time in the boosted frame since the lorentz transform is essentially a hyperbolic rotation of coordinates in minkowski space @xcite
. therefore lines of data in @xmath15 taken at fixed time from a lorentz boosted frame are rotated by the lorentz transform , i.e , @xmath47 .
the slope of each data line now becomes @xmath48 , where @xmath49 and each data line in the lab frame which belongs to the same point in time in lab frame is now spread over a range of @xmath50 and @xmath51 .
interestingly , when we connect the red end of each data line in the boosted frame it still has a slope of @xmath46 , i.e. @xmath52 . the range of data in the boosted frame is shown in figs .
[ fig : x1trange ] ( b ) , ( c ) and ( d ) . the data in fig .
[ fig : x1trange ] ( b ) , ( c ) corresponds to @xmath53 while that in fig .
[ fig : x1trange ] ( d ) corresponds to @xmath54 . in figs .
[ fig : x1trange ] ( b ) and ( c ) we also show the smallest area ( domain enclosed by dashed lines ) in @xmath55 space that includes the area needed to reconstruct the lab frame data for the two different values of @xmath56 .
this illustrates that the space - time area in the boosted frame can be minimized by using a moving window in this frame . in fig .
[ fig : x1trange ] ( b ) it is seen that this window moves to the left ( backwards ) ; while in fig .
[ fig : x1trange ] ( c ) the window moves to the right ( forwards ) .
we use such moving window in the boosted frame osiris simulations . currently , upic - emma boosted frame simulations in cartesian 2d/3d geometry uses a stationary window [ fig . [
fig : x1trange ] ( d ) ] @xcite .
range of important data in lab and boosted frame simulations .
( a ) range of data in lab frame ( stationary plasma ) simulation with and without a moving window , ( b ) range of data in a boosted frame simulation with @xmath53 including with a moving window , ( c ) range of data in a boosted frame simulation with @xmath54 with a moving window , and ( d ) range of data in a boosted frame simulation with @xmath53 without a moving window.,scaledwidth=100.0% ] from fig .
[ fig : x1trange ] it is evident that in lab frame simulations we usually dump data sparsely in time ( large time intervals between time outputs ) , but the data at each grid is dumped at each time output . on the other hand , in order to recover the equivalent lab frame data in a boosted frame simulation , we need to sample boosted frame data at a much higher rate in time , but only need a small number of spatial locations .
this can be seen by plotting a line across @xmath51 for a fixed @xmath50 .
this line only intersects the equivalent lab frame data at the same number of spatial locations as the number of time outputs .
we typically dump the boosted frame data in a standard form ( all grid points at small number of time steps ) as well as the data needed to transform the results back to the lab frame ( a small number of interpolated grid points at a large number of time steps ) .
we then post - process the later data by performing the inverse rotation back into lab frame for comparison with the lab frame data . when running in the lab frame we also plot the necessary data needed to reconstruct the data into a boosted frame .
this inverse construction method is useful during the development of a boosted frame code , as one can transform the lab frame data that has been extensively cross checked with theory , to the boosted frame , and compare the results against the results obtained by the boosted frame code . in fig .
[ fig : simsetup ] , we present a typical setup for a boosted frame simulation .
the moving window moves from right to left following the drifting plasma .
the moving antenna is also moving from right to left and injects the laser pulse from the left plasma boundary into the plasma .
we place a damping section at the rear ( right ) end of the moving window ( there is a gap between the plasma and the damping region ) to damp the em field to zero in this region .
this is done because periodic boundary conditions are applied in the @xmath15 direction when using the hybrid yee - fft solver , which requires that the em fields need to be zero at the rear end of the simulation window to match the fields at the opposite side ; otherwise the em field at the rear end will reappear at the starting end .
we note that there will be a low level of em reflection from the damping section . in an fft solver ,
the group velocity of light in vacuum is greater than the speed of light , however , since the simulation window is moving at the speed of light and the drifting plasma is drifting ultra - relativistically away from it , the reflected energy is not able to catch up with the drifting plasma .
hence the physics inside the plasma will not be affected by the reflecting em waves .
we have compared cases with the moving window plus the damping regions against cases without moving window to confirm that the moving window plus damping region works @xcite .
we also note that for high @xmath56 boosted frame simulations , we find that the modified pusher described in ref . @xcite is required in order to get the evolution of the bubble correct . as pointed out in ref .
@xcite the usual leap frog staggering leads to issues for the lorentz force when there is near cancellation of the electric and magnetic forces for relativistically moving particles .
determining at what @xmath56 the modified pusher in @xcite is needed is an area of future work .
simulation setup for a typical lwfa simulation in the boosted frame .
the moving window follows the drifting plasma moving from right to left .
a moving antenna injects laser pulse from left to right , and a damping region is located at the rear end of the moving window which also moves from left to right.,scaledwidth=100.0% ]
.parameters for the 3d and quasi-3d lwfa simulations in the lorentz boosted frame ( discussed in section [ sect : boostedframedata ] ) . the laser frequency @xmath57 and number @xmath58 in the lab frame are used to normalize simulation parameters .
the density is normalized to the critical density in the lab frame , @xmath59 .
normalized vector potential @xmath60 for the laser has been converted to linear polarization .
[ cols= " < , > " , ] comparison of simulation results in 3d and quasi-3d geometries for the @xmath61 ( converted to linear polarization ) 10 gev lwfa stage run .
all results are from boosted frame simulations . on the left
are the @xmath62 modes from @xmath63 and @xmath64 . on the right
are the @xmath65 modes for @xmath63 and @xmath64 .
results from a full 3d osiris case are compared against a quasi-3d osiris case where only @xmath66 modes were kept .
, scaledwidth=100.0% ] higher order @xmath14 modes for @xmath63 and @xmath64 obtained from a full 3d lwfa boosted frame data .
on the left are @xmath63 and @xmath64 for mode @xmath67 , while of the right are @xmath63 and @xmath64 for mode @xmath68.,scaledwidth=100.0% ] simulation results for a @xmath61 ( converted to linear polarization ) 10 gev lwfa stage run .
( a ) shows the comparison of envelope of @xmath69 field , which shows the evolution of laser driver as it propagates through the plasma ; ( b ) shows the corresponding comparison of the amplitude of @xmath70 , which shows how the wakefield of the bubble varies in the two frames due to the different self - injection results ; ( c ) , ( e ) , and ( g ) are comparisons of line out for the wakefield , laser envelope , and laser spot size respectively at lab frame time @xmath71 , while ( d ) , ( f ) , ( h ) are the corresponding plots at @xmath72 .
, scaledwidth=100.0% ] line outs of wakefield and laser pulse shape at various lab frame time for a @xmath73 case is shown .
since there are no self - injection in the lab frame for this case , much better agreements are obtained for the wake field part.,scaledwidth=100.0% ] in this section , we present two sample simulations .
we begin by comparing results from two boosted frame simulations where in one case we use full 3d osiris and in the second case we use quasi-3d osiris .
the parameters match those in ref .
@xcite whereby a 200 tw laser is focused to a spot size of @xmath74 at the entrance of a @xmath75 density plasma .
the fwhm pulse length of the laser was 35 fs and the normalized vector was @xmath76 for a linearly polarized laser or @xmath77 for a circularly polarized laser .
this corresponds to a 1.3 gev output electron energy according to the scaling laws in ref .
the numerical parameters are shown in table [ tab : lwfapara0 ] .
we then compare the output in the boosted frame for various azimuthal mode numbers .
this comparison requires the use of a post - processing algorithm which decomposes the full 3d data into azimuthal modes @xcite .
we then compare the data of a lwfa boosted frame simulation in quasi-3d lab with the corresponding quasi-3d boosted frame simulation . for this simulation
we explore parameters for which a full 3d lab frame simulation is not feasible due to the large cpu hours required .
the parameters correspond to a 1.8 pw laser focused to a spot size of @xmath78 at the entrance of a @xmath79 density plasma .
the fwhm pulse length of the laser was @xmath80 and the normalized vector was @xmath61 for a linearly polarized laser or @xmath81 for a circularly polarized laser .
this corresponds to a 10.4 gev output electron energy according to the scaling laws in ref .
the numerical parameters are shown in table [ tab : lwfapara1 ] .
the data from the boosted frame simulation is transformed back to the lab frame and it is compared against the data from the lab frame simulation . when modeling lwfa in quasi-3d geometry , whether it is in the lab frame or boosted frame , the accelerating ( @xmath63 ) and focusing fields ( @xmath64 and @xmath82 ) in the bubble are mainly in the @xmath62 modes of the em fields . on the other hand ,
the fields associated with the laser are associated with the @xmath65 mode of fields .
therefore , by keeping at least the @xmath83 modes the self - consistent evolution of the laser and wake fields can be examined when there is nearly azimuthal symmetry . for this comparison
we truncate the azimuthal harmonics keeping only the @xmath83 modes @xcite .
more modes can be kept in principle to study laser hosing and asymmetric spot size effects as well as to test the convergence of the results . in addition , the results and the needed truncation can be verified by comparing lwfa boosted frame simulation results from the full 3d and quasi-3d geometries . to verify the azimuthal mode truncation , we decompose the data from a full 3d osiris simulation into azimuthal harmonics and compare it against the corresponding quasi-3d simulation using the parameters listed in table [ tab : lwfapara0 ] . in fig .
[ fig:3dvsq3 ] , we plot the azimuthal decomposition of the 3d data for @xmath63 and @xmath64 at @xmath84 $ ] , and compare it against the corresponding quasi-3d data at the same time . for the @xmath83 modes , very good agreement is observed .
in addition , we plot the higher order @xmath85 modes from the 3d data in fig . [ fig : higherorder ] .
we can see that the higher order modes are at least one order of magnitude smaller than those of the @xmath86 modes , which verifies the truncation of azimuthal harmonics at @xmath83 in the quasi-3d simulations when the laser is nearly symmetric .
next , we compare data from a quasi-3d lwfa simulation in the lab frame against data lorentz transformed back to the lab frame from a quasi-3d simulation . a laser with normalized vector potential of @xmath61 ( converted to linear polarization ) with pulse length of 130 fs , and spot size of 45 @xmath3 propagates into a plasma column 20.8 cm long ( in the lab frame ) .
we use a boosted frame with @xmath87 , and use a moving window as described earlier that follows the relativistically drifting plasma .
a moving antenna injects the laser pulse into the plasma , and a damping region absorbs the em field at the rear end of the moving window . in the upper @xmath17 boundary of the simulation box we applied the perfectly - matched - layer boundary condition (
@xcite for more details ) .
the plasma density is uniform along the @xmath15 direction .
it is uniform in @xmath17 direction from @xmath88 $ ] ( where @xmath58 is the wave number of the laser in the lab frame ) , and then the density linearly ramps to zero at @xmath89 $ ] near the @xmath17 upper boundary ( an additional gap of @xmath90 $ ] is left between the @xmath17 upper plasma boundary and simulation box boundary ) .
the linear plasma density ramp is used to prevent reflection when the laser cross the upper @xmath17 plasma boundary into vacuum .
detailed simulation parameters are listed in table [ tab : lwfapara1 ] . as mentioned in section [
sect : simsetup ] , in the boosted frame each azimuthal mode of the em field is dumped frequently in time , while sparsely in space .
the results are transformed back to the lab frame in the post - processing . in figs .
[ fig : a4 ] ( a ) and ( b ) we present 2d plots of the @xmath62 mode of the @xmath63 field , as well as the @xmath65 mode of the @xmath64 fields , while in figs .
[ fig : a4 ] ( c ) and ( d ) we present the corresponding data from the lab frame simulation , respectively . as we can see from fig .
[ fig : a4 ] ( a ) and ( c ) the data from the two simulations agree well with each other , except for the area around the rear of the first bubble , which indicates that the two simulations give different self - injection results . on the other hand , the laser profiles from the two cases agree extremely well . in fig .
[ fig : a4 ] ( e ) line outs of the @xmath62 mode for @xmath63 at various time steps are plotted , and they show that in the transformed boosted frame data there is stronger beam loading , which indicates that more charge is self - injected into the bubble .
this is likely due to the difference in statistics between the lab frame simulation and boosted frame simulation . in the boosted frame
a macro - particle represents much more charge than in the lab frame .
particles in the boosted frame are `` fatter '' since the grid size in the boosted frame is larger , and this could affect the self - injection process . to confirm the differences are related to the self - injection process , we repeated the lab frame and boosted frame simulations in regimes with no self - injection , at @xmath73 ( converted to linear polarization ) , while keeping the other parameters listed in table [ tab : lwfapara1 ] the same . in fig .
[ fig : a3 ] ( a ) and ( b ) we show the line out of the wakefield at two different times in the lab frame , and in figs . [
fig : a3 ] ( c ) and ( d ) we show the corresponding amplitude of the laser profiles .
we see from fig .
[ fig : a3 ] that for this case where there is no self - injection in the lab frame simulation , the wake field results from the lab frame and boosted frame simulations agree very well .
it is challenging to accurately modeling the self - injection process in the lwfa blowout regime , and this is an area of future work .
as for the evolution of the laser profile , we plot the laser envelope and spot size obtained from the two cases from figs . [
fig : a4 ] ( e ) and ( f ) .
excellent agreement can be seen for the two times presented in fig .
[ fig : a4 ] .
excellent agreement is also seen for the evolution of the spot size , and laser amplitude of the laser driver as it propagates through the plasma column . in fig .
[ fig : spotamp ] we show a detailed time history of the laser spot size and amplitude at the position of the laser where its amplitude is largest . fig .
[ fig : spotamp ] clearly shows that very accurate results can be obtained when using lorentz boosted frame technique in quasi-3d geometry to study the evolution of laser driver in the plasma , with a large savings in cpu time .
evolution of the laser spot size and peak amplitude .
( a ) shows the comparison of laser spot size evolution as the laser propagates into the plasma for the two frames .
the laser spot size are defined at the location where the laser has the maximum amplitude .
the corresponding maximum laser amplitude evolution is shown in ( b).,scaledwidth=100.0% ]
in this paper , we described how it is possible to perform lwfa simulations in lorentz boosted frame using the quasi-3d algorithm .
the key to high fidelity lorentz boosted frame simulations in this geometry is the use of a hybrid yee - fft solver that solves the maxwell equation in @xmath16 space in the direction that the plasma drifts , while keeping the second order finite - difference operators in the transverse directions as in a conventional yee solver . using this maxwell solver we can then use the same strategies for eliminating nci in cartesian geometry to systematically eliminate the nci in the quasi-3d geometry . at the same time
all other features of osiris are also available including single core optimization and high parallel scalability . a current correction is applied to ensure the code rigorously conserves charge .
in addition , we analyzed the space - time area of the lab and boosted frame simulation data .
we showed how using a moving window which follows the drifting plasma in the boosted frame the further reduce the computational load .
we were able to combine lorentz boosted frame technique with quasi-3d algorithm , together with moving window technique to achieve unprecedented speed up for the modeling of lwfa .
we presented comparisons of lab frame against boosted frame simulation results for a 10 gev lwfa example that operates in the blowout regime .
it was shown that the evolution of the laser driver in the plasma can be very well reproduced by the boosted frame simulation .
we also found that the self - injection process is different in the boosted frame .
this is partly due to the difference in statistics between the simulations in the two frames since in the boosted frame each macro - particle represent many more real particles then in the corresponding lab frame simulation .
we found excellent agreement between the lab and boosted frame results for the wake fields when @xmath60 was reduced to avoid self - injection .
an area of future work is to systematically explore methods to accurately model self - injection process in the lorentz boosted frame simulation .
another area is the integration of this algorithm into our gpu and intel - phi enabled version of osiris .
this work was supported by us doe under grants de - sc0008491 , de - sc0008316 , de - fg02 - 92er40727 , and by the us national science foundation under the grant aci 1339893 , oci 1036224 , and by nsfc 11425521 , 11535006 , 11175102 , 11375006 , and tsinghua university initiative scientific research program , and by the european research council ( erc-2010-adg grant 267841 ) .
simulations were carried out on the ucla hoffman2 and dawson2 clusters , and on hopper cluster of the national energy research scientific computing center , and on blue waters cluster at national center for supercomputing applications at uiuc .
00 t. tajima , j.m .
dawson , phys .
43 ( 1979 ) 267 w. lu , et al .
rev . spec .
beams 10 ( 2007 ) 061301 .
gordon , w.b .
mori , t.m .
antonsen , ieee trans .
plasma sci .
28 ( 2000 ) 1135 .
p. mora , t.m .
antonsen , phys .
plasmas 4 ( 1997 ) 217 . c. huang , et al . , j. comp
217 ( 2006 ) 658 .
vay , phys .
98 , 130405 ( 2007 ) a.f .
lifschitz , et al .
, j. comput .
( 2009 ) 1803 .
a. davidson , et .
al . , j. comp .
281(2015 ) , 1063 .
al . , in proc .
advanced accelerator concept workshop 2014 , san jose , california , usa .
al . , in proc .
international particle accelerator conference 2015 , richmond , virginia , usa .
b. b. godfrey , j. comp .
15 , 504 ( 1974 ) p. yu et.al , in proc .
15th advanced accelerator concepts workshop , austin , tx , 2012 , in aip conf .
1507 , 416 ( 2012 ) . b. b. godfrey and j .- l .
vay , j. comp .
phys . , 248 , 33 - 46 ( 2013 ) x. xu , et .
al . , comp .
comm . , 184 , 25032514 ( 2013 )
godfrey , j .- l .
vay , i. haber , j. comput .
( 2014 ) 689 .
godfrey , j .- l
. vay , i. haber , j. comput .
( 2014 ) 1 .
al . , comp .
phys . comm . , 192 , 32 ( 2015 ) .
al . , comp .
, in press .
al . , j. comp .
266 , 124 ( 2014 ) .
k. yee , ieee trans .
antennas and propagation , 14 , 302 ( 1966 ) r. a. fonseca , et al .
, in : p.m.a .
sloot , et al .
( eds . ) , iccs , in : lect .
notes comput .
, vol . 2331 , 2002 , pp . 34351 ; r. hemker , ph.d .
thesis , ucla , 2000 ( arxiv:1503.00276 ) r. lehe , et .
al . , arxiv:1507.04790 ( 2015 ) .
t. esirkepov , comp .
comm . 135 ( 2001 ) , 144 . j. m. dawson , rev .
modern phys . , vol .
55 , no . 2 , 403 ( 1983 ) a. t. lin , j. m. dawson , h. okuda , phys .
fluids , 17 , 1995 ( 1974 ) . v.k .
decyk , comput .
commun . 177 ( 2007 ) 95 .
s. f. martins , r. a. fonseca , w. lu , w. b. mori and l. o. silva , nat .
phys . 6 , 311 ( 2010 ) j. -l .
vay , c. g. r. geddes , e. cormier - michel , d. p. grote , j. comp .
230 , 5908 ( 2011 ) .
s. f. martins , r. a. fonseca , l. o. silva , w. lu , w. b. mori , comp .
phys . comm . 181 , 869 ( 2010 )
vay , c.g.r .
geddes , e. esarey , c.b .
schroeder , w.p .
leemans , phys .
plasmas 18 ( 2011 ) 123103 . j .-
vay , et al . , in : proc .
14th advanced accelerator concepts workshop , annapolis , md , in : aip conference proceedings , 2010 , pp .
decker , w.b .
mori , phys .
72 ( 1994 ) 490 .
vay , c.g.r .
geddes , e. cormier - michel , phys .
plasmas 18 ( 2011 ) 030701 .
vay , phys .
plasmas 15 , 056701 ( 2008 ) .
j. p. boris , in proc .
4th conference on numerical simulation plasmas , naval research laboratory , washington , d.c .
, 1970 , pp .
367 . j .-
vay , i. haber , b. b. godfrey , j. comp .
( 2013 ) , 260 .
t. dalichaouch , et .
al . , bulletin of the american physical society 60 ( 2015 ) . | when modeling laser wakefield acceleration ( lwfa ) using the particle - in - cell ( pic ) algorithm in a lorentz boosted frame , the plasma is drifting relativistically at @xmath0 towards the laser , which can lead to a computational speedup of @xmath1 .
meanwhile , when lwfa is modeled in the quasi-3d geometry in which the electromagnetic fields and current are decomposed into a limited number of azimuthal harmonics , speedups are achieved by modeling three dimensional problems with the computation load on the order of two dimensional @xmath2 simulations . here
, we describe how to combine the speed ups from the lorentz boosted frame and quasi-3d algorithms .
the key to the combination is the use of a hybrid yee - fft solver in the quasi-3d geometry that can be used to effectively eliminate the numerical cerenkov instability ( nci ) that inevitably arises in a lorentz boosted frame due to the unphysical coupling of langmuir modes and em modes of the relativistically drifting plasma in these simulations . in addition , based on the space - time distribution of the lwfa data in the lab and boosted frame , we propose to use a moving window to follow the drifting plasma to further reduce the computational load .
we describe the details of how the nci is eliminated for the quasi-3d geometry , the setups for simulations which combine the lorentz boosted frame and quasi-3d geometry , the use of a moving window , and compare the results from these simulations against their corresponding lab frame cases .
good agreement is obtained , particularly when there is no self - trapping , which demonstrates it is possible to combine the lorentz boosted frame and the quasi-3d algorithms when modeling lwfa to achieve unprecedented speedups .
pic simulation , hybrid maxwell solver , relativistic plasma drift , numerical cerenkov instability , quasi-3d algorithm , lorentz boosted frame , moving window |
research into the optics of semicontinuous metal - dielectric films has been enjoying sustained interest due to a unique combination of novel physics and the practical applications offered by such composites .
it has been demonstrated , both theoretically and experimentally , that the electromagnetic ( em ) response of these structures is dominated by a non - trivial interplay between anderson - localized and delocalized surface plasmon polaritons.@xcite this results in unusual optical properties that include : greatly enhanced absorption , giant intensity fluctuations of local em fields , giant local chiral response , and strongly enhanced optical nonlinearities.@xcite in a related context , it has been recently shown that the reflectance of semicontinuous silver nanocomposites , chemically deposited on glass substrates , strongly depends on the direction of incident light.@xcite in particular , the reflectance of such a system irradiated from the substrate / film interface side can differ by as much as @xmath0 from its reflectance given film / air side incidence ( see fig.[filmfigure ] ) .
moreover , this large asymmetry in reflectance has been found to be extremely broadband , spanning most of the visible frequency spectrum .
for comparison , the reflectance asymmetry of thin , _ continuous _ silver films does not exceed @xmath1 when measured over the same range of optical frequencies , and does not exhibit any broadband characteristics .
it has been suggested that the origin of this large broadband asymmetry is in the enhanced optical absorbance which is often seen in percolation - type systems .
here we develop a quantitative description of the observed phenomenon .
the geometry of the system described in this manuscript is shown in figure [ filmfigure ] .
we approximate the silver percolation film as a uniform material with thickness @xmath2 [ in our calculations @xmath3 nm ( see section iii ) ] .
the microstructure of the film is characterized by the surface metal filling fraction @xmath4 ranging from @xmath5 for a bare glass substrate to @xmath6 for a substrate which is fully covered with metal . at the critical value @xmath7 , known as the _ percolation threshold _ , the @xmath8 conductivity response of the entire random metal - dielectric composite undergoes an insulator - conductor phase transition.@xcite the unique optical properties of our films are manifest particularly in the vicinity of the percolation threshold , and we therefore adopt the conventional description of the response as function of the parameter @xmath9 for the purpose of both modeling and data analysis .
all theoretical and experimental results in this work use @xmath10 , which correlates well with two dimensional site percolation on a square lattice ( @xmath11).@xcite , width=205 ] as mentioned above , the reflectance @xmath12 of the composite film , measured using light impinging from the air / metal interface , strongly differs from @xmath13 - the reflectance measured when light is incident from the substrate / film side . since the transmittance of our system , as the transmittance of any non - chiral homogeneous film is symmetric ( _ i.e. _ @xmath14)@xcite , the asymmetry in reflectance @xmath15 directly reflects the asymmetry in losses.@xcite as we show below , in contrast to vacuum - deposited percolation films , ( i ) @xmath16 as well as the computed combined losses exhibit a local minimum at @xmath17 , ( ii ) @xmath16 exhibits broadband response in the vicinity of @xmath18 , ( iii ) the reflectance exhibits a local maximum in the vicinity of @xmath17 , and ( iv ) the transmittance exhibits a local minimum near @xmath17 .
semi - continuous silver films with controllable filling fractions were deposited on microscope slides using a modified tollen s reaction as described previously.@xcite the amount of silver deposited on the substrates was controlled by monitoring deposition times , with reactions ranging between 1 - 6 h. ensuing deposition , coated substrates were rinsed with ultrapure water before being dried with filtered air and stored under nitrogen until tested .
figure [ semfig ] shows a scanning electron micrograph of a typical film with filling fraction @xmath19 .
these chemically deposited films appear as highly disordered polycrystalline aggregates , with large grain size distributions .
in addition , we note the non - uniform coating of the substrates by the metal , resulting in highly discontinuous morphologies . .
the scale bar is 500 nm .
[ semfig ] , width=302 ] optical reflectance and transmittance spectra were collected using a spectroscopic optical microscopy setup.@xcite a tungsten - halogen white - light source was used to illuminate the samples through an inverted microscope whose output was imaged on the entrance slit of a @xmath20 mm spectrometer with a resolution of 0.5 nm . a 10@xmath21 ( 0.25 n.a . )
objective was used to collect the normally incident light for spectroscopic imaging onto a liquid nitrogen - cooled detector . a high - reflectance mirror ( newport broadband supermirror , @xmath22 )
was used to normalize all signals .
in addition , data from @xmath23 m across the films were averaged to obtain each final trace , in order to eliminate spurious effects from local inhomogeneities in the rough films .
the spectral response of the film shown in fig.[semfig ] is depicted in fig.[rtafig ] .
many metal - dielectric composite systems are described by conventional effective medium techniques ( emts)@xcite by representing the composite as an effective homogeneous layer which successfully models the system s average optical properties .
however , it is known that the optical properties of these films close to the percolation threshold can not be adequately described by emts.@xcite the reason for the consistent failure of emts in this case is two - fold .
first , although the dimensions of the components in percolation films are much smaller than the free - space wavelength , the optical properties of the composites are dominated by the dynamics of resonant clusters that can be comparable in size to the wavelength .
second , as result of a _ dc _ metal - dielectric phase transition , the effective parameters of the percolation films in vicinity of @xmath24 become scale - dependent and therefore can not be described by quasi - static effective medium models . although some percolation films have been successfully described in terms of generalized ohm s law ( gol)@xcite , straightforward extensions of gol formalism to our system are not consistent with our experimental observations . .
solid lines represent the results of scaling theory calculations .
[ rtafig ] ] the only technique that explicitly accounts for the _ dc _ conductivity phase transition of percolation systems is known as _ scaling theory_.@xcite in this technique , the conductivity of the film is assumed to be explicitly dependent on the size of the cluster @xmath25 over which it is measured .
specifically , the average conductivity of a conductive cluster of size @xmath25 is given by @xmath26,\ ] ] and that of a dielectric ( insulating ) cluster is @xmath27.\ ] ] the expressions above explicitly assume that the conductivity of the conductive component of the film is given by the drude model , @xmath28 where @xmath29 is the dc conductivity , @xmath30 is the electron relaxation time , and @xmath31 is the angular frequency of the incident light .
the @xmath32 response of a dielectric film component is equivalent to that of a capacitor , @xmath33 where @xmath34 is the average capacitance between neighboring metal clusters .
the parameters @xmath35 @xmath36 , @xmath34 and @xmath37 coefficients are uniquely determined by the composition and micro - geometry of the percolation film .
the critical exponents for @xmath38d percolating films are @xmath39 and @xmath40.@xcite for @xmath41 percolation films are governed by dielectric conductivity , which is dominated by the capacitance coefficients @xmath42 and @xmath43 . for @xmath44 metallic conductivity ( governed by @xmath45 , @xmath46 ) dominates the optical properties of the system . despite the scale - dependence on the microscopic and mesoscopic levels
, the percolation film appears homogeneous when measured over a significantly large area .
the transition from the scale - dependent to the homogeneous @xmath8 response occurs at the scale known as the _ correlation length _ , @xmath47 , that characterizes the typical cluster size .
in the vicinity of the percolation threshold , the correlation length diverges as @xmath48 the constant @xmath36 represents the smallest metal cluster size , which occurs at @xmath49 . at finite frequencies ,
the oscillatory motion of electrons within conducting clusters leads to the length scale correction of the homogeneous response of the system , @xmath50 where @xmath51 , @xmath52 , and the free space wavelength is given by @xmath53.@xcite the @xmath32 conductivity of the percolation films , calculated using the expressions above can be directly related to an effective film index , which along with the film thickness @xmath2 can be used to determine the macroscopic optical properties of the film , including @xmath54 and @xmath55 . in our calculations
, we use the technique introduced in ref . [ ] . in this approach ,
the optical properties of the film are calculated as a weighted average of conductive ( dielectric ) film contributions , where the average conductivities are given by eqs.([sigmam ] ) and ( [ sigmad ] ) respectively to yield @xmath56 p(z ) dz \\
r_i = \int_{0}^{\infty } [ f r_{i,\sigma}(z \sigma_m ) + ( 1-f ) r_{i,\sigma } ( z \sigma_d ) ] p(z ) dz\end{aligned}\ ] ] where the parameter @xmath57,\ ] ] is the metal occupation probability . for small surface metal concentrations , @xmath58
$ ] , the occupation probability @xmath59 . when @xmath60 $ ] the occupation probability @xmath61 . for intermediate surface metal concentrations centered at @xmath7 with full - width @xmath62 ,
the occupation probability varies linearly as a function of p from unoccupied ( @xmath63 ) to occupied ( @xmath64 ) . as shown by the above inequalities , the range of metal surface coverage values for which scaled metal and dielectric optical properties
are averaged depends non - trivially on the correlation length , applied frequency , and film geometry .
the function @xmath65 gives the distribution of the conductivities of conductive [ dielectric ] clusters around their mean values given by eq.([sigmam ] ) [ eq.([sigmad ] ) ] .
following ref . [ ] and [ ] we assume that @xmath65 is adequately described by a log - normal distribution function with standard deviation of @xmath66 .
integrating over all scaled conductivities averages out the length dependent optical conductivity and allows for percolation films to be modeled by the contributions from planar homogeneous constituent layers .
the homogeneous - layer optical properties are given by,@xcite @xmath67 where the glass substrate index is @xmath68 , the effective film index is @xmath69 , and the phase parameter is @xmath70 .
as noted , all critical exponents in the expressions above are universal for all @xmath38d percolating networks , while the parameters @xmath71 , @xmath72 , and @xmath36 are unique for a given percolation film.@xcite in our calculations , we use @xmath73sec@xmath74 , frequency dependent relaxation time@xcite @xmath75 , where @xmath76fs , @xmath77fs , @xmath78 , @xmath79 , and @xmath80 nm .
a comparison of the experimentally obtained spectral response of the silver films with the predictions of scaling theory is shown in fig.[rtafig ] .
it is seen that both the broadband nature of the reflectance asymmetry and its non - monotonic behavior near the percolation threshold are are well reproduced by the theoretical model , as demonstrated in fig.[deltar ] .
we note however , that the model fails for the large metal concentrations @xmath81@xcite where the structure of the composite becomes substantially three - dimensional and can not be treated as a thin homogeneous film . to further illustrate the robustness of the presented technique we show in fig.[rtpdata ] a comparison of experimentally measured values of @xmath12 and @xmath82 , as well as the losses ( computed as @xmath83 , ) with our theoretical model . as mentioned before ,
both theoretical and experimental results clearly show that despite a strong reflectance asymmetry , the transmittance of the films remains symmetric .
therefore the asymmetry in reflectance is directly related to the asymmetry in losses ( @xmath84).@xcite ) for various incident wavelengths .
black circles @xmath85 nm , green triangles @xmath86 nm , red boxes @xmath87 nm . corresponding colored solid lines ( black solid @xmath85 nm , green short - dashed @xmath86 nm , red long - dashed @xmath87 nm ) represent the results of scaling theory reflectance calculations .
the inset shows the change in reflectance over the entire surface coverage range .
note that the 2d scaling model fails for large metal concentrations , where the three - dimensional structure of the composite dominates the optical response .
[ deltar ] ] nm ( solid line ) , @xmath88 nm ( dashed line ) , and @xmath89 nm ( dotted line ) . for all graphs
the incident wavelength is @xmath87 nm . [ rtpdata ] ] we now examine the loss ( or , alternately the reflectance ) in more detail .
we note that several previous experiments as well as theoretical models have observed absorption maxima in the vicinity of @xmath7 .
in contrast , our experimental data clearly demonstrate a local _ minimum _ in losses near the percolation threshold , as seen in fig.[rtpdata](e ) . additionally we observe a local _ maximum _ in reflectance and a local _ minimum _ in transmittance near @xmath7 , as seen in fig.[rtpdata](a , c ) .
we suggest that this anomalous behavior stems from the dramatically reduced correlation length in our solution - derived percolation films .
the smallest metallic particle size produced in our experiments is on the order of 2 nm , in contrast to 10 nm reported in ref .
in addition , the optical response of solution - derived metals is typically affected by the reduced electron mean free path@xcite , which may further reduce the effective particle size , described by the parameter @xmath36 .
figure [ rtpdata ] demonstrates the evolution of optical properties of percolation systems when the correlation length is reduced , corresponding to a change in @xmath36 from 10 nm to 2 nm .
it is clearly seen that at @xmath90 the absorption reaches a local minimum in the vicinity of @xmath7 , while at @xmath89 nm , the system recovers the absorption maximum at @xmath7 , as observed in previous studies . the dramatic difference between correlation lengths in our _
solution - derived _
films@xcite and other _ vacuum - deposited _
counterparts@xcite is consistent with the difference in fabrication technique : while thermal deposition under vacuum typically yields uniform metal films with almost perfect ohmic contacts between adjacent grains , solution - based deposition is routinely associated with quantitatively weaker contacts between conducting grains .
the latter results in films with reduced electron mean - free - paths and lower correlation lengths .
we have developed an analytical description of the phenomenon of broadband asymmetric reflection in percolation composites .
the developed technique , based on scaling theory , is not only capable of describing the spectral response of our films , but also explains that the reduced correlation length in our solution - derived composites is the primary cause of the experimentally observed anomalous optical properties near the percolation threshold .
our work demonstrates that the correlation length is an important factor that fundamentally affects the optical properties of percolation composites in the vicinity of the percolation threshold .
this research is sponsored in part by onr ( grant no . n00014 - 07 - 1 - 0457 ) , nsf ( grant no .
eccs-0724763 ) , prf ( grant no .
43924-g10 ) , nsf dmr-02 - 39273 , nsf dmr-08 - 04433 . | we develop a quantitative description of giant asymmetry in reflectance , recently observed in semicontinuous metal films . the developed scaling - theory based technique reproduces the spectral properties of semicontinuous composites , as well as provides insight into the origin of experimentally observed loss , reflectance , and transmittance anomalies in the vicinity of the percolation threshold . |
let @xmath0 be a closed orientable @xmath1-manifold of heegaard genus two with a non - trivial jsj - decomposition and @xmath2 be a finite group of orientation - preserving diffeomorphisms acting on @xmath0 which preserves each handlebody of heegaard splitting and each piece of the jsj - decomposition of @xmath0 .
let @xmath3 be the heegaard splitting and @xmath4 be the union of the jsj - tori . in @xcite
, there are two conditions _ `` condition a '' _ and _ `` condition b '' _ for the heegaard splitting and the jsj - tori , where each determines @xmath2 to be @xmath14 or @xmath6 up to isomorphism .
but these two conditions have strong restrictions - @xmath15 must consist of at most two disks or annuli .
how can we generalize these conditions ? in this article , we will determine the possible isomorphism types of @xmath2 by the following theorem .
[ main - theorem ] let @xmath0 be a closed orientable @xmath1-manifold with a genus two heegaard splitting @xmath16 and a non - trivial jsj - decomposition , where all components of the intersection of the jsj - tori and @xmath17 are not @xmath8-parallel in @xmath17 for @xmath18 . if @xmath2 is a finite group of orientation - preserving smooth actions on @xmath0 which preserves each handlebody of the heegaard splitting and each piece of the jsj - decomposition of @xmath0 , then @xmath19 or @xmath6 unless @xmath11 has three or more disk components for @xmath12 or @xmath13 .
acknowledgement this work was supported by the korea science and engineering foundation ( kosef ) grant funded by the korea government ( most ) ( no .
r01 - 2007 - 000 - 20293 - 0 ) .
we define two conditions for the heegaard splitting and the jsj - tori as follows . 1
. @xmath20 consists of disks , and the number of disks is at most two , 2 .
@xmath20 consists of annuli , and the number of annuli is at most two , + where we assume that each component of @xmath20 is not @xmath8-parallel in @xmath21 .
[ theorem - kjs2](@xcite , theorem 1.1 ) let @xmath0 be a closed orientable @xmath1-manifold with a genus two heegaard splitting @xmath16 and a non - trivial jsj - decomposition , where all components of the intersection of the jsj - tori and @xmath17 are not @xmath8-parallel in @xmath17 for @xmath18 . if @xmath2 is a finite group of orientation - preserving diffeomorphisms acting on @xmath0 which preserves each handlebody of the heegaard splitting and each piece of the jsj - decomposition of @xmath0 , then
@xmath19 or @xmath6 if @xmath11 consists of at most two disks or at most two annuli .
( @xcite , lemma 2.8)[lemma - gm-12 ] let @xmath2 be a finite group of non - trivial orientation preserving diffeomorphisms which acts on a genus two handlebody @xmath22 .
suppose that @xmath23 is a set of properly embedded surfaces in @xmath22 where @xmath23 cuts @xmath22 into two solid tori @xmath21 , @xmath24 and a @xmath1-ball @xmath25 .
if @xmath2 preserves each @xmath17 for @xmath26 , or @xmath2 preserves @xmath25 but exchanges @xmath21 and @xmath24 , then @xmath27 or @xmath6 .
( @xcite , lemma 2.9)[lemma - gm-2 ] let @xmath2 be a finite group of non - trivial orientation preserving diffeomorphisms which acts on a genus two handlebody @xmath22 .
suppose that @xmath23 is a set of properly embedded surfaces in @xmath22 where @xmath23 cuts @xmath22 into a @xmath1-ball @xmath21 and a solid torus @xmath25 .
if @xmath2 preserves both @xmath21 and @xmath25 , then @xmath28 or @xmath6 .
( @xcite , lemma 2.11)[lemma - gm - an-1 ] let @xmath2 be a finite group of non - trivial orientation preserving diffeomorphisms which acts on a genus two handlebody @xmath22 .
suppose that @xmath23 is a set of properly embedded surfaces in @xmath22 where @xmath23 cuts @xmath22 into two connected components @xmath21 and @xmath25 , where both @xmath21 and @xmath25 are noncontractible in @xmath22 , and @xmath21 or @xmath25 contains a meridian disk of @xmath22 . if @xmath2 preserves both @xmath21 and @xmath25 , then @xmath28 or @xmath6 .
( @xcite , lemma 2.12)[lemma - gm - an-2 ] let @xmath2 be a finite group of non - trivial orientation preserving diffeomorphisms which acts on a genus two handlebody @xmath22 .
suppose that @xmath23 is a set of properly embedded surfaces in @xmath22 where @xmath23 cuts @xmath22 into two solid tori @xmath21 and @xmath24 , and a genus two handlebody @xmath25 , where 1 .
@xmath29 is a genus two handlebody , 2 .
both @xmath21 and @xmath25 are non - contractible in @xmath30 and @xmath22 , and 3 .
there is a meridian disk @xmath31 in @xmath22 such that @xmath32 and @xmath33 is a non - empty set of meridian disks of @xmath30 .
a. @xmath34 cuts @xmath22 into a solid torus @xmath21 and a genus two handlebody @xmath25 and there is a complete system of meridian disks @xmath35 of @xmath25 such that @xmath36 and @xmath37 is an essential arc of @xmath34 ( see the upper of figure [ fig - lemma - ko-3 - 2 ] , ) or[lemma - ko-3 - 2 - 1 ] b. @xmath34 cuts @xmath22 into a genus two handlebody @xmath30 and there is a complete system of meridian disks @xmath35 of @xmath30 such that @xmath38 is an essential arc of @xmath34 ( see the lower of figure [ fig - lemma - ko-3 - 2].)[lemma - ko-3 - 2 - 2 ] a. [ lemma - ko-3 - 4 - 1]@xmath42 cuts @xmath22 into a solid torus @xmath21 and a genus two handlebody @xmath25 .
then @xmath43 , @xmath44 and there is a complete system of meridian disks @xmath35 of @xmath25 such that @xmath45 ( @xmath46 ) and @xmath47 ( @xmath18 ) is an essential arc of @xmath48 , b. [ lemma - ko-3 - 4 - 2]@xmath42 cuts @xmath22 into two solid tori @xmath21 , @xmath25 and a genus two handlebody @xmath24
. then @xmath49 , @xmath50 , @xmath51 and there is a complete system of meridian disks @xmath52 of @xmath24 such that @xmath45 @xmath53 and @xmath47 ( @xmath18 ) is an essential arc of @xmath48 or c. [ lemma - ko-3 - 4 - 3]@xmath42 cuts @xmath22 into a solid torus @xmath21 and a genus two handlebody @xmath25 .
then @xmath54 ( @xmath55 or @xmath13 , say @xmath56 ) , @xmath57 , @xmath58 and there is a complete system of meridian disks @xmath52 of @xmath25 such that @xmath59 is an essential arc of @xmath40 and @xmath60 ( @xmath18 ) is an essential arc of @xmath48 , see figure [ fig - lemma - ko-3 - 4 ] . since a jsj - torus is incompressible in @xmath0 , all components of @xmath15 are incompressible in @xmath21 .
so we will assume the incompressibility of all components of @xmath15 in @xmath21 . by the assumption of theorem [ main - theorem ]
, we assume that all components of @xmath15 are not @xmath8-parallel in @xmath21 .
since @xmath2 preserves @xmath15 and there is only one disk component @xmath31 in @xmath15 , @xmath2 must preserve @xmath31 .
so @xmath2 also preserves a small product neighborhood of @xmath31 in @xmath21 , say @xmath61 , where @xmath62 and @xmath63 meets no component of @xmath15 other than @xmath31 .
say that @xmath64 and @xmath65 .
then @xmath52 cuts @xmath21 into a @xmath1-ball @xmath66 and a solid torus ( when @xmath67 is non - separating in @xmath21 , ) or into a @xmath1-ball @xmath66 and two solid tori ( when @xmath67 is separating in @xmath21 . )
since @xmath2 preserves @xmath35 and @xmath66 is the only @xmath1-ball component of @xmath68 , @xmath2 preserves @xmath66 .
so @xmath19 or @xmath6 by lemma [ lemma - gm-2 ] or lemma [ lemma - gm-12 ] . since @xmath67 and @xmath69 are the only disk components in @xmath15 , @xmath2 must preserve @xmath52 in @xmath21
so @xmath2 preserves @xmath68 .
since the @xmath1-ball @xmath66 between @xmath67 and @xmath69 is the only @xmath1-ball component of @xmath68 ( see @xmath67 and @xmath69 in the proof of lemma [ lemma - d1cp ] ) , @xmath2 preserves @xmath66 .
so @xmath19 or @xmath6 using the arguments in the proof of lemma [ lemma - d1cp ] . + * case 2 . *
@xmath67 and @xmath69 are non - parallel in @xmath21 .
if one of @xmath67 and @xmath69 , say @xmath67 , is separating in @xmath21 , then @xmath69 is non - separating in @xmath21 .
so @xmath52 cuts @xmath21 into a @xmath1-ball @xmath70 and a solid torus @xmath71 .
since @xmath2 preserves @xmath52 , @xmath2 also preserves @xmath72 . but @xmath70 is not homeomorphic to @xmath71 , @xmath2 must preserves both @xmath70 and @xmath71 .
so @xmath19 or @xmath6 by lemma [ lemma - gm-2 ] .
if both of @xmath67 and @xmath69 are non - separating in @xmath21 , then @xmath35 cuts @xmath21 into a @xmath1-ball @xmath66 .
let @xmath73 and @xmath74 be two copies of @xmath75 for @xmath18 in @xmath76 after cutting @xmath21 along @xmath67 and @xmath69 , and @xmath77 be a non - disk component of @xmath15 . since any properly embedded incompressible surface in @xmath78 must be @xmath8-parallel , @xmath77 is @xmath8-parallel in @xmath66 .
moreover , no curve of @xmath79 is parallel to some @xmath80 for @xmath81 ( @xmath77 is incompressible in @xmath21 and does not intersect both @xmath67 and @xmath69 . )
let @xmath82 be @xmath83 .
then @xmath82 is a disk with three holes .
so each curve of @xmath79 must separate each two of @xmath84 .
in addition , the curves of @xmath79 must be pairwise parallel in @xmath82 after one of them is determined ( see figure [ bprime ] . ) since @xmath77 is a non - disk connected surface and all curves in @xmath79 are pairwise parallel in @xmath82 , @xmath79 must consist of two curves .
so @xmath77 is @xmath8-parallel to an annulus @xmath85 in @xmath66 , where @xmath34 is bounded by these two curves , i.e. @xmath77 is @xmath8-parallel in @xmath21 .
but this contradicts the hypothesis .
let @xmath77 be a non - annulus component of @xmath15 , @xmath88 be a jsj - torus which contains @xmath77 .
then @xmath77 is a disk with @xmath89 holes ( @xmath90 ) or a torus with @xmath91 holes ( @xmath92 ) since @xmath93 .
+ * claim . * _ @xmath94 must contain at least one disk component . _
+ so @xmath87 contains at least one disk component by the above claim .
since @xmath2 also preserves @xmath25 , if we substitute @xmath25 for @xmath21 , then we get @xmath19 or @xmath6 by lemma [ lemma - d1cp ] or lemma [ lemma - d2cp3morecp ] unless @xmath87 has three or more disk components .
now we remains only the cases when @xmath15 consists of annuli . since we already proved the cases when @xmath15 consists of at most two annulus components in `` condition b '' , it is sufficient to prove the cases @xmath15
consists of at least three annulus components .
( @xcite , lemma 3.5)[lemma - ko-3 - 5 ] let @xmath97 be a system of pairwise disjoint , non - parallel essential annuli in the genus two handlebody @xmath21 .
then @xmath98 cuts @xmath21 into two solid tori @xmath70 , @xmath71 and a genus two handlebody @xmath99 which satisfies 1 .
@xmath100 ( @xmath55 , @xmath13 or @xmath1 , say @xmath1 ) , @xmath39 , @xmath101 , @xmath39 , @xmath40 , @xmath102 , 2 .
there is a complete system of meridian disks @xmath52 of @xmath99 such that @xmath45 @xmath53 and @xmath47 ( @xmath18 ) is an essential arc of @xmath48 and 3 .
there is a meridian disk @xmath103 of @xmath71 such that @xmath104 ( @xmath26 ) is an essential arc of @xmath48 .
suppose that there is a system of pairwise disjoint , non - parallel essential four annuli in a genus two handlebody , say @xmath105 .
@xmath106 cuts @xmath21 into a genus two handlebody @xmath99 and two solid tori @xmath70 and @xmath71 as lemma [ lemma - ko-3 - 5 ] .
we know that any properly embedded incompressible annulus in a solid torus must be @xmath8-parallel in the solid torus from lemma 3.1 of @xcite .
if @xmath107 is embedded in @xmath70 or @xmath71 , then @xmath107 must be @xmath8-parallel in @xmath70 or @xmath71 . but
@xmath107 can not be parallel into any one of @xmath39 , @xmath40 and @xmath108 , @xmath107 must be @xmath8-parallel to a subset of @xmath109 or @xmath110 ( @xmath111 is in @xmath112 or @xmath113 . )
so @xmath107 is @xmath8-parallel in @xmath21 , this is a contradiction .
so @xmath107 is embedded in @xmath99 .
moreover , @xmath107 must be satisfy ( [ lemma - ko-3 - 2 - 1 ] ) of lemma [ lemma - ko-3 - 2 ] in @xmath99 ( @xmath107 must be properly embedded not only in @xmath99 , but also in @xmath21 , see figure [ fig - lemma - ko-3 - 2 ] and figure [ fig - threeannuli ] . )
say that @xmath41 bounds @xmath99
. then @xmath107 must be parallel to @xmath39 or @xmath40 in @xmath99 .
( @xmath107 is properly embedded not only in @xmath99 , but also in @xmath21 , too .
see figure [ fig - threeannuli ] . )
so we get a contradiction .
so we assume that @xmath15 consists of exactly three , non - parallel annuli @xmath39 , @xmath40 and @xmath108 .
use the same notations of @xmath39 , @xmath40 , @xmath108 , @xmath70 , @xmath71 and @xmath99 as lemma [ lemma - ko-3 - 5 ] .
since @xmath99 is the only genus two handlebody among @xmath70 , @xmath71 and @xmath99 , @xmath2 preserves @xmath99 , i.e. @xmath2 preserves @xmath114 .
so @xmath2 also preserves @xmath71 ( so also preserves @xmath70 . )
let @xmath30 be @xmath115 .
then @xmath30 is a genus two handlebody and @xmath116 . since @xmath2 preserves @xmath30 and @xmath70 , we get @xmath19 or @xmath6 by lemma [ lemma - gm - an-1 ] . by lemma [ lemma - ko-3 - 2 ] ,
@xmath114 cuts @xmath21 into a solid torus and a genus two handlebody ( if @xmath39 is non - separating in @xmath21 ) or two solid tori and a genus two handlebody ( if @xmath39 is separating in @xmath21 ) as figure [ figure - add - ad3 ] . if some component of @xmath15 except @xmath39 and @xmath40 intersects the solid torus bounded by @xmath39 and @xmath40 , then it must be parallel to @xmath39 ( also to @xmath40 ) in @xmath21 .
so if we choose nearest parallel annuli for @xmath39 and @xmath40 , then we can assume that no component of @xmath15 except @xmath39 and @xmath40 intersects the solid torus bounded by @xmath39 and @xmath40 .
if @xmath2 preserves the solid torus between @xmath39 and @xmath40 , then we get @xmath19 or @xmath6 using the same arguments in case 3 of section 6 in @xcite .
so now we assume that @xmath2 does not preserve the solid torus between @xmath39 and @xmath40 .
since there is no system of properly embedded , pairwise disjoint , non - parallel essential four or more annuli in a genus two handlebody from claim of lemma [ lemma-3moreannulicp ] , we can divide the proof into three cases as follows .
* @xmath15 consists of annuli parallel to @xmath39 .
if the number of components of @xmath15 is @xmath89 , then @xmath15 cuts @xmath21 into a genus two handlebody @xmath70 and @xmath117 solid tori ( if @xmath39 is non - separating , see ( ii ) of lemma [ lemma - ko-3 - 2 ] . ) or into a genus two handlebody @xmath70 and @xmath89 solid tori ( if @xmath39 is separating , see ( i ) of lemma [ lemma - ko-3 - 2 ] . ) in any case , the only genus two handlebody component @xmath70 of @xmath118 is preserved by @xmath2 . in the former case , the big solid torus @xmath119 must be preserved by @xmath2 , so we get @xmath19 or @xmath6 by lemma [ lemma - gm - an-1 ] . in the latter case ,
each of adjacently attached @xmath117 solid tori among @xmath89 solid tori meets two annuli of @xmath15 in its closure and the last solid torus @xmath99 meets only one annulus of @xmath15 in its closure ( the naming of @xmath99 is the same as ( i ) of figure [ figure - add - ad3 ] . ) therefore , @xmath99 must be preserved by @xmath2 , i.e. the big solid torus @xmath120 must be preserved by @xmath2 .
since @xmath121 is a genus two handlebody , we get @xmath19 or @xmath6 by lemma [ lemma - gm - an-1 ] ( in ( i ) of figure[figure - add - ad3 ] , we can say that the big solid torus @xmath122 is @xmath71 . ) * case 2 .
* the parallel class of @xmath15 in @xmath21 consists of two elements .
let the elements of the class be @xmath123 $ ] and @xmath124 $ ] .
then @xmath125 satisfies one of three conclusions of lemma [ lemma - ko-3 - 4 ] .
if @xmath125 satisfies ( [ lemma - ko-3 - 4 - 1 ] ) or ( [ lemma - ko-3 - 4 - 2 ] ) of lemma [ lemma - ko-3 - 4 ] , then we can choose a genus two handlebody component @xmath126 in @xmath127 which meets only two annuli of @xmath15 in its closure .
moreover , the other components of @xmath118 are all solid tori .
so @xmath2 preserves @xmath126 , and @xmath2 also preserves the big solid torus @xmath128 if @xmath125 satisfies ( [ lemma - ko-3 - 4 - 1 ] ) of lemma [ lemma - ko-3 - 4 ] or the set of two big solid tori @xmath128 if @xmath125 satisfies ( [ lemma - ko-3 - 4 - 2 ] ) of lemma [ lemma - ko-3 - 4 ] .
so we get @xmath19 or @xmath6 using lemma [ lemma - gm - an-1 ] or lemma [ lemma - gm - an-2 ] . if @xmath125 satisfies ( [ lemma - ko-3 - 4 - 3 ] ) of lemma [ lemma - ko-3 - 4 ] , then @xmath127 consists of a genus two handlebody and three or more solid tori ( one of the solid tori , say @xmath129 is directly induced by ( iii ) of figure [ fig - lemma - ko-3 - 4 ] , and the others are induced from the orbit of the solid torus between @xmath39 and @xmath40 by @xmath2 . )
since @xmath129 is the only solid torus component of @xmath127 which meets only one annulus of @xmath15 in its closure , say @xmath34 , and each of the other solid tori meets two annuli of @xmath15 in its closure , @xmath2 preserves @xmath129 . since @xmath130 is a genus two handlebody and preserved by @xmath2 , we get @xmath19 or @xmath6 by lemma [ lemma - gm - an-1 ] by the decomposition @xmath131 . * case 3 . * the parallel class of @xmath15 consists of three elements
. then @xmath15 cuts @xmath21 into a genus two handlebody @xmath132 and solid tori ( see figure [ fig - threeannuli ] . ) in particular , there is only one solid torus component of @xmath127 which meets only one annulus of @xmath15 as @xmath70 of figure [ fig - threeannuli ] .
let this solid torus be @xmath70 , then @xmath70 is preserved by @xmath2 . since the only genus two handlebody component ,
say @xmath99 , is also preserved by @xmath2 , @xmath2 preserves the big solid torus @xmath133 .
so we get @xmath19 or @xmath6 using the same arguments in lemma [ lemma-3moreannulicp ] . | let @xmath0 be a closed orientable @xmath1-manifold of heegaard genus two with a non - trivial jsj - decomposition and @xmath2 be a finite group of orientation - preserving diffeomorphisms acting on @xmath0 which preserves each handlebody of heegaard splitting and each piece of the jsj - decomposition of @xmath0 .
let @xmath3 be the heegaard splitting and @xmath4 be the union of the jsj - tori . in this article , we prove that @xmath5 or @xmath6 if all components of @xmath7 are not @xmath8-parallel in @xmath9 for @xmath10 unless @xmath11 has three or more disk components for @xmath12 or @xmath13 . |
this work was partially supported by the ministry of science and technology of taiwan r.o.c .
, under grant most-103 - 2112-m-006 - 004-my3 ( chc ) .
g. c. branco , p. m. ferreira , l. lavoura , m. n. rebelo , m. sher and j. p. silva , phys .
rept . * 516 * , 1 ( 2012 ) [ arxiv:1106.0034 [ hep - ph ] ] . y. h. ahn and c. h. chen , phys .
b * 690 * , 57 ( 2010 ) [ arxiv:1002.4216 [ hep - ph ] ] .
g. aad _ et al . _ [ atlas collaboration ] , phys .
b * 716 * , 1 ( 2012 ) [ arxiv:1207.7214 [ hep - ex ] ] .
s. chatrchyan _ et al . _
[ cms collaboration ] , phys .
b * 716 * , 30 ( 2012 ) [ arxiv:1207.7235 [ hep - ex ] ] .
m. magg and c. wetterich , phys .
b * 94 * , 61 ( 1980 ) ; g. lazarides , q. shafi and c. wetterich , nucl .
b * 181 * , 287 ( 1981 ) ; r. n. mohapatra and g. senjanovic , phys .
d * 23 * , 165 ( 1981 ) ; e. ma and u. sarkar , phys .
lett . * 80 * , 5716 ( 1998 ) [ hep - ph/9802445 ] .
w. konetschny and w. kummer , phys .
b * 70 * , 433 ( 1977 ) ; j. schechter and j. w. f. valle , phys . rev .
d * 22 * , 2227 ( 1980 ) ; t. p. cheng and l. -f .
li , phys .
d * 22 * , 2860 ( 1980 ) ; s. m. bilenky , j. hosek and s. t. petcov , phys .
b * 94 * , 495 ( 1980 ) . c. h. chen and t. nomura , phys .
d * 90 * , no . 7 , 075008 ( 2014 ) [ arxiv:1406.6814 [ hep - ph ] ] . c. h. chen and t. nomura , phys .
d * 91 * , 035023 ( 2015 ) [ arxiv:1411.6412 [ hep - ph ] ] .
r. benbrik , c. h. chen and t. nomura , phys .
d * 93 * , no .
9 , 095004 ( 2016 ) [ arxiv:1511.08544 [ hep - ph ] ] .
s. chatrchyan _ et al . _
[ cms collaboration ] , eur .
j. c * 72 * , 2189 ( 2012 ) [ arxiv:1207.2666 [ hep - ex ] ] .
g. aad _ et al . _ [ atlas collaboration ] , eur .
j. c * 72 * , 2244 ( 2012 ) [ arxiv:1210.5070 [ hep - ex ] ] .
g. aad _ et al . _ [ atlas collaboration ] , jhep * 1206 * , 039 ( 2012 ) [ arxiv:1204.2760 [ hep - ex ] ] .
g. aad _ et al . _ [ atlas collaboration ] , jhep * 1303 * , 076 ( 2013 ) [ arxiv:1212.3572 [ hep - ex ] ] .
g. aad _ et al . _ [ atlas collaboration ] , eur .
j. c * 73 * , no . 6 , 2465 ( 2013 ) [ arxiv:1302.3694 [ hep - ex ] ] . s. chatrchyan _ et al .
_ [ cms collaboration ] , jhep * 1207 * , 143 ( 2012 ) [ arxiv:1205.5736 [ hep - ex ] ] .
g. aad _ et al . _ [ atlas collaboration ] , jhep * 1503 * , 088 ( 2015 ) [ arxiv:1412.6663 [ hep - ex ] ] .
v. khachatryan _ et al . _
[ cms collaboration ] , jhep * 1512 * , 178 ( 2015 ) [ arxiv:1510.04252 [ hep - ex ] ] . v. khachatryan _ et al .
_ [ cms collaboration ] , jhep * 1511 * , 018 ( 2015 ) [ arxiv:1508.07774 [ hep - ex ] ] .
a. g. akeroyd , s. moretti , k. yagyu and e. yildirim , arxiv:1605.05881 [ hep - ph ] .
a. g. akeroyd _ et al .
_ , arxiv:1607.01320 [ hep - ph ] .
a. arhrib , r. benbrik and s. moretti , arxiv:1607.02402 [ hep - ph ] . c. degrande , r. frederix , v. hirschi , m. ubiali , m. wiesemann and m. zaro , arxiv:1607.05291 [ hep - ph ] .
v. khachatryan _ et al . _
[ cms collaboration ] , phys .
b * 759 * , 369 ( 2016 ) [ arxiv:1603.02991 [ hep - ex ] ] .
a. belyaev , n. d. christensen and a. pukhov , comput .
commun . * 184 * , 1729 ( 2013 ) [ arxiv:1207.6082 [ hep - ph ] ] .
m. nadolsky , h. l. lai , q. h. cao , j. huston , j. pumplin , d. stump , w. k. tung and c .- p .
yuan , phys .
d * 78 * , 013004 ( 2008 ) [ arxiv:0802.0007 [ hep - ph ] ] .
k. a. olive _ et al . _
[ particle data group collaboration ] , chin .
c * 38 * , 090001 ( 2014 ) .
j. alwall _ et al .
_ , jhep * 1407 * , 079 ( 2014 ) [ arxiv:1405.0301 [ hep - ph ] ] .
a. alloul , n. d. christensen , c. degrande , c. duhr and b. fuks , comput .
commun . * 185 * , 2250 ( 2014 ) [ arxiv:1310.1921 [ hep - ph ] ] .
t. sjostrand , s. mrenna , p. z. skands , jhep * 0605 * , 026 ( 2006 ) .
c. s. deans [ nnpdf collaboration ] , arxiv:1304.2781 [ hep - ph ] . | charged higgs in the type ii two - higgs - doublet model ( thdm ) has been bounded to be above a few hundred gev by the radiative @xmath0 decays . a higgs triplet extension of the thdm not only provides an origin of neutrino masses and a complete new doubly charged higgs decay pattern , but also achieves a light charged higgs with a mass of @xmath1 gev through the new scalar couplings in the scalar potential . it is found that this light charged higgs decays depend on its mass @xmath2 , @xmath3 , and mixing effect @xmath4 : at @xmath5 , if @xmath6 , @xmath7 , @xmath8 , and @xmath9 are the main decay modes ; if @xmath10 , the main decay modes then are @xmath11 and @xmath9 ; and at @xmath12 , @xmath9 mode overwhelms the other decays . in any circumstance , if @xmath13 is the dominant decay , we find that the atlas and cms recent upper bounds on the product @xmath14 can be directly applied and give a strict constraint on the correlation of @xmath2 and @xmath4 .
we study the potential to observe the charged higgs ; and it is found that the significance for discovering the charged higgs through @xmath15 is much lower than that through @xmath16 . with the luminosity of 100 fb@xmath17 at @xmath18 tev , the significance of the signal @xmath19 can reach around @xmath20 for @xmath21 .
two higgs doublet model ( thdm ) is one of minimal extensions of the standard model ( sm ) in the supersymmetric ( susy ) and non - susy frameworks .
the scalar bosons in the model are two cp - even scalars @xmath22 , one cp - odd pseudoscalar @xmath23 , and two charged higgs particles @xmath24 . to avoid the flavor - changing neutral currents at the tree level , a discrete symmetry is usually imposed .
thus , according to the couplings to the fermions in the yukawa sector , several types non - susy thdm are classified in the literature @xcite ; therefore , different types thdm may have different constraints on their masses and couplings . with the discovery of a new scalar at the atlas @xcite and cms @xcite , the mass of the sm - like higgs @xmath25 is determined to be @xmath26 gev . without fine - tuning the parameters in the scalar potential , the @xmath27 mass - square matrix ( @xmath28 ) elements of the cp - even scalars
are expected to have the same order of magnitude . due to the sizable off - diagonal entries of @xmath28 ,
the mass splitting between the two cp - even higgs bosons can be up to a few hundred gev . for instance , if the diagonal and off - diagonal elements of @xmath28 are @xmath29 gev@xmath30 and @xmath31 gev@xmath30 , respectively , the masses of cp - even bosons are 126 gev and 583 gev . among the classifications of the model , only type ii thdm has the same yukawa couplings as the minimal supersymmetric standard model ( mssm ) .
it is of importance that the charged higgs mass of this model has been bounded to be @xmath32 gev at @xmath33 confidence level ( cl ) @xcite by the precision measurement of @xmath34 @xcite .
additionally , the constraint is insensitive to the parameter @xmath35 , where @xmath36 is the vacuum expectation value ( vev ) of higgs doublet which couples to up(down)-type quarks . as a result ,
the charged higgs in type ii thdm can not be a light particle , unless the model is further extended .
one of unsolved puzzles in particle physics is the origin of neutrino masses .
if we regard that the neutrino masses arise from the spontaneous symmetry breaking ( ssb ) , like that in the sm and thdm , the minimal extension of the thdm is to add an @xmath37 higgs triplet @xmath38 @xcite . besides the new scalar bosons ,
doubly charged higgses @xmath39 and neutral scalars @xmath40 , a pair of new charged scalar bosons @xmath41 exist in such model . due to
the strict constraint from the precision measurement of the electroweak @xmath42-parameter @xcite , the vev of @xmath38 ( @xmath43 ) is limited to be @xmath44 gev .
that is , before electroweak symmetry breaking ( ewsb ) , the triplet scalar bosons are degenerate massive particles .
it can be easily seen that if only one higgs doublet and one higgs triplet involve in a model , e.g. the so - called type ii seesaw model @xcite , the mass splittings among the higgs triplet components are suppressed by @xmath45 , where the @xmath46 is the typical mass of higgs triplet . by assuming that @xmath47 are 100% decaying into leptons , the experimental lower bound on the @xmath46 currently is around 400 gev @xcite .
some interesting characteristics in the two - higgs - doublet ( thd ) type ii seesaw model are found @xcite : ( i ) the mass splittings among the triplet components can be of the @xmath48 ; ( ii ) the doubly charged higgses @xmath47 have a complete new dominant decay pattern , e.g. , @xmath49 , where @xmath50 are the lightest singly charged higgs ; ( iii ) the @xmath50 can be as light as @xmath1 gev .
these characteristics ascribe to the new interactions in the scalar potential , and we summarize them as @xcite : @xmath51 where @xmath52 are the higgs doublets , @xmath53 is the second pauli matrix , and @xmath54 are the mass dimension - one parameters and can be of the order of electroweak scale .
the vev of @xmath38 then can be simplified as : @xmath55 where the detailed expression can be found in @xcite . as a result
, the small @xmath43 can be achieved by taking proper values of @xmath54 . according to the results of @xmath56 ,
the off - diagonal singly charged higgs mass - square matrix can be compatible to @xmath57 and @xmath58 ; and following the earlier discussions on the case of two cp - even bosons , one indeed can obtain the light charged higgs with a mass of 100 gev , even @xmath59 gev .
the only price that we need to pay is to introduce a mixing effect between @xmath60 and @xmath41 .
direct searches for a light charged higgs were performed at the lhc with @xmath617 tev @xcite and 8 tev @xcite .
although significant events above backgrounds have not yet found , the search of a light charged higgs still continues and is an interesting issue at the lhc @xcite . based on the characters of thd type ii seesaw model , in this work we study the possible signatures of the light charged higgs at the lhc . since the triplet charged higgses @xmath41 can couple to @xmath62 at the tree level , the decays @xmath15 are sizable when @xmath63 ; therefore , we separate the light charged higgs mass to two ranges ( i ) @xmath64 and ( ii ) @xmath10 , where the light charged higgs predominantly decays to @xmath11 , @xmath65 , @xmath9 , and @xmath66 in the former range , while it decays to @xmath11 , @xmath9 , and @xmath66 in the latter range . since we do not have any information about the mass of pseudoscalar boson ( @xmath67 ) , we omit the discussions for the decays @xmath68 by assuming @xmath69 in the study range of @xmath2
this assumption is supported by the recent cms measurement , where the mass @xmath70 in the range @xmath71 gev at small @xmath3 has been excluded @xcite .
the decays @xmath72 could be interesting processes to probe the charged higgs , however , the vertices are suppressed by @xmath73 , which has been strictly limited by the current higgs data @xcite . although it is not necessary , for simplicity we adopt @xmath74 in this analysis . in order to study the light charged higgs production and decays , in the following we briefly introduce the relevant mixing and couplings . as discussed earlier , the singly charged higgses @xmath60 and @xmath41 can mix together due to the interactions in eq .
( [ eq : vphidelta ] ) .
we parametrize the physical charged higgs states as : @xmath75 where @xmath76 are identified as the lighter ( heavier ) charged higgses .
accordingly , the yukawa couplings of @xmath50 to the sm quarks are given by : @xmath77 d h^+ + h.c.\ , , \label{eq : qyu } % % % & + \sqrt{2 } \bar \nu_l v_{\rm pmns } \left [ -\frac{\tan\beta}{v } { \bf m_\ell } \right ] \ell_r h^+ + h.c.\,,\end{aligned}\ ] ] where we have suppressed the flvaor indices , @xmath78 stands for the cabibbo - kobayashi - maskawa matrix , and @xmath79 .
one can easily obtain the lepton yukawa couplings if the quark masses and ckm matrix are replaced by the lepton masses and pontecorvo - maki - nakagawa - sakata ( pmns ) matrix , respectively .
it is clear that the radiative @xmath0 decays via the charged higgs can be suppressed by @xmath80 in the decay amplitude ; the constraint on the light charged higgs then is relaxed .
we write the feynman rules for the gauge interactions to be @xmath81 where we have used the basis of @xmath60 and @xmath41 instead of @xmath50 and @xmath82 , @xmath83 can be @xmath60 and @xmath41 , @xmath84 denote the momenta of charged particles , @xmath85 , @xmath86 , and @xmath87 .
it can be seen that the @xmath88 coupling is suppressed by @xmath89 , however , when @xmath90 , its magnitude is compatible to @xmath91 the coupling of @xmath92 .
thus , in the mass range ( i ) , the relative branching ratios ( brs ) for @xmath15 and @xmath93 are sensitive to the @xmath4 .
after introducing relevant couplings , we then study the partial decay rates of the charged higgs for those channels which are kinematically allowed .
it is found that in the mass range ( i ) the partial decay rate indeed is dominated by the three - body decay @xmath94 , and its expression is written as : @xmath95 \nonumber \\ & \times \left [ 2p_t\cdot p_{b2 } + \frac{1}{m^2_w } \left ( ( q^2 -m^2_b)^2 -m^4_w\right ) \right ] d_3(ps)\ , , % \end{aligned}\ ] ] where @xmath96 and @xmath97 are the inner products of the particle momenta , and @xmath98 denotes the phase space factor of the three - body decay ; in sum , they are given by : @xmath99 as for the dominant two - body decays , they are formulated as : @xmath100\ , , % % \left [ 3 + \frac{m^4_{h^\pm}}{4m^2_w m^2_z}\lambda(r_z , r_w ) \right]\ , , % \end{aligned}\ ] ] where we have dropped the light fermion masses , @xmath101 is used , @xmath102 , @xmath103 , @xmath104 , and @xmath105 .
besides the @xmath9 and @xmath66 modes , in the mass range ( ii ) one of vector bosons in the vector - boson pair channels becomes off - shell .
although the contributions of the off - shell channels are small , we still express their partial decay rates as : @xmath106\ , , \nonumber \\
% \gamma(h^\pm \to w^\pm z^ * ) & \approx \frac { g^6 v^2_{\delta } c^2_{\pm } \xi_{va } } { 9\cdot 2^{8 } \pi^3 c^4_{w } } \int^{(m_{h^\pm}-m_w)^2}_0 dq^2 \frac{q^2 \sqrt{\lambda(r_w , r_{q})}}{|q^2 -m^2_z |^2 } \left [ 1 + \frac{e^2(m^2_w , q^2)}{2 m^2_w}\right ] % \end{aligned}\ ] ] with @xmath107 and @xmath108 . according to the obtained formulas for the @xmath50 decays , we show the br for each @xmath50 decay mode as a function of @xmath2 in fig .
[ fig : brch](a ) and that as a function of @xmath109 in fig .
[ fig : brch](b ) , where @xmath110 and @xmath111 gev are used in plots ( a ) and ( b ) , respectively , and @xmath112 is fixed in both plots . without further mentioning ,
@xmath113 gev is indicated in this work . from plot ( a )
, it can be clearly seen that when @xmath114 gev , the @xmath115 can be one order of magnitude larger than the other decay modes .
the @xmath116 is about a factor of two larger than the @xmath117 in the considered mass ranges ( i ) and ( ii ) .
when @xmath2 is above the threshold of @xmath118 mode , the brs of @xmath9 and @xmath118 decay channels are compatible . from plot ( b ) , it can be seen that the @xmath119 is enhanced when @xmath109 is decreased .
on the other hand , since @xmath109 dictates the single charged higgs production cross section , in order to produce the @xmath50 bosons with sizable cross sections , the values of @xmath109 can not be too small . in order to understand the situation with large @xmath120 , we present the results with @xmath121 in fig .
[ fig : brch_tb30 ] . from the results , it can be clearly seen that @xmath9 mode overwhelmingly dominates at large @xmath120 .
and ( b ) @xmath4 , where we take @xmath122 for both plots , @xmath123 for plot ( a ) , @xmath111 gev for plot ( b ) , and @xmath124 stands for @xmath11.,title="fig:",width=302 ] and ( b ) @xmath4 , where we take @xmath122 for both plots , @xmath123 for plot ( a ) , @xmath111 gev for plot ( b ) , and @xmath124 stands for @xmath11.,title="fig:",width=302 ] , but for the case with @xmath12.,title="fig:",width=302 ] , but for the case with @xmath12.,title="fig:",width=302 ] to numerically calculate the single charged higgs production cross section , we use the calchep @xcite associated with the cteq6l parton distribution functions ( pdfs ) @xcite .
the numerical analysis is done at the center of mass energy of @xmath18 tev .
it is found that there are four interesting channels to produce the single charged higgs in the @xmath125 collisions .
they are : @xmath126 , @xmath127 , @xmath128 , and @xmath129 jet , where the cp - conjugated processes are indicated , and the jet includes the gluon , light quark , and @xmath130 jets .
the main free parameters in this study are @xmath120 , @xmath109 , and @xmath2 . to show the correlations between the cross sections and @xmath2 , we plot the production cross sections of these channels as a function of @xmath2 in fig .
[ fig : x_ch](a ) , where @xmath112 and @xmath131 are used . by the results , we see that the cross sections of @xmath132 and @xmath133 channels are much larger than those of @xmath134 and @xmath129 jet channels .
similarly , we show the production cross sections as a function of @xmath109 in fig .
[ fig : x_ch](b ) , where @xmath135 and @xmath136 gev are used .
it can be seen that when the values of @xmath109 are around 0.2 , the @xmath137 and @xmath138 can still be over 100 fb . according to the results in fig .
[ fig : brch](b ) , where the @xmath139 is larger than the @xmath116 in the mass region ( i ) and small @xmath109 , it is of interest to explore the charged higgs through @xmath65 channel in such region . and
( b ) @xmath4 , where @xmath18 tev is used , we have fixed @xmath122 for both plots , @xmath123 for plot ( a ) , and @xmath2 for plot ( b).,title="fig:",width=302 ] and ( b ) @xmath4 , where @xmath18 tev is used , we have fixed @xmath122 for both plots , @xmath123 for plot ( a ) , and @xmath2 for plot ( b).,title="fig:",width=302 ] according to the charged higgs yukawa couplings shown in eq .
( [ eq : qyu ] ) , when @xmath120 increases , the contributions from up(down)-type quarks are decreased ( increased ) . since @xmath140-quark is much heavier than @xmath141-quark , it is expected that the single @xmath50 production cross sections should be lowered as the values of @xmath120 are augmented in the beginning .
however , when the values of @xmath142 are increased up to @xmath143 , due to @xmath144 being near @xmath145 and @xmath146 being larger than @xmath147 , the cross sections are enhanced close to the case with @xmath112 .
to illustrate the situation at large @xmath120 , we show the @xmath148 production cross sections as a function of @xmath2 and @xmath109 with @xmath121 in fig .
[ fig : x_ch_tb30](a ) and ( b ) , respectively . by the plots
, it can be seen that only the channel @xmath129 jet has a significant enhancement and the others are slightly smaller than the situation with @xmath112 .
the significant enhancement in @xmath129 jet channel arises from the result that @xmath149 is larger than @xmath147 by a factor of above two . , but for @xmath12.,title="fig:",width=302 ] , but for @xmath12.,title="fig:",width=302 ] in order to assess the discovery potential of the light charged higgs , we employ the event generator madgraph / madevent5 @xcite to generate the simulation events , where the relevant feynman rules and parameters of the model are provided by feynrules 2.0 @xcite , and the nnpdf23lo1 pdfs @xcite are used .
the generated events are passed through pythia6 @xcite to include the effects of hadronization and initial / final state radiation , and to deal with the sm particle decays , e.g. , the @xmath150-boson and top - quark decays .
furthermore , the events also run through the pgs4 for detector simulation @xcite .
as discussed before , the singly charged higgs @xmath50 is dominantly produced by the processes @xmath151 , @xmath152 , and @xmath153 jet ; and its production cross sections and decays depend on the parameters @xmath2 , @xmath109 , and @xmath120 .
therefore , for combining the charged higgs mass ranges ( i ) and ( ii ) and the schemes of @xmath154 , we classify the used schemes to be @xmath155 and @xmath156 , where the number in the subscript denotes the mass range and @xmath157 stands for @xmath158 . from fig . [
fig : brch_tb30 ] , it can be seen that the decay @xmath93 overwhelms the other decay modes at large @xmath120 ; that is , under such circumstance @xmath9 is the only channel to look for the @xmath50 . recently , atlas @xcite and cms @xcite reported the upper limits on the product @xmath159 .
the experimental situation can be applied to our schemes @xmath160 .
thus , instead of event simulation , we directly apply the constraints from the atlas and cms measurements for the scheme @xmath161 . according to the yukawa couplings in eq .
( [ eq : qyu ] ) , the decay width for @xmath162 can be easily formulated as : @xmath163\ , , \label{eq : thb } \\ & c_l = \frac{m_t } { v t_\beta } \
, , \quad c_r = \frac{t_\beta m_b } { v}\ , , \quad % \lambda(x , y ) = 1 + x^2+y^2 -2x-2y -2xy , \quad \tilde \lambda(x , y ) = 1+x^2+y^2 -2x + 2y -2xy \ , , \nonumber \end{aligned}\ ] ] where @xmath164 is used , @xmath165 , and @xmath166 . to estimate the br , the total decay width of top - quark
is taken as @xmath167 gev @xcite .
consequently , we show the @xmath159 as a function of @xmath2 in fig . [
fig : tdecay ] , where the atlas and cms upper limits are shown on the plots , the curves represent the different values of @xmath109 , and for comparison we plot the cases with @xmath112 ( left panel ) and @xmath121 ( right panel ) . from the results
, it can be seen that a lots of parameter space of @xmath109 has been excluded by the current lhc data ; and the constraint on the case with @xmath168 is stronger due to @xmath169 . in the following event simulation ,
we concentrate on the schemes @xmath170 . as a function of @xmath2 , where the upper limits from the atlas @xcite and cms @xcite measurements are included ,
the curves represent the different values of @xmath4 , and the left ( right ) panel is for @xmath171 .
, title="fig:",width=302 ] as a function of @xmath2 , where the upper limits from the atlas @xcite and cms @xcite measurements are included , the curves represent the different values of @xmath4 , and the left ( right ) panel is for @xmath171 .
, title="fig:",width=302 ] to estimate the ratio of signal event number ( @xmath172 ) to background event number @xmath173 , we have to understand the sm backgrounds that can mimic the signals of the charged higgs .
we list the main background processes as follows : 1 .
zz background : @xmath174 jets , 2 .
ww background : @xmath175 jets , 3 .
wz background : @xmath176 jets , 4 .
top background : @xmath177 , @xmath178 , where the number of jets is taken to be @xmath179 . in order to suppress the background events , we consider the kinematic cuts ( kcs ) , which are applied to all signals and backgrounds , as : @xmath180 where @xmath181 is the transverse momentum , @xmath182 is the pseudo - rapidity with @xmath183 being the scattering angle in the laboratory frame , and @xmath184 denotes the highest @xmath181 jet .
the further cuts can be proposed depending on the properties of each process . in the following ,
we discuss the simulation analysis for each scheme .
* scheme @xmath185 * : according to the results as shown in fig .
[ fig : brch](b ) , it is known that @xmath186 are the two main decay channels at small @xmath109 . if we take @xmath187 gev , the interesting signal processes are @xmath188 where the cp - conjugated processes are indicated . to display the signal events for the @xmath189 and @xmath15 channels
, we set the extra event selection conditions to be `` 2 leptons + @xmath190 jets '' and `` 3 leptons + @xmath191 jets '' with @xmath192 , respectively . in order to increase the significance of the signal from the channel @xmath15
, we further require the invariant mass of 3-lepton to satisfy the selection condition : @xmath193 to clearly show the signal and each background , we present the remaining event numbers after imposing the proposed kcs and extra selection conditions in table [ tab : nevents]([tab : nevents2 ] ) for the signal @xmath15 ( @xmath19 ) and backgrounds , where the luminosity of 100 fb@xmath17 is used , and the last column is the significance , which is defined as @xmath194 . since the @xmath8 mode can be over the @xmath9 mode at small @xmath109 , we take @xmath195 for the channel @xmath15 .
however , in order to obtain larger @xmath50 production cross section , we take @xmath131 for the channel @xmath16 . from table [
tab : nevents ] , it can be seen that with the proposed cuts , the background events from the top - quark and @xmath65 backgrounds are still much larger than the signal events . in principle
, we can further reduce the top - quark background by imposing the kc on the second highest @xmath181 of charged lepton ; however , the @xmath65 background does not significantly change .
the reason is that the kinematic distributions from the signal @xmath15 and from the @xmath65 background are very similar .
the proposed kc will reduce both events .
similar situation also happens in table [ tab : nevents2 ] for the signal from the channel @xmath19 .
the difference is that we can find a kc ( e.g. , invariant mass of @xmath196 ) to reduce the @xmath65 and @xmath197 backgrounds ; however , due to the similarity in the kinematic distributions between the signal @xmath16 and top - quark background , it is difficult to find an efficient cut to diminish the top - quark background without reducing the signal . from the results of both tables
, it can be clearly seen that the significance from @xmath15 is much smaller than that from @xmath198 .
the results can be easily understood as follows : the product @xmath199 at @xmath200 is close to that at @xmath131 , however , @xmath201 at @xmath131 is a factor of over 30 larger than that for @xmath202 . by the definition of significance , the @xmath203 of @xmath204 naively should be one order of magnitude smaller than that of @xmath16 . in sum , by combining all analyses , we conclude that it is difficult to search for a light charged higgs via @xmath15 channel .
.event number for the signal @xmath15 and backgrounds with the proposed kinematic cuts in the scheme @xmath185 , where we have used the luminosity of 100 fb@xmath17 , @xmath205 gev , and @xmath206 .
[ tab : nevents ] [ cols="<,^,^,^,^,^,^,^",options="header " , ] in summary , we have assessed the discovery potential of the light charged higgs in the two - higgs - doublet and one higgs triplet model at the center of mass energy of @xmath18 tev . if @xmath93 is the dominant decay channel , it is found that the current atlas @xcite and cms @xcite upper limits on the @xmath207 can be directly used , and the correlation between @xmath2 and @xmath4 is severely constrained .
although @xmath15 is allowed and its branching ratio is not suppressed in the model , the significance of the signal @xmath15 is much smaller than that of the @xmath16 .
we conclude that the better production processes to search for a light charged higgs are @xmath208 , where the charged higgs decay channel is @xmath19 . |
concave and convex billiards as prototypes of conservative chaotic systems have received a growing theoretical and experimental interest in the fields of nonlinear dynamics and statistical mechanics [ 1 ] .
these billiards are nowadays fabricated as quantum dots or antidots in ballistic microstructures where the system size is much less than the mean free path @xmath0 m ) and larger than the fermi wavelength ( @xmath1nm)[2 ] .
since the striking experiments on the magneto - conductance of nanoscale stadium and circle billiards with a pair of conducting leads [ 2 ] , quantum transport phenomena have been intensively investigated [ 3 ] .
some salient features of experiments well agreed with the semiclassical theory , where a semiclassical conductance formula was derived for two - dimensional ( 2-d ) chaotic billiards by combining the semiclassical green function with the landauer formula [ 4,5 ] .
subsequently , the semiclassical conductance formula was extended for integrable 2-d billiards [ 6 ] , and detailed numerical computation on circle billiards indicated that the conductance also depends on the lead orientation [ 7 ] .
recently , the dependence of the conductance on the lead orientation was found experimentally in a ballistic square billiard [ 8 ] . at present , another important object of research on quantum transport is the three - dimensional ( 3-d ) billiard [ 9 ] .
the ballistic 3-d quantum dots will be fabricated , for instance , by exploiting drying etching processess with focused - ion beams applied to @xmath2 @xmath3 double hetero - structures , which will be less than both the elastic mean free path and phase coherence length .
the subject of this paper is to extend the semiclassical conductance formula for 2-d billiards to 3-d billiards and to investigate the quantum transport in open 3-d billiards from a viewpoint of the arnold diffusion , a key concept in high - dimensional nonlinear dynamical systems .
while arnold diffusion is a classical phenomenon , an anomalous phenomenon like this in classical mechanics should have a quantum counterpart , which can only be analytically revealed by using the semiclassical theory .
the conductance for open 3-d billiards connected to a pair of rectangular parallel - piped lead wires can be expressed by the transmission coefficient @xmath4 as @xmath5 , where @xmath6 is the * s*-matrix element between the incoming mode @xmath7 and the outgoing mode @xmath8 , and the double summation is taken over all propagating modes .
the transmission amplitude connecting those modes at the fermi energy @xmath9 reads as [ 10,11 ] @xmath10 where @xmath11 and @xmath12 are the longitudinal velocities of electrons for the incoming and outgoing modes , respectively . @xmath13 and @xmath14 are the @xmath15 coordinates for the transverse ( @xmath16 ) and longitudinal ( @xmath17 ) directions ( inward the billiard for @xmath18 and outward for @xmath19 ) of the incoming and outgoing leads , respectively , and @xmath20 is the green function for an electron propagating from the entrance to the exit . @xmath21 and @xmath22 are the transverse components of the wave functions at the leads . to simplify the problem , the cross section of the leads is assumed to be a square with side length @xmath23
. then @xmath24 and @xmath25 . in the ballistic regime ,
the green function from entrance ( @xmath26 ) to exit ( @xmath27 ) can be replaced by the gutzwiller semiclassical path - integral expression [ 12 ] @xmath28 and @xmath29 are the action and maslov index for a trajectory @xmath30 , and @xmath31 is the fermi velocity .
the above semiclassical green function was widely applied to derive the level density of both chaotic and regular billiards [ 12,13 ] . here , to avoid direct transmission without any bouncing at the billiard wall we may place a stopper inside the billiard . in the integration in eq.(1 ) we introduce the entrance angle @xmath32 between the initial direction of the orbit and the @xmath18 axis , and the angle @xmath33 between the projection of the initial direction onto the @xmath34 plane and the @xmath35 axis .
angles @xmath36 and @xmath37 are defined for the exit in the same way .
then , we can rewrite @xmath38 for each trajectory @xmath30 as @xmath39 where @xmath40 . for large mode numbers
, the integration over the transverse coordinates in eq.(1 ) can be performed in the stationary - phase approximation , and the transmission amplitude becomes @xmath41 with @xmath42\nonumber\\ & \times&[\sin2\varphi^ { ' } ( \frac{\partial\theta^{'}}{\partial{y}^{'}}\frac{\partial\theta^ { ' } } { \partial{x}^{'}}\cos^2 \theta^{'}- \frac{\partial\varphi^{'}}{\partial{y}^{'}}\frac{\partial\varphi^ { ' } } { \partial{x}^{'}}\sin^2\theta^ { ' } ) + \sin2\theta^ { ' } ( \frac{\partial\varphi^{'}}{\partial{y}^{'}}\frac{\partial\theta^ { ' } } { \partial{x}^{'}}\cos^2\varphi^{'}- \frac{\partial\theta^{'}}{\partial{y}^{'}}\frac{\partial\varphi^ { ' } } { \partial{x}^{'}}\sin^2\varphi^{'})]|,\end{aligned}\ ] ] where @xmath43,@xmath44 , @xmath45,@xmath46 , and @xmath47 , and @xmath48 is the maslov index which includes an extra phase coming from the possible sign change in each of the four fresnel integrals .
the above result means that only those isolated trajectories connecting a pair of transverse planes at discrete angles @xmath49 , and @xmath50 dominate the conductance . using eqs.(3,4 )
, we shall first analyze a completely chaotic 3-d billiard with a pair of conducting leads . in this case
, the electron injected from the incoming lead will bounce at the billiard boundary in an ergodic way before reaching the exit , leading to amplitudes of the same order for both transmission and reflection coefficients . in the large mode number case ,
the summations over modes are replaced by integrations and the transmission coefficient @xmath51 is rewritten as @xmath52 noting the dimension of @xmath53 to be @xmath54 , and using the diagonal approximation , the above integration can be evaluated as @xmath55 .
the conductance for 3-d chaotic billiards is thus given by @xmath56 without any dependence on the lead orientations .
however , the situation will be dramatically changed for the case of completely or partially broken - ergodic 3-d billiards . as examples
, we choose so(2)-symmetric billiards : a completely integrable (: broken - ergodic ) 3-d billiard is available by rotating , e.g. , the 2-d ellipse billiard ( on the @xmath57 plane ) around the @xmath58 axis ( see fig.1 ) . note : @xmath59 are the @xmath60 coordinates ; a partially broken - ergodic 3-d billiard is also available in a similar way by rotating , e.g. , the 2-d stadium .
all these so(2)-symmetric 3-d billiards have the angular momentum @xmath61 as a constant of motion . in an open system version of each of these billiards , one may place in its inside a smaller stopper of the similar shape , to avoid direct transmission of electrons .
the resultant 3-d @xmath62 billiards still retain the so(2 ) symmetry .
firstly , consider the vertical case ( i ) when the incoming and outgoing leads are connected with the billiard vertically at points @xmath63 and @xmath64 , respectively ( see fig.1 ) .
electrons to reach the exit @xmath64 should have the vanishing angular momentum , @xmath65 .
therefore , only the electrons with the initial velocity vector at the entrance @xmath63 lying in the @xmath57 plane can reach @xmath64 because of the conservation of @xmath61 ; other incoming electrons falling into the trajectories out of this plane can not reach @xmath64 and should return to @xmath63 . in deriving the transmission coefficient for this lead orientation , the integration prior to eq.(5 )
is reduced to the line integration over @xmath32 and @xmath36 only , yielding the conductance @xmath66 ( linear in @xmath67 ! ) .
( the reflection coefficient , which is obtained by applying the same procedure as in the chaotic billiard , is proportional to @xmath68 . )
we shall then consider the parallel case ( ii ) when the incoming and outgoing leads are parallelly placed at points @xmath69 and @xmath64 , respectively ( see fig.1 ) .
each electron incoming from the entrance @xmath69 has the vanishing angular momentum , @xmath65 , and thereby can reach @xmath64 or return to @xmath69 with probability of the same order .
therefore , for this lead orientation , @xmath70 . generally , in completely or partially broken - ergodic 3-d billiards , we have @xmath71 with @xmath72 , and the exponent @xmath73 is determined by how the billiard symmetry is broken by the lead orientation .
it should be noted : open 2-d billiards always share @xmath74 [ 4 ] , independently of both integrability and lead orientations .
the magneto - resistance is also a very important problem in 3-d billiards .
it is well known that in ballistic 2-d billiards the decrease of resistance with increasing magnetic field (: negative magneto - resistance ) occurs due to the weak localization based on the quantum interference between a pair of time - reversal symmetric orbits[2 - 5 ] .
what kind of additional novel phenomena will be expected in so(2)-symmetric 3-d billiards with the vertical lead ( @xmath75 ) orientation when a symmetry - breaking weak magnetic field @xmath76 will be applied along the @xmath77 axis ?
in contrast to phase - space structures in 2-d systems where each of chaotic zones generated by weak @xmath76 field is mutually separated by the kolmogorov - arnold - moser ( kam ) tori , those in 3-d systems consist of chaotic zones mutually connected via narrow chaotic channels called the arnold web ( aw ) [ 14 - 16 ] . in this case ,
electrons incoming in an arbitrary direction , which permeate ( chaotic ) zone by zone in an ergodic way through the aw , can show a global diffusion , i.e. , arnold diffusion ( ad ) .
therefore , a @xmath76 field increases the transmission channels and the number of stationary points in the integration in eq.(3 ) , and the resultant integration in eq.(5 ) over @xmath33 and @xmath37 as well as @xmath32 and @xmath36 indicates @xmath70 . in general we shall expect a more general result , @xmath78 with @xmath79 , and the exponent @xmath80 depends on the width of aw .
a crucial point is that a @xmath76 field has increased the exponent in the power - law behavior of the semiclassical conductance in case of the vertical lead orientation .
this anomalous phenomenon is just the negative magneto - resistance beyond the weak localization .
we here keep to employ the terminology of the @xmath81 @xmath82 wherever a @xmath76 field reduces the zero - field resistance . on the other hand , for the parallel lead ( @xmath83 ) orientation where each electronic trajectory lies almost in the vertical plane
, @xmath56 is retained , irrespective of the absence or presence of @xmath76 field ; we can expect only the @xmath76 field - induced increment of the proportionality constant , which implies a conventional negative magneto - resistance caused by suppression of the weak localization . before moving to the next section
, we should note : originally , the negative magneto - resistance was evidenced in the numerical analysis of the conductance of 2-d quantum dots [ 4 ] . the accompanying theory , based on the resistance , manipulated backscattering orbits and attributed this phenomenon to the weak localization [ 4 ] .
later works [ 5 ] based on the conductance , which introduced the small - angle - induced diffraction effect together with an idea of interference between a pair of @xmath84 time - reversal symmetric orbits [ 5 ] , swept away a cloud ( i.e. , breakdown of unitarity ) hanging over the above theory .
all our results in the present section have been obtained just within the same primitive analysis of the conductance as the original one for 2-d dots .
to numerically verify the above prediction , we shall choose a so(2)-symmetric 3-d elliptic shell billiard with the vertical lead orientation .
this ellipsoid is a typical example of completely integrable so(2)-symmetric 3-d billards .
the system in the presence of a symmetry - breaking field @xmath76 parallel to @xmath77 axis will be examined . to understand the classical dynamics of electrons , we analyze as a poincar surface of section the @xmath85 plane with longitude @xmath86 and velocity component @xmath87 as birkhoff coordinates [ 17 ] .
( @xmath88 is the velocity unit corresponding to the lowest mode . ) for @xmath89 , the trajectory from the entrance @xmath63 with an arbitrary initial velocity vector lying out of @xmath90 plane proves to be confined to a torus ( see fig.2(a ) ) , failing to reach the exit @xmath64 , i.e. , @xmath91 .
let @xmath92 ( @xmath93 ) be the ratio between the cyclotron radius @xmath94 and the characteristic length @xmath95 of the cross - sectional ellipse ( see the caption of fig.1 ) .
then @xmath76 field can be expressed as @xmath96 with @xmath97 ( @xmath98 corresponds to @xmath95 ) . in case of a weak field ( @xmath99 ) ,
poincar sections for the trajectory with the same initial velocity vector as in fig.2(a ) are given in figs.2(b ) and ( c ) for increasing numbers of bouncings at the outer billiard wall , and similarly those for the trajectory with the initial velocity vector lying in the @xmath100 plane are given in figs.2(d)-(f ) .
one can find : as the @xmath76 field is switched on , the orbit trajectory first leaves the initial torus , enters an outer layer , and , repeating a similar process , diffuses over more and more distant layers , exhibiting a phenomenon of the arnold diffusion ( ad ) .
even though a symmetry - breaking field is weak enough to keep the orbit almost straight , ad makes it possible for the trajectories with initial angular momenta @xmath101 to reach the exit @xmath64 , i.e. , @xmath91 ( see figs.2(c ) and ( f ) ) and eventually to contribute to conductance . to proceed to the numerical calculation of semiclassical transmission amplitudes in eqs.(3,4 ) for the 3-d elliptic billiard with the vertical lead orientation ,
let us define the @xmath102 monodromy matrix @xmath103 as @xmath104 .
at the entrance we introduce a pair of mutually orthogonal coordinates @xmath105 and @xmath106 lying in the transverse plane perpendicular to the initial direction of each orbit .
similarly @xmath107 and @xmath108 are chosen at the exit . using elements ( @xmath109 ) of @xmath103 matrix , we rewrite the necessary factors for the numerical computation .
for example , @xmath110/({{m } } _ { 11}{{m}}_{22}-{{m}}_{12}{{m}}_{21})$ ] and @xmath111 .
while @xmath112 and @xmath113 ( maslov index ) can be computed by the method suggested in refs.[9,18 ] , the first and last rotation angles in @xmath103 should be determined by using @xmath105 , @xmath106 and @xmath107 , @xmath108 chosen as above , since , after folding the orbit , the initial and final coordinates should become identical . combining all these factors with eqs.(3,4 ) , the semiclassical conductance is calculated explicitly .
the result , which includes both the classical contribution and the quantum correction in an inseparable way , is plotted in fig.3 . although the data show small fluctuations , the fitted curves for their means are described by @xmath56 and @xmath66 in cases of @xmath99 and @xmath89 , respectively , confirming the analytic issue suggested in the previous section .
the increase of the exponent in the @xmath67 dependence of @xmath114 upon switching on a weak field , which can not be seen in 2-d cases , provides a numerical evidence that the negative magneto - resistance (: a difference between two lines in fig.3 ) in 3-d systems comes from ad as well as the weak - localization correction .
we have investigated semiclassical quantum transport in three - dimensional ( 3-d ) ballistic quantum billiards , and explored novel phenomena that can not be seen in 2-d billiards . for partially or completely broken - ergodic 3-d billiards such as so(2 ) symmetric billiards , the dependence of the conductance @xmath114 on the fermi wavenumber @xmath67 is dramatically changed by the lead orientation . in case of the vertical orientation , a symmetry - breaking weak magnetic field turns out to increase the exponent in @xmath67 dependence of @xmath114 . in marked contrast to the 2-d case , the anomalous increase of the conductance should be caused by the arnold diffusion ( ad ) as well as a weak localization correction .
we should note the following:(i)in an extremely weak field case when the width of arnold web ( aw ) is thin enough , we shall see @xmath78 with @xmath115 .
our numerical calculation in the previous section suggests that , to have @xmath116 in a very large @xmath67 regime , @xmath117 should be no less than @xmath118 ; ( ii ) role of a stopper is also essential .
reflection at an inner convex wall should increase the instability of orbits , which is favourable for a genesis of ad ; ( iii ) to clearly observe ad , the symmetry - breaking field should destroy all constants of motion except for the energy . from this viewpoint ,
it is more advantageous to choose a so(2)-symmetric 3-d chaotic shell billiard generated from a fully - chaotic 2-d billiard with a hollow by its rotation around the @xmath58 axis .
the billiards of this kind , which are partially broken - ergodic , are experimentally more accessible than a ( completely broken - ergodic ) elliptic shell billiard , and all conclusions for the latter should hold for the former as well . finally , we point out the role of temperature in observing the effect of ad on the conductance . at finite temperatures ,
the conductance is related to the scattering matrix as @xmath119\sum_{\bf n , m}|t_{\bf n , m}|^2 $ ] with @xmath120 the fermi distribution function .
then the integration over energy will give a temperature factor @xmath121 with @xmath122 the length of trajectories .
for the vertical lead orientation , this factor suppresses very long trajectories reaching the outgoing lead along the arnold web , and thereby the negative magneto - resistance is described only by the weak - localization correction . on the other hand , if the temperature is lowered , the self - averaging over @xmath67 is not permissible , leading to a suppression of the weak - localization correction .
then the increment of the magneto - resistance is mainly caused by the effect corresponding to the arnold diffusion .
we hope that a future experiment on 3-d ballistic cavities can verify our prediction .
n. s. krylov , _ works on the foundation of statistical physics _
( princeton university press , princeton , 1979 ) ; ya .
g. sinai , russ .
25 * , 137 ( 1970 ) ; p. gaspard and g. nicolis , phys .
lett . * 65 * , 1693 ( 1990 ) . c. m. marcus , a. j. rimberg , r. m. westervelt , p. f. hopkins and a. c. gossard , phys .
lett . * 69 * , 506 ( 1992 ) ; a. m. chang , h. u. baranger , l. n. pfeiffer , and k. w. west , phys .
lett . * 73 * , 2111 ( 1994 ) . | we investigate a semiclassical conductance for ballistic open three - dimensional ( 3-d ) billiards . for partially or completely broken - ergodic 3-d billiards such as so(2 ) symmetric billiards , the dependence of the conductance on the fermi wavenumber is dramatically changed by the lead orientation .
application of a symmetry - breaking weak magnetic field brings about mixed phase - space structures of 3-d billiards which ensures a novel arnold diffusion that can not be seen in 2-d billiards .
in contrast to the 2-d case , the anomalous increment of the conductance should inevitably include a contribution arising from arnold diffusion as well as a weak localization correction .
discussions are devoted to the physical condition for observing this phenomenon . |
one of the most promising theories for physics beyond the standard model is supersymmetry . in the minimal supersymmetric exstension of the sm , the mssm ,
an additional quantum number , defined as @xmath6 is introduced .
sm particles carry r - parity of 1 while supersymmetric particles have r - parity of -1 .
r - parity violating ( rpv ) decays can happen through the following interaction terms : @xmath7 where the indices i , j and k denote the generations .
the fields in eq .
[ eq : w_rp_odd ] are superfields @xcite . the 48 rpv couplings in eq .
[ eq : w_rp_odd ] are constrained from theoretical and phenomenological points of view ( @xcite , @xcite ) .
the @xmath8 collaboration searched for heavy sneutrino decaying into @xmath9 in 1.0 @xmath3 of data @xcite .
the previous search for high - mass resonances decaying into @xmath9 channel by the cdf collaboration used 344 @xmath10@xcite . in susy models , with an extra @xmath11 , lepton flavor violation ( lfv ) interactions are allowed by the lagrangian density : @xmath12x_{\mu}$ ] , where @xmath13 are generation indices , @xmath14 is the @xmath11 gauge coupling , @xmath15 is electroweak mixing angle and @xmath16 is referred to as the `` charges '' @xcite .
the lepton charges @xmath17 are constrained by low energy experiments @xcite .
cdf run ii performed a search for @xmath18 and excluded a portion of the @xmath19 vs.@xmath20 plane @xcite .
we use cdf run ii data corresponding to an integrated luminosity of 1 @xmath21 .
we select events with two identified leptons of different flavor and opposite electric charge . to distinguish the flavor , we select only hadronic taus .
the kinematic cuts of the leptons are the following : electron @xmath22 , muon @xmath23 and tau visible @xmath24 .
candidate events in @xmath0 channel and @xmath1 channel are collected by the high@xmath25 single electron trigger path , which requires an electron candidate with @xmath26 and with pseudorapidity @xmath27 1.1 .
candidate events in @xmath28 channel are collected by the high@xmath25 single muon trigger path , which requires a muon candidate with @xmath29 and pseudorapidity @xmath30 1.0 .
lepton candidates above the kinematic threshold are required to pass lepton identification cuts . in this analysis , we select muons which pass through cmup ( @xmath30 0.6 ) and cmx ( @xmath27 1.0 ) muon detectors @xcite . they are required to pass the standard cdf muon i d cuts @xcite .
electron and hadronic tau reconstruction and identification algorithms have been improved to increase the acceptance at large lepton energy . a standalone calibration of the central showermax detector ( ces )
@xcite has been performed , providing a better resolution of the shower energy . as a consequence ,
the i d efficiency of electrons improved by @xmath31 ( figure [ fig : leptonacc ] ) .
it also improves the energy measurement of @xmath32 contributing to the visible momentum of the hadronic tau and @xmath33 ) are reconstructed by the central outer tracker ( cot ) ; neutral pion ( @xmath32 ) energies and positions are measured by the calorimeter and central showermax detector ( ces ) .
we measure @xmath32 energy using the calibrated ces energy measurement . ] .
as shown in figure [ fig : vismass100500 ] , the efficiency of the standard fixed upper cut on the tau mass ( @xmath34 ) decreases as the energy of the tau increases .
the fixed cut on tau visible mass ( mass of the @xmath35s and tracks originating in the tau decay ) is replaced by an energy dependent cut ( figure [ fig : slidingcut ] ) .
the upper cut is selected to yield a constant 95% efficiency when energy of the tau increases .
tau quantities related to the hadronic shower are poorly simulated and therefore tuned to data . due to the improvements described above , the i d efficiency of @xmath36 improved by factor of 2 ( figure [ fig : leptonacc ] ) .
there are several sm processes having the same signature as the signal : drell yan ( @xmath37 ) , diboson ( @xmath38 ) and @xmath39 .
there are also backgrounds due to mis - identified leptons .
we consider @xmath40jet(s ) and drell yan@xmath41jet(s ) , where the gauge boson leptonically decays into e/@xmath42/@xmath36 and a jet is mis - identified as a lepton .
in addition , leptons from drell yan ( @xmath43 , @xmath44 ) can be mis - identified ; typically the electron is mis - identified as tau and the muon is mis - identified as electron or tau .
these backgrounds are estimated using pythia monte carlo samples . the pythia event generator with cteq5l parton distribution function ( pdf ) and the cdf run ii detector simulation based on geant 3
are used to generate the simulated samples . along with the backgrounds listed above , qcd events with jets faking either leptons , or with one jet faking a lepton ( muon , tau ) and one @xmath45 faking an electron are background to the signal . to estimate this background , we use events with same charge in data since there is no correlation between the charge of the leptons in dijet and @xmath46jet events .
we need to subtract the contribution of same charge events originating from the backgrounds other than dijets and @xmath45 + jets .
the background estimates in whole mass region are shown in fig [ figemusig ] .
we check the consistency between the expectations from sm backgrounds and data in the control region defined as : @xmath47 ( for instance , fig [ figcontrol_1 ] ) . to estimate the number of expected sm events
, we use the same selection as applied to the data .
the events selected in the monte carlo background samples are scaled by the lepton i d scale factors and the trigger efficiencies .
we performed a search for high - mass resonances decaying into two leptons of different flavor : @xmath0 , @xmath1 and @xmath2 .
the search is based on 1 @xmath3 of tevatron run ii data .
we first improved electron i d efficiency and tau i d efficiency .
we find agreement between the sm prediction and observation in data in the control regions .
v. m. abazov _ et al . _
[ d0 collaboration ] , phys .
lett . * 100 * , 241803 ( 2008 ) [ arxiv:0711.3207 [ hep - ex ] ] .
a. abulencia _ et al .
_ [ cdf collaboration ] , phys . rev .
lett . * 96 * , 211802 ( 2006 ) [ arxiv : hep - ex/0603006 ] .
p. langacker and m. plumacher , phys .
d * 62 * , 013006 ( 2000 ) [ arxiv : hep - ph/0001204 ] . | we present a search for high - mass resonances decaying into two leptons of different flavor : @xmath0 , @xmath1 and @xmath2 . these resonances are predicted by several models for physics beyond the standard model , such as r - parity - violating mssm .
the search is based on 1 @xmath3 of tevatron run ii data collected with the cdf detector at @xmath4 = @xmath5 tev in proton anti - proton collisions . |
profiles of x - ray emitting hot gas in clusters of galaxies have been studied often by using the so - called @xmath0-model , which is an isothermal hydrostatic gas model consisting of a core and envelope . from the @xmath0-model analyses of 121 clusters including nearby clusters ( @xcite ) , ota and mitsuda ( 2002 ) obtained an interesting result that the distribution of the core radii of the intracluster gas exhibits two distinct peaks at @xmath3 kpc and @xmath4 kpc for @xmath5 ( see also @xcite ) .
akahori and masai ( 2005 ) ( hereafter am05 ) investigated correlations of the core radii with various properties of the clusters , and found that the radii in the larger core group around the peak @xmath4 kpc are marginally proportional to the virial radii of the clusters and therefore the origin may be attributed to simple self - similar collapse . on the other hand , in the smaller core group around @xmath3 kpc , no clear correlation is found between the core and virial radii , suggesting some other origin of the small core formation .
am05 investigated several possibilities of the origin of the small cores , not only by examining correlations among the observational quantities but by simulating the @xmath0-model with a hydrodynamics code .
many of the small core clusters possess central cd s or giant ellipticals , but the simulation shows that the gas core under their presence is @xmath6 kpc at most , and is too small to account for the observed range @xmath680 kpc .
am05 suggested another possibility : the effects of radiative cooling on the core size .
radiative cooling time is shorter in the core than in the ambient region , because the density is higher . as the core cools , the ambient gas then could inflow to compensate the pressure loss inside .
although the asca , chandra and xmm - newton observations ( e.g. , @xcite ; @xcite ; @xcite ) suggested much smaller amount of cooled mass than expected for the classical cooling flow model ( see @xcite for a review ) , this process increases the gas density toward the cluster center and is likely responsible for the small cores observed .
am05 found that the central cooling time , @xmath7 , is significantly shorter than the hubble time for 48 out of 50 small core clusters .
thermal evolution of clusters has been studied by many authors .
most of the papers addressed processes that would slow or inhibit the onset of radiative cooling instability , such as electron conduction and activities of active galactic nuclei . until an elapsed time compareble to @xmath7
, however , the gas appears to be cooling rather slowly with the temperature @xmath8 kev at the cluster center .
for example , ruszkowski and begelman ( 2002 ) studied the heating , conduction , and minimum temperature in cooling flows with a simple spherical model and showed that line and free - free cooling in the cluster center leads to slow cooling until the initial central cooling time has elapsed . at this stage feedback
is not yet very important .
masai and kitayama ( 2004 ) showed that when the intracluster gas is cooling with quasi - hydrostatic balancing , the temperature of the cooling gas appears to approach a constant value toward the cluster center . in the present paper
, we investigate thermal evolution of clusters , paying attention to the core radius of the intracluster gas undergoing radiative cooling .
our interest is the origin of the small cores or the observed core - size distribution which exhibits two distinct peaks .
we simulate cooling clusters using a hydrodynamics code ( am05 ) , and analyze the core radii applying the @xmath0-model , on which the observed results are based . in section 2
we describe the model . in section 3 , we present the results of calculations and discuss the properties of the cooling cores .
we give some concluding remarks in section 4 .
we calculate the evolution of radiatively cooling intracluster gas which is initially in hydrostatic balance under the presence of dark matter .
we analyze the resultant clusters using the @xmath0-model in the same manner as the observations ( ota , mitsuda 2002 ) : the core - radius distribution having two distinct peaks was obtained by analyses using the @xmath0-model .
since the @xmath0-model is based originally on the king model of the collisionless matter ( see e.g. , @xcite ) , we apply primarily the king model for the dark matter and galaxies of a cluster .
we also mention the results in the case of the nfw dark matter model ( navarro et al .
1996 ) , because such a cuspy profile has been suggested for dark halo mergers or cluster formation by numerical simulations . as for the thermal evolution of the intracluster gas ,
the two dark matter models show the similar results to each other . in the following we describe our models and calculations along the king dark matter model . with the gravitational potential @xmath9 of the total mass within radius @xmath10 ,
gravitational balance of a spherically symmetric cluster is written as , @xmath11 where dust " means collisionless components , i.e. dark matter and galaxies , @xmath12 is the boltzmann constant , @xmath13 the proton mass , @xmath14 the mean molecular weight for which we take @xmath15 , @xmath16 the mass density for each component , @xmath17 the gas temperature , and @xmath18 is the radial velocity dispersion which is equal to the line - of - sight velocity dispersion for the isotropic dust . if isothermal distribution is attained for every component , i.e. @xmath19 and @xmath20 , @xmath21 follows equation ( [ eq:1 ] ) .
for comparison with the observations of core sizes , in modelling clusters we apply the density profile function based on the @xmath0-model ( @xcite ) , @xmath22^{-3\beta/2},\ ] ] where @xmath23 is the central density , @xmath24 is the core radius , and @xmath0 is a parameter to represent the envelope slope . for the consistency with the @xmath0-model for the gas , we adopt the king model for the collisionless components .
although the original king model ( @xcite ) is not of simple analytic form , the profile may be approximated with equation ( [ eq:3 ] ) , which is sometimes called approximate king model ( see e.g. , @xcite ) .
table [ ta:1 ] shows the best - fit parameters for the king profile with equation ( [ eq:3 ] ) in the range @xmath25@xmath26 , where @xmath27 is the core radius of the king model .
cccccc range & & + @xmath10/@xmath27 & & @xmath28 & @xmath29 & @xmath30 + 0.01.0 & & 1.04 & 1.00 & 1.13 + 0.02.0 & & 1.05 & 1.00 & 1.14 + 0.05.0 & & 1.10 & 1.00 & 1.21 + 0.010.0 & & 1.34 & 0.98 & 1.46 + we investigate four models of clusters which consist of dark matter , galaxies , and gas with their mass ratio @xmath31 .
the galaxies and dark matter are represented by their gravitational potential components in the momentum equation ( equation ( 16 ) in am05 ) .
we employ the original king model for their density distributions , and then @xmath32 with @xmath33 for the initial gas profile ( equation ( [ eq:2 ] ) ) .
the initial profiles are shown in figure [ fig:1 ] , where the tidal radius @xmath34 , i.e. @xmath35 , and the core radius @xmath36 . the core radius of the gas well reflects the gravitational potential consistently in the context of the @xmath0-model ( am05 ) .
( 77mm,54mm)figure1.eps the four model - clusters have different initial core radii of the gas , @xmath37 160 , 200 , 250 , and 300 kpc , and satisfy the self - similar relation with their virial radii , as @xmath38 , or @xmath39 2.40 , 3.00 , 3.75 , and 4.50 mpc ( listed in table [ ta:2 ] ) .
these values are on the dashed line in figure 5 of am05 , which represents the @xmath40 relation obtained for the large core group peaked at @xmath4 kpc in the observed core - size distribution .
the initial gas temperature ( isothermal ) , @xmath17 , under the hydrostatic equilibrium is higher than the virial temperature , @xmath41 , as @xmath42 for @xmath43 ( see am05 ) .
cccccccccc & & & & & & & + & & @xmath44 & @xmath45 & @xmath17 & & @xmath46 & @xmath47 & @xmath0 & @xmath48 + model & & ( mpc ) & @xmath49 & ( kev ) & & ( kpc ) & @xmath50 & & ( gyr ) + ( a ) ........ & & 2.40 & 0.555 & 3.12 & & 171 & 2.15 & 0.83 & 0.689 + ( b ) ........ & & 3.00 & 1.08 & 4.87 & & 209 & 2.16 & 0.82 & 0.800 + ( c ) ........ & & 3.75 & 2.62 & 7.61 & & 260 & 2.10 & 0.81 & 1.086 + ( d ) ........ & & 4.50 & 3.66 & 11.0 & & 310 & 2.04 & 0.80 & 1.303 + + we carry out hydrodynamical simulations using a smoothed particle hydrodynamics ( sph ) code ( @xcite ; @xcite ) with 95,000 bodies . the smoothing length is estimated to be @xmath51 kpc and @xmath52 kpc for the core region ( @xmath53 ) and cooled cluster center ( @xmath54 ) , respectively , and sufficiently small compared with the core size concerned here . here , @xmath55 is the number density of the gas . for the general artificial viscosity
we adopt the monaghan - gingold value ( @xcite ; see also @xcite ) with the coefficients ( @xmath56 , @xmath0 ) = ( 1.0 , 2.0 ) , and for the gravitational softening we take @xmath57 with @xmath58 .
radiative cooling is taken into account for the energy equation as @xmath59 where @xmath60 is the internal energy per unit mass , @xmath61 is the velocity , and @xmath62 is the cooling function for the @xmath63-th sph - particle .
no additional heating processes are considered because the evolution of cooling cores is of our interest .
neither is electron conduction , which is not effective within cooling cores .
we use the cooling function of sutherland and dopita ( 1993 ) , which includes line emission with the metalicity @xmath64 , by approximation as @xmath65 @xmath66 in the temperature range @xmath81.510 kev . actually , until @xmath67 of practical interest , the temperatures of the cooling gas are still above @xmath68 kev where bremsstrahlung dominates . here
, @xmath69 is the initial cooling time at the cluster center . before going to calculations of evolution with radiative cooling
, we examine the hydrostatic balance of the gas with @xmath70 in equation ( [ eq:5 ] ) , as shown in figure [ fig:2 ] .
gravity on the gas and the pressure gradient balance with each other , and the initial gas profile is kept at least until 3.2 gyr , which is enough long compared with the dynamical ( free - fall ) timescale , @xmath71 gyr at the cluster center ( @xmath72 @xmath73 ) .
the gas is kept nearly isothermal so that the profile is well represented by the @xmath0-model , although the temperature declines somewhat at @xmath74 due likely to adiabatic expansion of the outermost envelope . at @xmath75 of practical interest ,
the hydrostatic gas is represented well by the @xmath0-model as follows . from equations ( [ eq:2 ] ) and ( [ eq:3 ] )
, the @xmath0-model should give a relation @xmath76 or @xmath77 .
actually , @xmath780.83 ( table [ ta:2 ] ) is consistent with @xmath79 for @xmath80 ( table [ ta:1 ] ) and @xmath81 .
we apply the double @xmath0-model as well as the single @xmath0-model to simulated clusters .
the double @xmath0-model is the superposition of two single @xmath0-models as @xmath82^{-3{\beta_1}/2}+\rho_2[1+(r / r_2)^2]^{-3{\beta_2}/2}.\ ] ] while in observations the parameters are inferred from the surface brightness profile , which depends substantially on the density profile as @xmath83 , in our calculations they are obtained straightforwardly from the density profile .
we obtain the best - fit parameters of the outer " components ( @xmath84 , @xmath85 , and @xmath86 ) using the data in the range @xmath87 , and then obtain the inner " ones ( @xmath88 and @xmath89 ) , assuming @xmath90 as done by ota and mitsuda ( 2002 ; 2004 ) , using the data of @xmath91 including the outer component with @xmath84 , @xmath85 and @xmath86 fixed .
figure [ fig:3 ] shows the time evolution of the parameters for the single and double @xmath0-models after the gas starts to cool .
we see in the bottom panel that the outer core radii ( the thin solid lines ) decrease very slowly compared with the core radii of the single @xmath0-model ( the dotted lines ) .
this implies that the outer component keeps the initial value since thermal evolution of the gas is predominant in the inner component .
actually , @xmath86 , which is determined by the outer component , is roughly kept the initial value , although @xmath0 of the single @xmath0-model decreases by about 10% within @xmath7 ( the top panel ) . the behavior is also seen in the central density . while the increase in the central density of the outer component is as small as a factor of @xmath92 , the central density with the single @xmath0-model increases exponentially ( the middle panel ) .
( 68mm,100mm)figure3.eps from the time evolution of the core radii , we calculate the time during which a cluster would have the core radius between @xmath10 and @xmath93 and estimate the population or relative number of clusters that would fall into a certain range of core radius .
the core size distribution thus obtained for the outer and inner cores is shown in figure [ fig:4 ] .
our calculations reproduce the observed core - size distribution or two distinct peaks except for the details such as their widths or the tails .
this implies that radiative cooling is a possible cause of small cores of clusters , while cooling is yet minor in large core clusters .
it is interesting that even the small core peak is produced by the clusters of @xmath94 , i.e. moderately cooled clusters that are keeping quasi - hydrostatic balance .
( 77mm,54mm)figure4.eps in the observed distribution of figure [ fig:4 ] one may notice four clusters which have very small cores .
it is unlikely that such small cores as @xmath95 kpc are reproduced by radiative cooling within @xmath96 .
our cooling function underestimates by @xmath6 % at @xmath97 kev compared with the function of sutherland and dopita ( 1993 ) .
if applied their original function , the small core peak might have a tail toward the small core end .
however , this is unlikely because such a very small core is quite transient ( figure [ fig:3 ] ) .
this may be related with the fact that clusters of @xmath98 are out of quasi - hydrostatic balance .
another possible explanation of the very small cores is the presence of central gravitational source ; the four clusters evidently have central cd s or giant ellipticals .
the presence of such galaxies is likely responsible for the very small cores ( see am05 ) .
the lack of clusters of core radii @xmath99 kpc is seen in the calculated distribution as well .
although the outer cores of @xmath8 100300 kpc of simulated clusters can explain the population around the peak of the observed large core clusters , the latter exhibits larger core tail extended to @xmath100 1 mpc .
if we simulate clusters which have initial huge cores @xmath101 mpc , we might get the cooled cores of @xmath102 kpc following the self - similarity ( see below )
. however , if such huge core clusters are in hydrostatic equilibria as represented by the @xmath0-model , the clusters would have unlikely virial radii as large as @xmath103 mpc .
in fact , most of the huge core clusters deviate from the self - similar relation , as shown in figure 5 of am05 .
a considerable factor for the origin of such huge cores is mergers .
when merger takes place , the gas could form a large flat core with a steep envelope ( large @xmath0 ) .
actually , in am05 sample , 6 out of 7 clusters of @xmath104 have cores of @xmath105 mpc , and the average of @xmath0 in the 17 clusters ( including 14 irregulars ) of @xmath106 mpc is 0.98 , which is significantly greater than the average 0.65 in 121 clusters . we find in figure [ fig:3 ] that the single @xmath0-model represents the intermediate profile between the inner and outer profiles of the double @xmath0-model . if cooling is minor and the self - similarity remains , the profile would be represented by the single @xmath0-model of a large core . with cooling the core , but yet minor in the outer , the profile is being that represented better but transiently by the double @xmath0-model , and eventually comes to that represented by the single @xmath0-model of a small core . as a result , the core - size distribution exhibits two distinct peaks along the thermal evolution .
two distinct peaks are not seen for the cluster sample of redshift between 0.4 and 1.3 ( @xcite ) .
the reason may be that their analysis was done with the single @xmath0-model ; otherwise , the cooling might be yet minor in such very distant clusters .
we confirm that two distinct peaks do not appear if applied the single @xmath0-model to the simulated clusters .
the estimation of the inner core radius depends on @xmath86 obtained from the outer slope . if @xmath86 is taken to be 10 % smaller than that in the present calculations , the resultant clusters might have 10 % smaller cooling cores , as expected from the @xmath0@xmath24 correlation described in figure [ fig:1 ] .
thus , the small - core distribution merely shifts toward the smaller side .
we also investigate clusters of the nfw dust profile ( navarro et al .
1996 ) , where the parameters of the nfw profile are determined so that the clusters have nearly the same initial density and temperature profiles of the gas , and analyze the core sizes in the same manner as above .
the obtained core - radius distribution is similar except for a few % tail extended to @xmath107 kpc ( figure [ fig:4 ] ) .
there is no appreciable difference between the nfw and king dark matter models in the cooling gas distribution .
( 46mm,30mm)figure5a.eps ( 38mm,27mm)figure5b.eps ( 38mm,27mm)figure5c.eps ( 38mm,27mm)figure5d.eps the evolutions of the density and temperature profiles ( figure [ fig:5 ] ) imply the self - similar evolution in the core region , since the evolution is predominant in the inner component .
we try to approximate the evolution of inner - core radii with a parameter , @xmath108 , and the normalized time , @xmath109 , as @xmath110 figure [ fig:6 ] shows that equation ( [ eq:7 ] ) can represent the inner - core evolution , i.e. thermal evolution of the small core can be described by the self - similar relation , normalized by @xmath7 and @xmath111 which are obtained from the initial outer component .
this self - similarity may be seen also in the clusters having different initial central densities , because in the @xmath0-model the central density determines the amplitude of the profile while the flat - core region clearly exists ; the core radius is related mainly to the envelope slope ( see figure [ fig:1 ] ) .
( 77mm,54mm)figure6.eps such a self - similarity is , however , likely lost in the observed small cores . at the initial state , the inner - core size can be represented by the outer - core size , as @xmath112 where @xmath113 is obtained from figure [ fig:3 ] . while the thermal evolution is predominant in the inner component , outer core is roughly constant , i.e. @xmath114 , so that @xmath115 is written with equations ( [ eq:7 ] ) and ( [ eq:8 ] ) as @xmath116 from equations ( 10 ) and ( 11 ) in am05 , the initial central cooling time is given by @xmath117 therefore , equation ( [ eq:9 ] ) is written as @xmath118 for the clusters with @xmath119 ( am05 ) .
equation ( [ eq:11 ] ) suggests that the self - similar relation in the small cores may be lost for not only the different age of the cluster from the last major merger , i.e. the time when the cluster formed to the current size and down to the hydrostatic equilibrium , which is labeled by @xmath120 , but the various values of @xmath0 from @xmath121 to @xmath122 in observed clusters as mentioned in am05 .
calculations show that the cooling gas approximately keeps the hydrostatic balance between the gravitational force and pressure gradient at each step until @xmath123 ( figure [ fig:7 ] ) .
we compare our result with quasi - hydrostatic cooling discussed by masai and kitayama ( 2004 ) .
they suggest that @xmath124 at every @xmath10 or @xmath125 where @xmath126 is the continuity equation , @xmath127 and @xmath128 are the mass of a uniform gas sphere and the cooling time for the local values at radius @xmath10 , respectively .
as shown in figure [ fig:8 ] , our calculation shows @xmath129 is nearly constant . until @xmath130 , constant @xmath131 is seen commonly for all the model clusters , not only the king case but also the nfw case .
( 77mm,54mm)figure7.eps masai and kitayama ( 2004 ) show some properties of quasi - hydrostatic cooling by approximating the density / temperature profiles in the power - law form .
we examine our results in the same manner with @xmath132 , @xmath133 and inflow velocity @xmath134 .
since @xmath135 , quasi - hydrostatic cooling means @xmath136 or @xmath137 such a relation is actually found at @xmath6 kpc : @xmath138 is comparable to @xmath139 with ( @xmath140 , @xmath141)=(0.28 , 0.45 ) . at @xmath142 kpc , however , @xmath143 is smaller than @xmath144 with ( @xmath140 , @xmath141)=(0.25 , 0.34 )
. therefore , quasi - hydrostatic balance may be satisfied marginally .
when the cluster center cools rapidly at @xmath145 ( figure [ fig:5 ] ) , the balance breaks because a large amount of inflow is required to maintain the balance .
this also leads to the fact that the spatial resolution in calculations becomes practically worse than estimated simply from the number of sph particles .
( 77mm,54mm)figure8.eps at @xmath146 kpc , the gas inflow velocity @xmath147 , where @xmath148 is the sound speed . the mass inflow rate @xmath149 is decreasing to @xmath150 , which are much smaller than expected for the classical cooling flow model and consistent with the quasi - hydrostatic cooling model by masai and kitayama ( 2004 )
. it should be noted that they consider @xmath151 for smooth inflow , but we find in figure [ fig:8 ] that the inflow velocity gradually decreases toward the cluster center , i.e. @xmath152 rather than @xmath153 .
it is , however , consistent with their prediction that @xmath154 is well satisfied in @xmath155 kpc .
@xmath152 implies that the gas is heated by viscosity .
we confirmed the convergence of the solution by using sph - particles greater than @xmath156 , which is also sufficient resolution to remove improperly accelerating cooling ( @xcite ) .
the gas inflow might be exaggerated by underestimating the viscous heating at the dead center of the cluster , though the influence reaches up likely to @xmath157 from the center because sph - particles within @xmath158 at a given point contribute to the smoothed estimate in the present work .
therefore , at each step until @xmath130 , quasi - hydrostatic balance is likely attained without some heat sources such as electron conduction and/or agn activities .
we investigate the thermal evolution of cluster cores : the core size of the intracluster gas can vary as the gas cools radiatively and flows toward the cluster center . in order to compare with the observed core radii , we apply the @xmath0-model or double @xmath0-model , as done by the observational studies , to analyze the gas profile , although the model is not always good for cooling cores of more or less center - peaked profiles .
thermal evolution of the gas may be classified into the following three stages : ( i ) at @xmath159 , the profile is represented by the single @xmath0-model with a large core or by the outer - core dominated double @xmath0-model , ( ii ) at @xmath160 , the profile is represented well by the double @xmath0-model , and ( iii ) at @xmath161 , the profile is represented by the single-@xmath0 model with a small core or by the inner - core dominated double @xmath0-model . until @xmath130 , which is the central cooling time for the initial gas profile
, the gas cools keeping the hydrostatic balance between the gravity on the gas and the gradient of the thermal pressure .
this evolution is explained by the quasi - hydrostatic cooling model proposed by masai and kitayama ( 2004 ) .
the properties of the cooling gas in our calculations , such as the mass inflow rate , radius dependence of the local cooling time and constant @xmath131 , are consistent with this model in the regime where quasi - hydrostatic structure is attained .
we find that the inflow velocity of the gas decreases toward the cluster center , and its mass flow rate is about @xmath162 at 20 kpc .
quasi - hydrostatic condition is marginally satisfied up to @xmath163 , and then the cluster center cools rapidly . regarding the relation with the core - size distribution , it may be an important clue for understanding thermal properties of clusters that even the small core peak is produced by clusters of @xmath94 , in other words , by clusters that are still keeping quasi - hydrostatic balance . analyzing the simulated clusters with the single and double @xmath0-models , we demonstrate the core size distribution exhibits two distinct peaks at @xmath2300 kpc and at @xmath164100 kpc with a valley at @xmath99 kpc until @xmath7 and is in fairly good agreement with the observed distribution ( ota and mitsuda 2002 ; akahori and masai 2005 ) .
this implies that the origin of small cores can be explained by radiative cooling or thermal evolution of the gas .
however , a question remains : the self - similarity is kept in simulated clusters through the thermal evolution with cooling , but seems to be lost in the observed small core clusters for the various values of @xmath0 from @xmath121 to @xmath122 ( akahori and masai 2005 ) .
99 akahori , t. , & masai , k. 2005 , , 57 , 419 cavaliere , a. , & fusco - femiano , r. 1976 , , 49 , 137 ettori , s. , tozzi , p. , borgani , s. , & rosati , p. 2004
, , 417 , 13 fabian , a. c. 1994 , ara&a , 32 , 277 hernquist , l. , & katz , n. 1989 , , 70 , 419 king , i. r. 1966 , , 71 , 64 lewis , a. d. , stocke , j. t. , & buote , d. a. 2002 , , 573 , l13 makishima , k. , et al .
2001 , , 53 , 401 masai , k. , & kitayama , t. 2004 , , 421 , 815 mohr , j. j. , mathiesen , b. , & evrard , a. e. 1999 , , 517 , 627 monaghan , j. j. 1992 , ara&a , 30 , 543 monaghan , j. j. , & gingold , r. a. 1983 , j. comput .
, 52 , 374 navarro , j. f. , frenk , c. s. , & white , s. d. m.,1996 , 462 , 563 ota , n. , & mitsuda , k. 2002 , , 567 , l23 ota , n. , & mitsuda , k. 2004 , a&a , 428 , 757 peterson , j. r. , et al .
2001 , , 365 , l104 ruszkowski , m. , & begelman , m. c. 2002 , , 581 , 223 sarazin , c. l. 1986 , rev .
58 , 1 springel , v. , & hernquist , l. 2002 , , 333 , 649 sutherland , r. s. , & dopita , m. a. 1993 , , 88 , 253 | we investigate the core structure of radiatively cooling intracluster gas , using a hydrodynamics code .
we calculate evolution of model clusters of the initial core radii 160300 kpc until the initial central cooling time , and analyze the resultant clusters using the double @xmath0-model as done by observational studies .
it is found that the core - size distribution thus obtained shows two peaks @xmath1100 kpc and @xmath2300 kpc and marginally can reproduce the observed distribution which exhibits two distinct peaks around @xmath3 kpc and @xmath4 kpc .
this result may suggest radiative - cooling origin for small cores , while cooling is yet insignificant in the clusters of large cores .
it should be noted that the small core peak is reproduced by clusters that are still keeping quasi - hydrostatic balance before the initial central cooling time has elapsed . |
as@xmath0 with the tetragonal thcr@xmath0si@xmath0-type structure .
the structure consists of alternating feas and ba layers stacked along the crystallographic @xmath14 axis .
( b ) crystal structure of eupd@xmath0sb@xmath0 with the origin of the unit cell shifted by ( 1/4 1/4 1/4 ) compared to that in the space group @xmath15 , for comparison purposes .
the structure consists of alternating pdsb and eu layers stacked along the crystallographic @xmath14 axis similar to the bafe@xmath0as@xmath0 shown in ( a ) . however , half of the pdsb layers are inverted ( the pd and sb atoms switch positions ) with respect to the feas - type layers.,width=316 ] the recent discovery of high - temperature superconductivity in @xmath16feaso@xmath17f@xmath18 ( @xmath16 = la , ce , pr , nd , sm , gd , tb , and dy)@xcite compounds with superconducting transition temperatures @xmath19 as high as 55 k has sparked a lot of interest in the search for new superconductors .
these materials crystallize in the tetragonal zrcusias - type structure with space group @xmath15.@xcite the structure consists of alternating feas and @xmath16o layers stacked along the crystallographic @xmath14 axis .
the parent compounds @xmath16feaso exhibit spin density wave ( sdw ) transitions at temperatures @xmath20 k.@xcite upon doping with f , the sdw is suppressed and superconductivity appears.@xcite another group of structurally related parent compounds with the chemical formula @xmath21fe@xmath0as@xmath0 ( @xmath21 = ca , sr , ba , and eu ) was soon discovered to show superconductivity upon doping or application of pressure . these compounds crystallize in the tetragonal thcr@xmath0si@xmath0-type structure with space group @xmath22 ( no . 139 ) .
the structure consists of alternating feas and @xmath21 layers stacked along the @xmath14 axis as shown in fig .
[ struc](a ) . in the feas layers ,
the fe atoms form a square planar lattice
. the @xmath21fe@xmath0as@xmath0 compounds also show sdw and structural transitions at high temperatures@xcite which are suppressed by doping with k , na , and cs at the @xmath21 site and accompanied by the onset of superconductivity.@xcite in both classes of @xmath16feaso@xmath17f@xmath18 and @xmath21fe@xmath0as@xmath0 compounds described above , feas layers that are stacked along the @xmath14 axis are evidently a key building block yielding superconductors with relatively high @xmath19 .
this gives a strong motivation to investigate similarly structured compounds in a search for additional high-@xmath19 superconductors .
the compound eupd@xmath0sb@xmath0 crystallizes in the cabe@xmath0ge@xmath0-type structure with space group @xmath15 ( no . 129),@xcite as shown in fig .
[ struc](b ) .
the structure is closely related to the @xmath21fe@xmath0as@xmath0 structure .
alternating pdsb and eu layers are stacked along the @xmath14 axis , similar to the @xmath21fe@xmath0as@xmath0 structure . however , there is a distinct difference between the two structures . in half of the pdsb layers in the eupd@xmath0sb@xmath0 structure ,
the pd atoms are arranged in a planar square lattice with two sb layers on either side of each pd layer , resulting in a tetrahedral coordination of pd by sb as in the feas - type layers . however , alternating with these layers are layers in which the pd and sb positions are switched , as shown in fig .
[ struc](b ) .
there have been reports of structural instabilities and antiferromagnetic ordering in some compounds forming in the cabe@xmath0ge@xmath0-type structure .
ucu@xmath23sn@xmath0 orders antiferromagnetically at 110 @xmath24c which is very high among uranium intermetallics.@xcite cepd@xmath0ga@xmath0 undergoes a tetragonal to monoclinic second order structural transition at 125 k and orders antiferromagnetically at 2.3 k.@xcite lapd@xmath0ga@xmath0 is superconducting below 1.9 k. @xcite eu was reported to be in a mixed valent state between eu@xmath25 ( spin @xmath26 ) and eu@xmath27 ( spin @xmath28 ) in polycrystalline samples of eupd@xmath0sb@xmath0.@xcite in this paper , we report the synthesis and structure of single crystals of eupd@xmath0sb@xmath0 and their physical properties including magnetic susceptibility , magnetization , specific heat , and electronic transport measurements .
sb@xmath0 . the grid size in 1 mm.,width=240 ] single crystals of eupd@xmath0sb@xmath0
were grown using pdsb self - flux which melts at @xmath29 805 @xmath24c .
the eu ( 99.999% pure ) was obtained from the ames laboratory materials preparation center .
the pd ( 99.95% pure ) and sb ( 99.999% pure ) were obtained from alfa - aesar .
pd and sb powders were thoroughly mixed inside a helium - filled glove box , and then poured on top of a chunk of eu ( @xmath29 0.1 g ) that was placed at the bottom of a 2 ml alumina crucible .
the elements were in the atomic ratio eu : pd : sb = 1:5:5 .
the top of the crucible was packed with quartz wool .
the crucible was then sealed in a quartz tube under vacuum and was placed vertically in a box furnace and heated to 1000 @xmath24c at a rate of 76 @xmath24c / h and held there for 6 h. the tube was then cooled to 850 @xmath24c at the rate of 1.5 @xmath24c / h and at this temperature the tube was removed from the oven and centrifuged to partially separate the flux from the crystals .
a single conglomerated chunk ( about 0.4 g ) was found in the crucible after removing the quartz wool .
plate - like gold - colored crystals were isolated mechanically .
the largest crystals had dimensions @xmath29 2@[email protected] mm@xmath31 .
the crystals are brittle and are easily broken into smaller pieces .
figure [ pic ] shows an as - grown crystal on a mm grid .
single crystal x - ray diffraction measurements were done using a bruker ccd-1000 diffractometer with mo k@xmath32
( @xmath33 = 0.71073 ) radiation .
magnetic measurements on the crystals were carried out using a quantum design superconducting quantum interference device ( squid ) magnetometer in the temperature @xmath6 range 1.8350 k and magnetic field @xmath7 range 05.5 t. heat capacity , electrical resistivity , and hall coefficient measurements were done using a quantum design physical property measurement system ( ppms ) . for the heat capacity measurements , apiezon
n grease was used for thermal coupling between a sample and the sample platform .
the heat capacity was measured in the temperature range 1.8300 k in @xmath7 = 0 , 2 , 5 , 7 , and 9 t. for electrical resistivity and hall coefficient measurements , platinum leads were attached to the crystals using silver epoxy .
electrical resistivity measurements were carried out using the standard ac four probe method with 10 ma excitation current in the temperature range 1.8300 k and magnetic field range 08 t. hall coefficient measurements were carried out using the five - wire configuration supported by the ppms act@xcite option with 100 ma excitation current in the temperature range 1.8310 k and magnetic field range 08 t. the hall voltage was computed at each temperature from the odd part of the measured transverse voltage upon reversing the sign of the applied magnetic field .
the even part was much smaller than the odd part at each measured temperature .
.crystal data and structure refinement of @xmath34 at a temperature of 173 k. here r1 = @xmath35@xmath36@xmath36@xmath37@xmath38@xmath36 @xmath39 @xmath36@xmath37@xmath40@xmath36@xmath36/@xmath35@xmath36@xmath37@xmath38@xmath36 and wr2 = ( @xmath35 [ @xmath41(@xmath36@xmath37@xmath38@xmath36@xmath42 @xmath39 @xmath36@xmath37@xmath40@xmath36@xmath42)@xmath42]/@xmath35 [ @xmath41(@xmath36@xmath37@xmath38@xmath36@xmath42)@xmath42])@xmath44 , where @xmath37@xmath38 is the observed structure factor and @xmath37@xmath40 is the calculated structure factor . [ cols= " < , < " , ] a comparison of structural and magnetic parameters of eupd@xmath0sb@xmath0 with those of some eu compounds which form in the related thcr@xmath0si@xmath0-type structure is given in table [ comparison ] .
of the compounds listed in table [ comparison ] , only eufe@xmath0as@xmath0 shows superconducting behavior under pressure@xcite as well as under doping at the eu site.@xcite @xmath45 for eupd@xmath0sb@xmath0 with @xmath46 and a @xmath47 ratio of 2.28 is 6.1 k compared to 20 k for eufe@xmath0as@xmath0 which has a smaller @xmath48 and a much larger @xmath47 = 3.10 .
thus the significant difference in @xmath45 values is probably due at least in part to these structural differences . in eu@xmath49k@xmath49fe@xmath0as@xmath0
which is superconducting below 32 k , however , af ordering of the eu spins still takes place below 10 k. the calculated effective moment of the eu spins in eupd@xmath0sb@xmath0 is close to that of eu spins in eufe@xmath0as@xmath0 .
the hall coefficient of eupd@xmath0sb@xmath0 remains negative between 1.8 300 k like that in the superconducting ba(fe@xmath17co@xmath18)@xmath0as@xmath0 and ba(fe@xmath17cu@xmath18)@xmath0as@xmath0.@xcite however , in eufe@xmath0as@xmath0 , the hall coefficient changes sign from negative to positive at @xmath50 k. at 300 k , @xmath51 in eufe@xmath0as@xmath0,@xcite which suggests that the charge carriers comprise both electrons and holes .
probably , the magnetic nature of the feas layers in eufe@xmath0as@xmath0 is an important factor behind the superconducting behavior at high pressures and low temperatures
. it will be very interesting to grow single crystals of eupd@xmath52fe@xmath18sb@xmath0 and study their physical properties .
the fe - doping at the pd site will eventually make the pd(fe)-sb layers structurally similar to the feas layers in eufe@xmath0as@xmath0 .
klauss , h. luetkens , r. klingeler , c. hess , f. j. litterst , m. kraken , m. m. korshunov , i. eremin , s .- l .
drechsler , r. khasanov , a. amato , j. hamann - borrero , n. leps , a. kondrat , g. behr , j. werner , and b. bchner , phys .
* 101 * , 077005 ( 2008 ) . | single crystals of eupd@xmath0sb@xmath0 have been grown from pdsb self - flux .
the properties of the single crystals have been investigated by x - ray diffraction , magnetic susceptibility @xmath1 , magnetization @xmath2 , electrical resistivity @xmath3 , hall coefficient @xmath4 , and heat capacity @xmath5 measurements versus temperature @xmath6 and magnetic field @xmath7 .
single crystal x - ray diffraction studies confirmed that eupd@xmath0sb@xmath0 crystallizes in the cabe@xmath0ge@xmath0-type structure .
the @xmath8 measurements suggest antiferromagnetic ordering at 6.0 k with the easy axis or plane in the crystallographic @xmath9 plane .
an additional transition occurs at 4.5 k that may be a spin reorientation transition .
the @xmath10 data also show the two transitions at 6.1 k and 4.4 k , respectively , indicating the bulk nature of the transitions .
the 4.4 k transition is suppressed below 1.8 k while the 6.1 k transition moves down to 3.3 k in @xmath11 t. the @xmath12 data show metallic behavior down to 1.8 k along with an anomaly at 5.5 k in zero field .
the anomaly is suppressed to 2.7 k in an 8 t field .
the @xmath4 measurements indicated that the dominant charge carriers are electrons .
the @xmath13 isotherms show three field - induced transitions at 2.75 t , 3.90 t , and 4.2 t magnetic fields parallel to the @xmath9 plane at 1.8 k. no transitions are observed in @xmath13 for fields parallel to the @xmath14 axis . |
throughout this paper , we assume that alice and bob , located some distance apart , each has initial control of one qudit from a two - qudit system in a pure state @xmath1 , with schmidt representation @xmath15 here , @xmath16 , assuring normalization , and we assume without loss of generality that @xmath17 let @xmath18 be the hilbert space with orthonormal basis @xmath19 , let @xmath20 be the hilbert space with orthonormal basis @xmath21 , and let @xmath22 .
@xmath23 is the state space for the system that alice and bob share , with orthonormal basis @xmath24 , @xmath25 , where the latter form for basis elements ( letting order distinguish alice and bob s kets ) will be used for notational convenience . in the deterministic dense - coding protocol , originated by bennett and wiesner @xcite , alice prepares messages for bob by applying to her qudit a physical operation actualizing a unitary operator ( on @xmath18 ) chosen from a family @xmath26 of _ encoding unitaries _ , and then she sends her qudit to bob via a noiseless @xmath27-dimensional quantum channel .
when bob receives alice s qudit , he measures the pair in an appropriate basis , obtaining a message . in order that the messages alice sends be perfectly distinguishable by bob , the states @xmath28 must be ( pairwise ) orthogonal in the hilbert space @xmath23 .
thus alice s encoding unitaries are necessarily orthogonal in the sense that they generate orthogonal states in @xmath23 , each corresponding to a message alice may send to bob .
if @xmath1 ( given by ( [ gqds ] ) ) is maximally entangled , i.e. , @xmath29 for @xmath30 , then alice may send @xmath31 perfectly distinguishable messages to bob ( @xcite ; see , also @xcite ) . for non - maximally entangled states @xmath1 , numerical work of mozes _ et al . _
@xcite suggests that for each @xmath32 in @xmath33 there is a collection of states that will support the sending of @xmath32 messages but not @xmath34 . as in @xcite , we quantify entanglement level of a bipartite system in state @xmath1 of ( [ gqds ] ) using @xmath35 which is the von neumann entropy of either of the reduced density operators @xmath36 or @xmath37 .
we call @xmath38 the _ entanglement entropy _ of @xmath1 . much of the work herein derives from numerical studies by mozes _
@xcite , designed to indicate the maximum number of perfectly distinguishable messages that alice can send to bob as a function of the schmidt coefficients of the initial state they share . before these studies were conducted , little was known about the message - bearing capacity of non - maximally entangled systems used for deterministic dense coding .
based on their results , both numerical and analytic , mozes _
et al . _
@xcite formulate several conjectures , one of which is discussed in detail below .
also , their results suggest the following conjecture , not explicitly formulated in @xcite : for any two - qudit state @xmath1 that supports a set of @xmath39 encoding unitaries , @xmath10 .
the truth of this conjecture is a consequence of the following result due to wu , cohen , sun , and griffiths @xcite : suppose that alice and bob share a two - qudit entangled state @xmath1 with maximum schmidt coefficient @xmath9 .
let @xmath40 .
if @xmath1 permits alice to send @xmath8 messages to bob via deterministic dense coding , then @xmath41 the work of @xcite suggests that the upper bounds on @xmath42 provided by inequality ( [ wi ] ) can not be decreased when @xmath43 , but that for @xmath12 , the bound should instead be the smaller value @xmath44 . additional analytic work , supporting the validity of the numerical results in @xcite , is provided by ji _
et al . _
@xcite , who have proved that when @xmath1 is non - maximally entangled the greatest number of messages that @xmath1 can support is @xmath45 . also , @xcite itself contains explicit constructions of some families of encoding unitaries that are consistent with numerical data .
for example , the state @xmath46 is shown to support the sending of @xmath47 messages and @xmath48 is shown to support the sending of @xmath49 messages .
numerical work in @xcite suggests that the state ( [ dpo ] ) is the state with minimal entanglement entropy ( and largest @xmath42 value ) that supports the sending of @xmath47 messages and that the state ( [ oh ] ) is the state with minimal entanglement entropy ( and largest @xmath42 value ) that supports the sending of @xmath49 messages .
more generally , based on their numerical work , mozes _
et al _ @xcite conjecture the following : [ mc ] the state ( of a two qudit system ) with minimal entanglement entropy supporting the sending of @xmath50 messages , @xmath51 , is @xmath52 for a fixed value of @xmath53 , it is easy to check that the entanglement entropy ( [ es ] ) of a system in state ( [ gqds ] ) is minimized when @xmath54 ( and @xmath55 .
in addition , the entanglement entropy of @xmath56 decreases as @xmath42 increases ( @xmath53 ) .
thus for a @xmath57 , the state @xmath58 of ( [ mes ] ) represents the state with minimum entanglement entropy for which @xmath59 .
note that the wcsg bounds are pertinent to conjecture m , telling us that if alice wishes to send @xmath60 messages to bob for some @xmath57 , then @xmath58 of ( [ mes ] ) is the minimum - entanglement - entropy state that can possibly allow alice to succeed .
the complete resolution of conjecture m thus depends on showing that for @xmath61 in @xmath62 , there are @xmath50 encoding unitaries for the state @xmath58 . in general ,
when there exists a @xmath1 with @xmath63 that supports @xmath8 encoding unitaries , we will call the wcsg bound @xmath64 _ saturated _ ( a term used in @xcite ) . recall that
_ @xcite have constructed an orthogonal family of @xmath49 encoding operators at @xmath65 .
this construction shows that the wcsg bound @xmath66 is saturated and settles conjecture m in the @xmath67 case .
numerical work in @xcite suggests that all the wcsg bounds provided by inequality ( [ wi ] ) may be saturated except for @xmath68 ( shown not be saturated by ji _
@xcite ) and @xmath69 , which is the largest bound of interest in that it provides information about where the transition from what is possible without entanglement ( sending @xmath27 messages with a qudit ) to what is possible with entanglement ( sending more than @xmath27 messages with a qudit ) .
this paper s principal contributions to deterministic dense - coding theory include the following : * ii : we present an alternate proof of the validity of the wcsg bounds , which provides useful information about any family of @xmath8 encoding unitaries for @xmath1 ( of form ( [ gqds ] ) ) , given that @xmath1 supports @xmath8 fully distinguishable messages and @xmath63 .
this `` boundary information '' is recorded as corollary [ kc ] . using corollary [ kc ] ,
we prove that the wcsg bound @xmath70 is not saturated . * iii : we prove that there are two - qutrit systems with states @xmath71 and @xmath72 for which the entanglement entropy of the first is lower than that of the second , yet the first supports the sending of a greater number of perfectly distinguishable messages than the second .
this outcome , somewhat counterintuitive , was strongly suggested by numerical work of mozes _ et al .
the authors of @xcite attribute this possibility to the fact that the von neumann entropy is an asymptotic quantity whereas the deterministic dense coding process can be carried out with a single entangled pair .
of course , it is the vector of schmidt coefficients of the bipartite - system state @xmath73 that ultimately determines the number of dense - coding messages @xmath1 will support , and in general this vector is not determined by @xmath38 . *
iii & iv : we establish that conjecture m is valid in dimensions @xmath74 and @xmath3 via explicit construction of families of encoding unitaries .
we also prove a non - extendibility result for certain families of encoding unitaries constructed using powers of the @xmath27-dimensional ( unitary ) shift operator @xmath75 , defined by @xmath76 we show that if @xmath77 is any collection of unitary diagonal operators , then the collection @xmath78 , which may serve as a collection of encoding unitaries for any two - qudit state @xmath1 , can not be extended to be a part of a larger family of encoding unitaries for @xmath1 if @xmath79 .
this result , as well as corollary [ kc ] , helped us discover our families of encoding unitaries relevant to conjecture m. * iv : we prove that the @xmath6 and @xmath80 cases of conjecture m are valid for all dimensions ( again , via explicit construction of encoding unitaries ) .
we continue to work with @xmath81 the pure state of a two - qudit system shared by alice and bob , where @xmath1 s schmidt coefficients appear in decreasing order ( [ orderasmpt ] ) . recall that if @xmath82 is a family of encoding unitaries acting on @xmath18 , then @xmath83 is an orthonormal set in @xmath23 , the members of which we call _ orthogonal messages_. throughout the paper , there are occasions when it will be useful to us to make additional assumptions about elements of sets of encoding unitaries .
the following well - known lemma provides an illustration .
[ iif ] if there is a family of @xmath8 encoding unitaries , then there is also a family of @xmath8 encoding unitaries one of which is the identity operator .
suppose that the two - qudit system used for dense coding is in state @xmath1 , given by ( [ gqdsns ] ) .
suppose that there is a family of @xmath8 orthogonal encoding unitaries @xmath82 acting on @xmath18 , so that @xmath84 constitute an orthonormal set in @xmath23 . for each @xmath85
, let @xmath86 because unitary operators preserve inner products , we see that @xmath87 is also an orthonormal set in @xmath23 .
in other words , @xmath88 is also a family of @xmath8 orthogonal encoding unitaries .
thus we have a family of @xmath8 orthogonal encoding unitaries , one of which , @xmath89 , is the identity operator .
wu _ et al . _
@xcite use properties of partial - trace operators to establish the wcsg bounds .
these bounds will be derived differently here , in a way that enables us to obtain information about any family of @xmath8 encoding unitaries for @xmath1 if @xmath63 .
assume there are @xmath90 unitary operators @xmath91 acting on @xmath18 such that @xmath92 , form an orthonormal set in @xmath23 .
we show @xmath10 .
let @xmath93 be the subspace of @xmath23 spanned by the orthonormal set @xmath94 so that @xmath95 is the projection operator for states in @xmath23 onto @xmath93 .
let @xmath96 .
consider @xmath97 from which it follows that the square of the norm of @xmath98 is given by @xmath99 sum both sides of the preceding equation from @xmath100 to @xmath101 : @xmath102 where to obtain the final equality we have used @xmath103 , which follows from the unitarity of @xmath104 . because @xmath105 is a projection operator , @xmath106 ; thus , @xmath107 so that @xmath10 , the desired result . note that by equation ( [ tgp ] ) if there are @xmath8 encoding unitaries and @xmath63 , then @xmath108 since @xmath109 for each @xmath110 , this can happen only if @xmath111 for each @xmath110 .
since @xmath112 is a unit vector , @xmath113 implies @xmath114 ; equivalently , @xmath112 belongs to @xmath93 .
thus the following is a corollary of our proof of the wcsg bounds .
[ kc ] let @xmath115 , let @xmath63 , and let @xmath116 where @xmath9 is the largest schmidt coefficient of @xmath1 .
suppose that there are @xmath8 unitary operators @xmath91 acting on @xmath18 such that @xmath117 is an orthonormal set in @xmath23 .
then the linear span @xmath93 of the set @xmath118 contains @xmath112 for each @xmath119 .
as our first application of the preceding corollary , we prove the wcsg bound @xmath70 is not saturated . [ bns ] if the state @xmath120 supports the sending of @xmath121 orthogonal messages via deterministic dense coding , then @xmath1 s largest schmidt coefficient , @xmath9 , must satisfy @xmath122 assume there are @xmath12 encoding unitaries @xmath123 , so that @xmath124 form an orthonormal set in @xmath23 . by lemma [ iif ] , we can and
do assume that @xmath125 .
suppose , in order to obtain a contradiction , that @xmath126 then corollary [ kc ] tells us that the @xmath47 dimensional subspace @xmath93 spanned by @xmath127 contains the @xmath27 orthonormal vectors @xmath128 .
the subspace @xmath93 also contains @xmath129 subtracting @xmath130 , which is in @xmath93 , from @xmath131 , we see @xmath132 belongs to @xmath93 and that @xmath133 , being an orthogonal set of @xmath121 elements in @xmath93 , must form an orthogonal basis for @xmath93 .
let @xmath32 be an arbitrary element in @xmath134 .
consider @xmath135 since @xmath136 is in @xmath93 and @xmath137 is also in @xmath93 ( being a linear combination of @xmath138 ) , we see , by subtraction , @xmath139 belongs to @xmath93 . since @xmath140 is an orthogonal basis for @xmath93 and @xmath141 is orthogonal to all elements of @xmath140 except @xmath142 , there is a scalar @xmath143 such that @xmath144 ; that is , @xmath145 since @xmath146 is not zero @xmath147 , the preceding equation shows that the second column of the matrix @xmath148 , representing @xmath149 with respect to the basis @xmath150 , consists of @xmath151 s except for its second entry , which is @xmath143 .
hence @xmath152 .
since the second column of @xmath148 is orthogonal to its first column , we see the first column of @xmath148 must have a zero as its second entry . in other words , @xmath153 .
since @xmath154 is arbitrary , and @xmath155 is the identity , we have @xmath156 for @xmath157 .
it follows that @xmath158 for @xmath159 ; and thus @xmath160 is orthogonal to @xmath93 and in @xmath93 , a contradiction , completing the proof .
in this section , we focus on deterministic dense coding based on a two - qutrit system . in this context
, we prove that a state with lower entanglement entropy can support the sending of a greater number of perfectly distinguishable messages via deterministic dense coding than a state with higher entanglement entropy .
we consider a two - qutrit system , shared by alice and bob , in state @xmath161 where as usual , we assume @xmath162 via primarily numerical methods , mozes _
et al . _
@xcite assess how many orthogonal messages to bob , alice can generate , given their shared system is in state @xmath1 .
their results are nicely summarized in figure [ mozesd ] , reproduced from @xcite . '' of orthogonal encoding unitaries that the state @xmath1 of ( [ gqts ] ) will support , as a function of @xmath42 and @xmath146 .
the equation @xmath163 defines the lower side of the triangular region bounding the numbered areas while region s upper side is piecewise defined as @xmath164 for @xmath165 and @xmath166 for @xmath167 .
[ mozesd],height=240 ] the figure suggests that if @xmath42 equals , say , 0.51 , then @xmath1 of ( [ gqts ] ) will support 5 orthogonal unitaries independent of the values of @xmath146 and @xmath168 ( satisfying ( [ sum1 ] ) ) .
note the role that the wcsg bounds play in the figure : it appears that @xmath169 and @xmath170 can be saturated ( as we explained in the introduction , @xmath65 is already known to be saturated ) and that @xmath171 is not saturated ( which is proved in proposition [ bns ] above ) .
note also that the boundary line between the 5 region and the 4 region appears to be curved slightly allowing its top point , @xmath172 , to possibly be in the 5 region while the point on the bottom boundary line directly below it , @xmath173 , is not in the 5 region .
this is precisely what we prove below , and since the state corresponding to the top point , @xmath174 has lower entanglement entropy than the bottom - line point @xmath175 we have a proof that a state with lower entanglement entropy can support more messages than one with higher entanglement entropy .
a plot of entanglement entropy over the triangular region of figure [ mozesd ] appears as figure [ entropyplot ] .
computed from equation ( [ es ] ) for the state @xmath176 as @xmath42 and @xmath146 vary over the triangular region displayed in figure [ mozesd],height=259 ] a matrix point of view will facilitate our work . as in the preceding sections ,
let @xmath177 be an encoding unitary operator on @xmath18 , a space which has orthonormal basis @xmath178 , where we have dropped the subscripts of @xmath179 on the basis vectors .
_ we will not distinguish between @xmath177 and its representation as a @xmath180 matrix with respect to @xmath140_. we will consider the @xmath181 entry of @xmath177 to be the entry in the @xmath85-th row and @xmath61-th column of @xmath177 so that @xmath182 we denote the collection of all @xmath180 matrices with complex entries @xmath183 .
consider @xmath42 , @xmath146 , and @xmath168 to be fixed nonnegative ( real ) numbers satisfying ( [ sum1 ] ) .
let @xmath184}.\ ] ] we say that the unitary matrices @xmath185 and @xmath177 in @xmath183 are @xmath186-_orthogonal _ provided that @xmath187 we say a set of unitary matrices is @xmath186-orthogonal provided its elements are pairwise @xmath186-orthogonal . the following proposition is easily verified ( and appears as equation ( 7 ) in @xcite ) . [ op ] the matrices @xmath185 and @xmath177 in @xmath183 are @xmath186-orthogonal if and only if @xmath188 and @xmath189 are orthogonal vectors in @xmath23 .
we consider unitary matrices in @xmath183 to correspond to encoding operators that alice may apply to her qutrit from the pair she shares with bob .
given their qutrit pair is in state @xmath1 of ( [ gqts ] ) , proposition [ op ] shows that the messages alice encodes with unitary matrices @xmath185 and @xmath177 will be perfectly distinguishable by bob if and only if @xmath185 and @xmath177 are @xmath186-orthogonal .
consider , for example , the situation where the two - qutrit system shared by alice and bob is fully entangled . here
@xmath186 would be the diagonal matrix with diagonal entries @xmath190 and it is easy to check that @xmath191 is a @xmath186-orthogonal family of 9 encoding unitaries , where @xmath192 is the @xmath180 identity matrix , while @xmath193 and @xmath194 are , respectively , shift and phase operators given by @xmath195 } \ ; \ \ z = { \left[\begin{array}{ccc}1&0&0\\#4&\exp\left(\frac{2\pi i}{3}\right)&0\\#7&0&\exp\left(\frac{4\pi i}{3}\right)\end{array}\right]}.\ ] ] also easy to check is that @xmath196 is @xmath186-orthogonal independent of the values of @xmath42 , @xmath146 , and @xmath168 .
in fact , @xmath186-orthogonality for @xmath196 arises in the simplest possible way : @xmath193 , @xmath197 and @xmath198 ( which equals @xmath193 ) each have main diagonal consisting of all zeros ; when one multiplies such matrices by a diagonal matrix ( on either the right or left ) , zeros remain on the main diagonal so that the @xmath186-orthogonality of @xmath196 is clear .
as is pointed out in @xcite , the special properties of @xmath196 can be exploited to build @xmath186-orthogonal families of six unitaries if @xmath199 : specifically , for each @xmath186 of the form ( [ lform ] ) with @xmath199 , there is a unitary diagonal matrix @xmath200 such that @xmath201 will be @xmath186-orthogonal .
our first result in the qutrit context shows , however , that if @xmath79 then the family @xmath196 can not be expanded to a larger @xmath186-orthogonal family of unitaries .
we prove something a bit more general : namely that if @xmath202 and @xmath203 are diagonal unitary matrices and @xmath79 , then @xmath204 can not be extended to a @xmath186-orthogonal family of 4 or more unitaries .
one can think of @xmath205 and @xmath206 as the two basic forms that a @xmath180 unitary matrix may assume if its main diagonal consists only of @xmath151 s .
the family of five @xmath186-orthogonal unitaries that we construct below , for @xmath172 , contains no matrix having all zeros along its main diagonal .
suppose that @xmath207 and that @xmath202 and @xmath203 are diagonal unitary @xmath180 matrices ; then there is no unitary matrix @xmath177 such that the family @xmath208 is @xmath186-orthogonal .
let @xmath79 .
let @xmath202 and @xmath203 be diagonal unitary matrices with @xmath202 having diagonal entries @xmath209 , @xmath210 , @xmath211 and @xmath203 having diagonal entries @xmath212 , @xmath213 , @xmath214 . because @xmath202 and @xmath203 are unitary , each of the @xmath215 s and @xmath216 s has modulus 1 .
suppose , in order to obtain a contradiction , that @xmath177 is a unitary matrix such that @xmath217 is @xmath186-orthogonal .
then , we have @xmath218 , @xmath219 , @xmath220 , and these three equations expand , respectively , to @xmath221 the preceding system of equations may be written in the following vector form @xmath222 where , @xmath223 } , { { \rm \bf v}}_2 = { \left[\begin{array}{c}u_{22}\\#2\\#3\end{array}\right]},\ { \rm and}\ { { \rm \bf v}}_3 = { \left[\begin{array}{c}u_{33}\\#2\\#3\end{array}\right]}.\ ] ] since @xmath177 is unitary , each of its columns is a unit vector and it follows that @xmath224 , and @xmath225 are also unit vectors . taking the norm of both sides of ( [ vf ] ) and applying the triangle inequality , we obtain @xmath226 but we are assuming @xmath79 , which , since @xmath227 , makes @xmath228 . thus ( [ pc ] ) provides a contradiction , completing the proof .
\(1 ) the preceding non - extendability result , as well as the argument that yields it , easily generalizes to dimensions @xmath229 .
thus if @xmath75 is the @xmath27-dimensional shift given by ( [ xd ] ) and if @xmath77 is any collection of unitary diagonal operators , then the family @xmath78 , which may serve as a collection of encoding unitaries for any two - qudit state @xmath1 , can not be extended to be a part of a larger family of encoding unitaries for @xmath1 if @xmath79 .
other results in this section , such as lemma [ til ] and proposition [ fr ] below , also have obvious higher - dimensional generalizations .
( 2 ) for the case @xmath74 , cohen @xcite has established the following non - extendability result : suppose that @xmath230 and @xmath200 is a diagonal unitary matrix chosen so that @xmath231 is @xmath186-orthogonal ; then there is no unitary matrix @xmath177 such that @xmath232 is also @xmath186-orthogonal .
[ til ] suppose that @xmath233 ; then no @xmath186-orthogonal family of unitary matrices may contain the identity and a diagonal matrix distinct from the identity .
suppose that @xmath233 and @xmath192 and @xmath177 are @xmath186-orthogonal for a diagonal unitary matrix @xmath177 .
then @xmath234 since @xmath177 is unitary and diagonal , each of @xmath235 , and @xmath236 has modulus @xmath237 . rewriting equation ( [ iuo ] ) as @xmath238 , taking absolute values of both sides , and applying the triangle inequality , we obtain @xmath239 , which can not happen if @xmath79 . in our current context ,
lemma [ iif ] implies the following : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ if there exists a @xmath186-orthogonal family of @xmath8 unitary matrices in @xmath183 , then there also exists a @xmath186-orthogonal family of @xmath8 unitary matrices in @xmath183 one of which is the identity matrix .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ we depend upon a refinement of the preceding result that holds when @xmath240 .
( note the condition @xmath240 defines the lower line of the triangle in figure [ mozesd ] . )
[ fr ] suppose that @xmath240 and that there exists a @xmath186-orthogonal family of @xmath8 unitary matrices in @xmath183 .
then there also exists a @xmath186-orthogonal family of @xmath8 unitary matrices in @xmath183 one of which is the identity and another of which has @xmath151 as its @xmath241 entry . under the hypotheses of this proposition , we know that there is a family of @xmath8 , @xmath186-orthogonal unitaries in @xmath183 containing the identity matrix .
let this family be @xmath242 as before , let @xmath186 be the diagonal matrix with diagonal entries @xmath42 , @xmath146 , and @xmath168 . because @xmath240 , it is easy to check that @xmath186 commutes with any matrix of the form @xmath243}.\ ] ] set @xmath244}\ ] ] where @xmath245 } \
{ \rm is\ a\ unit\ vector\ chosen\ to
\ be\ orthogonal\ to}\ { |u\rangle } \equiv{\left[\begin{array}{c}u_{12}\\#2\end{array}\right]}.\ ] ] here , of course , @xmath246 is the @xmath247 entry in the matrix @xmath177 from the family @xmath248 .
note that @xmath249 is unitary and the @xmath241 entry of @xmath250 is @xmath251 .
we claim that the family obtained from @xmath248 by left multiplication by @xmath249 and right multiplication by @xmath252 satisfies the requirements of the proposition : @xmath253 consists of unitary matrices , the first element listed is the identity , the second has @xmath151 as its @xmath241 entry , and we claim that its elements are @xmath186-orthogonal .
consider , for example , @xmath254 where we have used the fact that @xmath249 commutes with @xmath186 , the cyclicity property of the trace , and the @xmath186-orthogonality of @xmath248 to obtain the final three equalities . set @xmath255 } \
{ \rm and}\ \lambda_l = { \left[\begin{array}{ccc}3/5&0&0\\#4&2/5&0\\#7&0&0\end{array}\right]}.\ ] ] where we have used the subscripts @xmath256 and @xmath257 because , as we indicated earlier , of the corresponding pair of states @xmath72 and @xmath71 , the state @xmath72 has the higher entanglement entropy and @xmath71 , the lower .
the following two propositions are the major results of this section , the first showing that @xmath71 supports transmission of five orthogonal messages and the second showing that @xmath72 supports at most four .
[ fivefam ] there are five @xmath258-orthogonal unitary matrices .
let @xmath259 } , u = { \left[\begin{array}{ccc}-\frac23&0&\frac{\sqrt{5}}{3}\\#4&1&0\\#7&0&-\frac{2}{3}\end{array}\right]},\ { \rm and}\ m = { \left[\begin{array}{ccc}-\frac13&-\frac{\sqrt{3}}{2}i&\frac{\sqrt{5}}{6}\\#4&\frac12&-\frac{\sqrt{5}}{2\sqrt{3}}i\\#7&0&\frac23\end{array}\right]}.\ ] ] let @xmath260 , where @xmath192 is the @xmath180 identity matrix and @xmath261 is the matrix obtained from @xmath185 by taking the complex conjugate of each of its entries .
the reader may verify that @xmath248 is @xmath258-orthogonal ( and consists of unitaries ) , completing the proof .
the family @xmath248 of five @xmath258-orthogonal matrices presented in the proof of the preceding proposition was constructed with the assistance of corollary [ kc ] .
we began with the assumption that there is a family @xmath262 of five @xmath258-orthogonal unitaries , where @xmath192 and @xmath179 are as in the proof of proposition [ fivefam ] and @xmath263 , @xmath264 , and @xmath265 were to be constructed ( if possible ) .
thus we knew that the linear span @xmath93 of the unitaries in @xmath248 applied to @xmath71 had to include the vectors @xmath266 , @xmath267 , and , via corollary [ kc ] , the vectors @xmath268 and @xmath269 . because @xmath93 is closed under linear combinations , @xmath270 $ ] and @xmath271 $ ] , were also necessarily in @xmath93 .
we concluded that @xmath272 is an orthonormal basis for the ( five dimensional ) subspace @xmath93 .
it follows that @xmath273 must be orthogonal to @xmath93 , which means in particular that for @xmath274 , @xmath275 so that every encoding matrix in @xmath248 necessarily had @xmath151 as its @xmath276 entry . with this knowledge and with some algebraic work , we arrived at the family @xmath248 of proposition [ fivefam ] .
the number of @xmath277-orthogonal unitary matrices is less than or equal to four . because @xmath170 for a system in state @xmath72 , the inequality ( [ wi ] ) for the wcsg bounds shows that there are at most @xmath278 @xmath277-orthogonal unitary matrices .
suppose , in order to obtain a contradiction , that there is a @xmath277-orthogonal family of 5 unitary matrices : @xmath279 .
we assume that @xmath177 has @xmath151 as its @xmath241 entry , which we may do by proposition [ fr ] .
let @xmath93 be the linear span of @xmath280 and note that @xmath93 is a five - dimensional space . by corollary [ kc ]
, @xmath93 must contain @xmath281 , and @xmath269 , and hence @xmath93 also contains @xmath282 now consider @xmath283 , which too belongs to @xmath93 ; thus , @xmath284 also belongs to @xmath93 .
recall @xmath285 .
we claim that one of @xmath286 must be nonzero .
otherwise @xmath287 and @xmath236 are the only nonzero entries in , respectively , the second and third columns of @xmath177 and , since @xmath177 is unitary , it follows that @xmath287 and @xmath236 are unimodular and that @xmath177 is diagonal .
this contradicts lemma [ til ] .
thus at least one of @xmath286 is nonzero and it follows that @xmath288 is linearly independent .
being a linearly independent set of 5 elements in the 5 dimensional space @xmath93 , we see that @xmath140 is a basis for @xmath93 .
thus we may express @xmath289 as a linear combination of the elements of @xmath140 : @xmath290 it follows that @xmath185 must have the form @xmath291}.\ ] ] we claim that both @xmath143 and @xmath292 are nonzero . because @xmath185 is unitary ,
its columns form an orthonormal basis of @xmath293 .
thus @xmath143 and @xmath292 can not both be zero , because that would make both the second and third columns of @xmath185 the zero vector .
suppose that @xmath294 .
then @xmath185 would have to be a diagonal matrix , contradicting lemma [ til ] .
suppose that @xmath295 .
then the rightmost two columns of @xmath185 are simply @xmath143 times the corresponding columns of @xmath177 . since both @xmath177 and @xmath185 are unitary , @xmath152 .
the first column of @xmath185 is determined to within a multiplicative constant by the other two , and thus can be written as a unimodular constant @xmath296 times the first column of @xmath177 .
checking the @xmath186-orthogonality of @xmath185 and @xmath177 gives @xmath297 since @xmath298 , the preceding trace can not be @xmath151 , a contradiction .
therefore , we may assume that @xmath185 has the form ( [ mform ] ) , where both @xmath143 and @xmath292 are nonzero .
we can say more about the form of @xmath185 .
because @xmath185 and @xmath177 are both @xmath186-orthogonal to @xmath192 , we have @xmath299 and @xmath300 combining these equations yields @xmath301 because the second and third columns of @xmath185 are orthogonal , we have @xmath302 because the second and third columns of @xmath177 are orthogonal , we conclude from the preceding equation that @xmath303 , from which follows that @xmath304 , where we have used the fact that both c and q are nonzero . since the second column of @xmath177 is a unit vector and @xmath285 , @xmath305 so that @xmath306 .
it follows that @xmath307 .
thus , because @xmath308 is also 0 , the inner product of the first two columns of @xmath177 is @xmath309 , which must equal 0 .
there are two possibilities : either @xmath310 or @xmath311 .
we show each of these possibilities leads to a contradiction , which completes the proof of the theorem .
suppose that @xmath310 ; then the only nonzero entry in the first column of @xmath177 is the first , which makes @xmath312 .
this is a contradiction because if @xmath313 is unimodular , then @xmath177 can not be @xmath277-orthogonal to @xmath192 .
if it were , @xmath314 which implies @xmath315 , a contradiction .
thus we must have @xmath311 and @xmath177 takes the form @xmath316}.\ ] ] note that @xmath317 and @xmath318 must have the same form as does @xmath185 : @xmath319},\ ] ] where @xmath143 , @xmath292 , and @xmath320 are constants that depend on which of @xmath185 , @xmath317 , and @xmath318 is so represented .
now , view the @xmath180 matrices @xmath321 , and @xmath318 as vectors in the nine dimensional vector space @xmath183 .
because @xmath322 is a @xmath186-orthogonal set , the vector subspace @xmath93 of @xmath183 spanned by @xmath323 is four dimensional .
however , given the forms of these matrices from ( [ uform ] ) and ( [ mform ] ) , @xmath93 is also spanned by the 3 vectors in the following set @xmath324 } , { \left[\begin{array}{ccc}u_{11}&0&u_{13}\\#4&u_{22}&0\\#7&0&u_{33}\end{array}\right ] } , { \left[\begin{array}{ccc}0&0&0\\#4&0&0\\#7&0&0\end{array}\right]}\right\},\ ] ] which is the contradiction that completes the proof .
proposition [ fivefam ] of the preceding section shows that the state @xmath325 is the two - qutrit state with minimum entanglement entropy supporting the sending of 5 orthogonal messages through deterministic dense coding .
thus the proposition resolves conjecture m in dimension @xmath2 and shows that the wcsg qutrit bound @xmath170 is saturated .
we now resolve conjecture m for @xmath326 , which entails showing the wcsg bounds of @xmath329 and @xmath330 are saturated .
let @xmath331 we need to exhibit @xmath332 orthogonal encoding unitaries for @xmath333 with @xmath334 and to exhibit @xmath335 orthogonal encoding unitaries for @xmath333 with @xmath330 .
we represent these encoding unitaries as @xmath336 matrices ( with respect to the basis @xmath337 ) .
the encoding unitary matrices for @xmath333 , say @xmath263 and @xmath264 , need to be @xmath186-orthogonal ; i.e. , @xmath338 , where @xmath186 is the @xmath336 diagonal matrix whose diagonal entries , in order , are @xmath339 . because only the initial two diagonal entries of @xmath186 are nonzero , only the first two columns of unitary encoding matrices are relevant to @xmath186-orthogonality calculations .
we record the full matrices , however , in order for the reader to begin to see patterns that lead to higher - dimensional generalizations . in all constructions below ,
corollary [ kc ] was used to obtain information about the form of members of encoding families of unitaries . for @xmath334
, our family of six @xmath186-orthogonal encoding unitaries for @xmath333 is given by @xmath340 , where @xmath192 is the identity and the remaining members of the family are @xmath341 , u_1(4 ) = \left[\begin{array}{cccc}-\frac12&0 & \frac{\sqrt{3}}{2}&0 \\
0&1&0&0\\ -\frac{\sqrt{3}}{2}&0&-\frac12&0\\ 0&0&0&1\end{array}\right ] , u_2(4 ) = \left[\begin{array}{cccc}-\frac12&0 & \frac{\sqrt{3}}{2}&0\\ 0&1&0&0\\ \frac{\sqrt{3}}{2}&0&\frac12 & 0\\ 0&0&0&1\end{array}\right],\ ] ] @xmath342 , v_2(4 ) = \left[\begin{array}{cccc}0&1 & 0&0\\ -\frac12&0&\frac{\sqrt{3}}{2}&0\\ 0&0&0&1\\ \frac{\sqrt{3}}{2}&0&\frac12&0\\\end{array}\right].\ ] ] for @xmath330 , the following is a @xmath186-orthogonal family of 7 unitary matrices : @xmath343 , where @xmath192 is the identity and the remaining members of the family are @xmath344 , a_2 = \left[\begin{array}{cccc } 0&1&0&0\\ 0&0&1&0\\1&0&0&0\\0&0&0&1\end{array}\right ] , u = \left[\begin{array}{cccc } -\frac{3}{4}&0&0&\frac{\sqrt{7}}{4}\\ 0&1&0&0\\0&0&1&0\\-\frac{\sqrt{7}}{4}&0&0&-\frac{3}{4}\end{array}\right],\ ] ] @xmath345 , \
{ \rm where}\ \omega_3 = \exp(2\pi i/3)\ { \rm and}\ j = 0 , 1 , 2.\ ] ] we now generalize the preceding constructions , i.e. those for the @xmath327 and @xmath328 wcsg bounds , to all dimensions @xmath346 .
we continue to assume that @xmath186 is a diagonal matrix whose 1,1 entry is @xmath42 , whose 2,2 entry is @xmath347 , and all of whose other entries are 0 s . in this subsection , we construct @xmath186-orthogonal families of @xmath348 unitaries for @xmath349 . in the next , we construct @xmath186-orthogonal families of @xmath350 unitaries for @xmath351 .
let @xmath352 . for @xmath3 ,
the family @xmath353 , exhibited above , saturates the @xmath328 bound .
note that in @xmath353 , we can view the triple @xmath354 , and @xmath355 as @xmath356 and @xmath355 where @xmath357 is the block diagonal matrix having the @xmath61-th power of the @xmath358 shift matrix , @xmath359 , in its upper left - hand corner and the @xmath360 matrix @xmath361 $ ] in its lower right - hand corner ( with zeros elsewhere ) .
the @xmath362-dimensional shift operator @xmath363 , defined by ( [ xd ] ) , will play a similar role in our constructions for @xmath364 .
we are now in a position to show that the wcsg bound @xmath349 is saturated for every @xmath365 .
the construction for odd dimensions differs from that for even dimensions . in what follows
we will frequently describe only the first two columns of unitary encoding matrices . as long as
these two columns are orthonormal , there is a unitary matrix that contains them .
moreover , since only the first two entries of @xmath186 are nonzero , only the first two columns of an encoding matrix are relevant to @xmath186-orthogonality calculations .
suppose that @xmath365 is odd .
we claim that a family of @xmath370 orthogonal encoding unitaries for the state @xmath371 is given by @xmath372 , where @xmath357 is the @xmath373 block diagonal matrix with @xmath374 occupying its upper left - hand corner and the matrix @xmath361 $ ] occupying its lower right - hand corner ( with zeros in other locations ) and where the first two columns of @xmath177 and @xmath375 are given by @xmath376 note that all but the first and last entries in the first column of @xmath375 may be defined by the following formula : @xmath377}\frac{1}{\sqrt{d}}\ , i^k{{\omega_{d-1}}}^{kj } , \ \ k = 1 , 2 , \ldots , d-2,\ ] ] where @xmath378 $ ] represents the greatest integer less than or equal to @xmath379 . note also that the second column of @xmath375 is determined by the first column as follows ; @xmath380 and for @xmath381 , @xmath382 ( while @xmath383 ) .
observe that these relationships between the entries in the first and second columns of @xmath375 are precisely what make @xmath375 , for any given @xmath384 , @xmath186-orthogonal to each @xmath385 , @xmath386 , where @xmath349 .
it is easy to see that @xmath387 is @xmath186-orthogonal , in fact @xmath388 will have zeros as its @xmath389 and @xmath390 entries as long as @xmath391 .
the form of @xmath177 makes it easy to see that @xmath392 independent of @xmath61 . also easy to see
is that @xmath393 independent of @xmath394 ( @xmath395 will have @xmath151 s as its 1,1 and 2,2 entries for @xmath396 ) .
clearly the first two columns of @xmath177 are unit vectors and are orthogonal so that @xmath177 extends to be a unitary @xmath373 matrix . also , it s easy to see that the first two columns of @xmath375 are orthogonal unit vectors .
( for the orthogonality calculation , it is easier to take the conjugate - transpose of the 2nd column times the first ; terms will go to zero pairwise , the first by using the fact that the complex conjugate of @xmath397 is @xmath398 ) . thus , to complete our verification of the claim that @xmath399 is a @xmath186-orthogonal family of unitaries , we must check that for distinct @xmath394 and @xmath400 , @xmath401 is @xmath186-orthogonal to @xmath402 .
we have @xmath403^m + \frac{2d-1}{d^2}\right ) \\+ \frac{d-1}{2d-1}\left(\frac{1}{(d-1)^2 } + \frac{d}{(d-1)^2}\sum_{m=1}^{d-2 } \left[{{\omega_{d-1}}}^{\ell - k}\right]^m\right ) .
\end{split}\ ] ] the sums over @xmath110 on the right of the preceding equality are @xmath404 by ( [ cmplx ] ) and the right side thus reduces to @xmath405 , as desired .
we continue to assume @xmath349 .
the construction of a @xmath186-orthogonal family of @xmath370 encoding unitaries is more difficult when @xmath27 is even : the @xmath336 case does nt entirely fit the general pattern and roots of unity of order d-3 as well as d-1 appear in the `` @xmath375 '' construction ( because the pairwise cancellation that made the columns of @xmath375 orthogonal in the case of odd @xmath27 can not work when @xmath27 is even ) .
let @xmath406 be even .
we claim that a family of @xmath370 , @xmath186-orthogonal encoding unitaries is given by @xmath407 , where , as before , @xmath357 is the @xmath373 block diagonal matrix with @xmath374 occupying its upper left - hand corner and @xmath361 $ ] occupying its lower right - hand corner , and where the first two columns of @xmath177 are given in ( [ adodd ] ) and the first two columns of @xmath375 are given by @xmath408 note that all entries in the first column of @xmath375 except the first and the last are given by the following formula @xmath409 the second column of @xmath375 is determined by the first in the same way as before ( i.e. , in the @xmath27 is odd case ) , and , just as before , this means that @xmath375 , for any given @xmath384 , is @xmath186-orthogonal to each @xmath385 , @xmath410 . just as before the @xmath357 s are ( pairwise ) @xmath186-orthogonal as are the @xmath357 s and @xmath177 as well as @xmath177 and the @xmath375 s .
it is easy to check that the first two columns of @xmath375 are unit vectors . to verify that they are orthogonal we compute the conjugate transpose of the second column times the first : @xmath411 \rule{1in}{0 in } \\ - \frac{1}{d-1}{{\omega_{d-1}}}^j \left(\sum_{m=0}^{\frac{d}{2}-2}{{\omega_{d-3}}}^m + \sum_{m=1}^{\frac{d}{2}-2}({{\omega_{d-3}}}^*)^m\right ) .
\end{split}\ ] ] the expression in square brackets is zero since @xmath412 . also , using @xmath413 we see that the second sum over @xmath110 in ( [ evenp ] ) equals @xmath414 and so both sums over @xmath110 in ( [ evenp ] ) combine to the sum of @xmath415 from @xmath100 to @xmath416 , which is 0 by ( [ cmplx ] ) .
thus the columns of @xmath375 are indeed orthogonal . to complete the verification that @xmath417 is @xmath186-orthogonal , we must check that for distinct @xmath394 and @xmath400 , @xmath401 is @xmath186-orthogonal to @xmath402 .
the calculation here is the same as that for the case of odd @xmath27 , the point being that in the product @xmath418 all @xmath419 roots of unity will be multiplied by their conjugates and thus reduce to @xmath237 .
we now turn to the task of establishing that the wcsg bound of @xmath420 is saturated for all @xmath365 . for @xmath421 , the saturation of this bound is quite well known , e.g , the @xmath422 identity matrix , together with the pauli matrices constitute a @xmath186-orthogonal family of @xmath423 encoding unitaries . recall that we have already established saturation for the @xmath74 ( proposition [ fivefam ] ) and @xmath3 cases ( via the family @xmath424 exhibited earlier in this section ) . because our construction of orthogonal unitaries for the @xmath425 bound is inductive , we modify our notation to reflect dependence on dimension @xmath27 . for the remainder of this section , in what follows , we use @xmath428_j$ ] to denote the @xmath61-th column of the matrix @xmath185 , @xmath429 to denote the @xmath430 identity matrix , @xmath431 to denote the matrix obtained by interchanging the first two columns of @xmath429 , and @xmath432 to denote a column containing @xmath32 zeros
. a permutation matrix plays a role in our work : @xmath433 is a ( unitary ) permutation matrix such that @xmath434 moves the second row of @xmath185 to the last row and shifts all rows initially beneath the second row up one .
finally , in all constructions below , we continue to exhibit only the first two columns of our matrices ; if these columns are orthonormal , the given matrix may be extended to be unitary .
having already defined @xmath424 , for each even @xmath406 we define , inductively , @xmath435 where @xmath436_1&{{\bf 0}}_{\frac{d}{2}-1}\\ { { \bf 0}}_{\frac{d}{2}-1 } & { { \bf 0}}_{\frac{d}{2}-1 } \end{array}\right .
, v_{d/2}(d)=\left[\begin{array}{cc}0&1\\-\frac{2}{d}&0\\ { { \bf 0}}_{\frac{d}{2 } - 1 } & { { \bf 0}}_{\frac{d}{2 } - 1}\\ \sqrt{1-\left(\frac{2}{d}\right)^2}\ \left[i_{\frac{d}{2 } - 1}\right]_{\frac{d}{2}-1 } & { { \bf 0}}_{\frac{d}{2}-1}\end{array } \right.,\ ] ] and where the remaining @xmath437 matrices in the @xmath177 and @xmath438 collections are constructed from the corresponding collections from the family @xmath439 as follows : for @xmath440 , @xmath441_1&{{\bf 0}}_{d-2 } \end{array}\right . ,
v_j(d)=\left[\begin{array}{cc}0&1\\-\frac{2}{d}&0\\\sqrt{1-(\frac{2}{d})^2}{{r}}\left[v_j(d-2)\right]_1&{{\bf 0}}_{d-2}\end{array}\right .. \ ] ] to illustrate how this inductive process works , we build the family @xmath442 using the family @xmath424 presented earlier in this section .
the family @xmath442 consists of @xmath443 , @xmath444 , @xmath445 & \begin{array}{c}0\\0\\0\\0\end{array } \end{array}\right.\ ] ] @xmath446 & \begin{array}{c}0\\0\\0\\0\end{array } \end{array}\right .
, v_1(6 ) = \left[\begin{array}{cc}0 & 1 \\ -\frac13&0\\ \frac{\sqrt{8}}{3}\left[\begin{array}{c}0\\0\\-\frac{\sqrt{3}}{2}\\ -\frac12 \end{array}\right ] & \begin{array}{c}0\\0\\0\\0\end{array } \end{array}\right . , v_2(6 ) = \left[\begin{array}{cc}0 & 1 \\ -\frac13&0\\ \frac{\sqrt{8}}{3}\left[\begin{array}{c}0\\0\\ \frac{\sqrt{3}}{2 } \\ -\frac12 \end{array}\right ] & \begin{array}{c}0\\0\\0\\0\end{array } \end{array}\right .
.\ ] ] the reader may verify directly that the family @xmath442 exhibited above is @xmath447-orthogonal and consists of unitary matrices .
we now turn to the proof that the family @xmath448 defined by ( [ ddte ] ) is @xmath426 orthogonal for every even @xmath406 .
our formal , inductive argument begins with the @xmath3 case .
we have exhibited a family @xmath424 of six @xmath449-orthogonal encoding unitaries for @xmath329 of the form @xmath450 .
suppose that for some even @xmath451 the collection @xmath452 is @xmath426-orthogonal and consists of unitaries .
we claim that the family @xmath453 is a @xmath454-orthogonal family of unitaries .
establishing this claim , completes the proof .
clearly the first two columns of @xmath455 and @xmath456 , defined by ( [ ufvf ] ) , are orthonormal ( and hence these columns may be augmented to create unitary matrices ) . for @xmath457 and @xmath458 , defined by ( [ ujvj ] ) orthogonality of the first two columns is obvious and normality follows from the fact that @xmath459_1 $ ] and @xmath460_1 $ ] are unit vectors .
we must verify @xmath454-orthogonality of the members of @xmath461 .
clearly @xmath462 and @xmath463 are @xmath454-orthogonal .
it is easy to see , due to the placement of zeros , that each @xmath457 matrix will be @xmath454-orthogonal to each @xmath464 matrix ; @xmath465 .
it is easy to check that each @xmath457 as well as each @xmath464 matrix is @xmath454-orthogonal to both @xmath462 and @xmath463 .
the remaining orthogonality checks require a little more effort .
let @xmath466 ( so that , e.g. , the 1,1 entry in the first column of each @xmath457 matrix becomes @xmath467 ) .
let @xmath468 , @xmath54 , and let @xmath85 and @xmath61 be distinct elements in @xmath469 .
we have @xmath470_1)^\dagger({{r}}[u_j(d)]_1)\right)+\lambda_1\\ & = & \frac{d+2}{d+4}\left(\frac{1}{(n+1)^2}+\left(\frac{n^2 + 2n}{(n+1)^2}\right)([u_i(d)]_1^\dagger [ u_j(d)]_1\right)+\frac{2}{d+4}.\label{oce}\end{aligned}\ ] ] since @xmath471 and @xmath472 are @xmath473-orthogonal at @xmath474 , @xmath475_1^\dagger[u_j(d)]_1\right)+\frac{2}{d+2}=0,\ ] ] from which it follows that @xmath476_1^\dagger[u_j(d)]_1=-\frac{1}{n}$ ] . substituting @xmath477 for @xmath476_1^\dagger[u_j(d)]_1 $ ] in ( [ oce ] ) and doing
a bit of algebra reveals that @xmath478 , as desired .
an argument essentially the same as the preceding one shows that @xmath479 and @xmath458 are @xmath454-orthogonal ( given @xmath85 and @xmath61 are distinct elements in @xmath469 ) .
let @xmath480 .
it remains to test the @xmath454-orthogonality of @xmath481 and @xmath457 ( as well as of @xmath482 and @xmath458 ) .
we have @xmath483 and essentially the same calculation shows @xmath484 .
this completes the verification that @xmath461 is a @xmath454-orthogonal family of unitary matrices .
it follows that for every even @xmath485 , the bound @xmath425 is saturated .
we now argue that the bound @xmath425 is saturated when @xmath346 is odd .
we construct @xmath426-orthogonal families of @xmath350 unitary matrices based on an inductive argument that begins with the @xmath486 case ; i.e. , with the family @xmath487 .
the members of this family will be exhibited in a moment .
first , however , for odd @xmath488 , define , inductively , the family @xmath489 of @xmath350 matrices of size @xmath373 by @xmath490 where @xmath491_1&\bold0_{\frac{d}{2}-\frac{1}{2 } } \\ \bold0_{\frac{d}{2}-\frac{3}{2}}&\bold0_{\frac{d}{2}-\frac{3}{2}}\end{array}\right . , v_{\frac{d}{2}-\frac{3}{2}}(d)=\left[\begin{array}{cc}0&1\\-\frac{2}{d}&0\\\bold0_{\frac{d}{2}-\frac{1}{2}}&\bold0_{\frac{d}{2}-\frac{1}{2}}\\\sqrt{1-(\frac{2}{d})^2}\left[i_{\frac{d}{2}-\frac{3}{2}}\right]_{\frac{d}{2}-\frac{3}{2}}&\bold0_{\frac{d}{2}-\frac{3}{2}}\end{array}\right.,\ ] ] and the remaining @xmath492 matrices in @xmath493 and @xmath494 matrices in @xmath495 are constructed from corresponding matrices in the family @xmath496 , where @xmath497_1&\bold0_{d-2}\end{array}\right .
, v_k(d)=\left[\begin{array}{cc}0&1\\-\frac{2}{d}&0\\\sqrt{1-(\frac{2}{d})^2}{{r}}[v_k(d-2)]_1&\bold0_{d-2}\end{array}\right . .\ ] ] for odd @xmath27 , the family @xmath489 will always contain two `` @xmath185 '' matrices , @xmath498 and @xmath499 , where @xmath500_1&\bold0_{d-2}\end{array}\right.\ .\ ] ] our inductive proof begins with @xmath501 , consisting of @xmath502 @xmath503&\begin{array}{c}0\\0\end{array}\\0&0\end{array}\right . , u_2(5)=\left[\begin{array}{cc}-\frac{2}{5}&0\\0&1\\\frac{\sqrt{21}}{5}&0\\0&0\\0&0\end{array}\right .
, v_1(5)=\left[\begin{array}{cc}0&1\\-\frac{2}{5}&0\\0&0\\0&0\\\frac{\sqrt{21}}{5}&0\end{array}\right . , \ ] ] @xmath504&\begin{array}{c}0\\0\\0\end{array}\end{array}\right .
, m(5)^*=\left[\begin{array}{cc}-\frac{1}{5}&-\frac{\sqrt{3}}{2}i\\\frac{\sqrt{3}}{5}i&\frac{1}{2}\\\frac{\sqrt{21}}{5}\left[\begin{array}{c}-\frac{1}{3}\\-\frac{\sqrt{5}}{3}\\\frac{\sqrt{3}}{3}i\end{array}\right]&\begin{array}{c}0\\0\\0\end{array}\end{array}\right . .\ ] ] the reader may verify that this family is @xmath505-orthogonal and consists of unitary matrices .
suppose that for some odd @xmath506 the family @xmath507 is @xmath426-orthogonal and consists of unitaries .
we claim that the family @xmath508 is a @xmath454-orthogonal family of unitaries , and establishing this claim completes the proof . the first two columns of @xmath509 and @xmath510 , defined by ( [ ukvk ] ) , are easily seen to be orthonormal , and thus can be augmented to create unitary matrices . for @xmath457 and @xmath464 ,
defined by ( [ ulvl ] ) , and @xmath511 and @xmath512 , defined by ( [ m ] ) , orthogonality of the first two columns can be easily verified , and normalization follows from the fact that @xmath513_1 , { { r}}[v_k(d)]_1 $ ] , and @xmath514_1 $ ] are unit vectors .
now we must verify @xmath454-orthogonality among the members of @xmath515 .
just as in the case for even @xmath27 , the matrices @xmath462 and @xmath463 are @xmath454-orthogonal , and once again , due to the placement of zeros , each @xmath457 matrix will be @xmath454-orthogonal to each @xmath464 matrix , for @xmath516 .
it is easy to check that @xmath517 as well as each @xmath457 and @xmath464 matrix is @xmath454-orthogonal to both @xmath462 and @xmath463 .
our next step is the verification that @xmath518 is a @xmath454-orthogonal family .
let @xmath85 and @xmath61 be distinct elements in @xmath519 ; we have @xmath520_1)^\dag{{r}}[u_j(d)]_1\right)+\lambda_1\\ & = & \frac{d+2}{d+4}\left(\frac{4}{(d+2)^2}+\frac{d(d+4)}{(d+2)^2}[u_i(d)]_1^\dag[u_j(d)]_1\right)+\frac{2}{d+4}.\label{uij}\end{aligned}\ ] ] since @xmath471 and @xmath472 are @xmath473 orthogonal at @xmath474 , @xmath475_1^\dagger[u_j(d)]_1\right)+\frac{2}{d+2}=0,\ ] ] and it follows that @xmath521_1^\dagger[u_j(d)]_1=-\frac{2}{d}$ ] . substituting @xmath522 for @xmath521_1^\dagger[u_j(d)]_1 $ ] into ( [ uij ] ) and simplifying the resulting equation will show that @xmath523 and so for @xmath524 , each @xmath457 matrix is @xmath454-orthogonal to every other @xmath457 .
a similar approach may be used to prove @xmath479 and @xmath458 are @xmath454-orthogonal for two distinct elements @xmath525 . as for @xmath454-orthogonality of @xmath509 and @xmath457 , for @xmath524 : @xmath526 which simplifies to 0 .
a similar calculation shows that @xmath527 for each @xmath528 .
thus the set @xmath529 is @xmath454-orthogonal , and all that remains to be shown is that this set , augmented with @xmath530 , remains @xmath454-orthogonal .
we begin by showing that @xmath512 and @xmath511 are @xmath454-orthogonal .
we have @xmath531_1^\dag[m(d)]_1)\right)+\frac{2}{d+4}\left(-\frac{1}{2}\right)\label{m1m2}. \end{split}\ ] ] since @xmath498 and @xmath499 are @xmath426-orthogonal when @xmath532 @xmath533_1^\dag[m(d)]_1)+\frac{2}{d+2}\left(-\frac{1}{2}\right)=0\ ] ] so @xmath534_1^\dag[m_1(d)]_1=\frac{1}{d}$ ] . if we substitute @xmath535 for @xmath534_1^\dag[m(d)]_1 $ ] in ( [ m1m2 ] ) and simplify we see @xmath536 , as desired .
we now establish the @xmath454-orthogonality of the @xmath457 matrices with @xmath511 and @xmath512 .
let @xmath524 ; we have @xmath537_1^\dag[m(d)]_1\right)+\frac{2}{d+4}\left(\frac{1}{2}\right ) . \end{split}\ ] ] because @xmath472 and @xmath498 are @xmath426-orthogonal at @xmath538 , @xmath539_1^\dag[m(d)]_1)+\frac{2}{d+2}\left(\frac{1}{2}\right)=0.\ ] ] hence @xmath540_1^\dag[m(d)]_1=-\frac{1}{d}$ ] , and by making this substitution in ( [ um ] ) and simplifying , @xmath457 is seen to be @xmath454-orthogonal to @xmath511 . also , @xmath541 which simplifies to 0
; so we have shown that every @xmath457 matrix is @xmath454-orthogonal to @xmath511 .
since @xmath457 and @xmath454 are real , @xmath542 immediately implies @xmath543 ; thus , every @xmath457 matrix is also orthogonal to @xmath512 .
we complete the argument by establishing the @xmath454-orthogonality of the @xmath458 matrices to @xmath511 and @xmath512 .
let @xmath528 ; we have @xmath544_1^\dag[m(d)]_1\right)+\frac{2}{d+4}\left(\frac{\sqrt{3}}{2}i\right ) .
\end{split}\ ] ] we are assuming that @xmath545 and @xmath498 are @xmath426-orthogonal at @xmath546 , so @xmath547_1^\dag[m(d)]_1)+\frac{2}{d+2}\left(\frac{\sqrt{3}}{2}i\right)=0.\ ] ] then @xmath548_1^\dag[m(d)]_1=-\frac{\sqrt{3}i}{d}$ ] , and by making this substitution in ( [ vm ] ) and simplifying , @xmath458 is seen to be @xmath454-orthogonal to @xmath511 . also , @xmath549 which simplifies to 0 , so we have shown that each @xmath458 matrix is @xmath454-orthogonal to @xmath511 . because @xmath454 and @xmath458 are real , it follows that every @xmath458 matrix is also @xmath454-orthogonal to @xmath512 .
we have shown that @xmath511 and @xmath512 are orthogonal to each other and to each matrix in the set @xmath550 .
therefore , the family @xmath508 is composed of d+4 @xmath454-orthogonal matrices .
this completes the argument that wcsg bound @xmath425 is saturated for every odd @xmath27 .
99 c. h. bennett and s. j. wiesner , phys .
lett . * 69 * , 2881 ( 1992 ) .
s. mozes , j. oppenheim , and b. resnik , phys .
a * 71 * , 012311 ( 2005 ) .
z. ji , y. feng , r. duan , and m. ying , phys .
rev . a * 73 * , 034307 ( 2006 ) . s. wu , s. cohen , y. sun , and r. griffiths , phys .
rev . a * 73 * , 042311 ( 2006 ) . s. cohen , duquesne university ( private communication ) . | we present an analytical study of the standard two - party deterministic dense - coding protocol , under which communication of perfectly distinguishable messages takes place via a qudit from a pair of non - maximally entangled qudits in pure state @xmath0 .
our results include the following : ( i ) we prove that it is possible for a state @xmath1 with lower entanglement entropy to support the sending of a greater number of perfectly distinguishable messages than one with higher entanglement entropy , confirming a result suggested via numerical analysis in mozes _
et al .
_ [ phys . rev .
a * 71 * , 012311 ( 2005 ) ] .
( ii ) by explicit construction of families of local unitary operators , we verify , for dimensions @xmath2 and @xmath3 , a conjecture of mozes _ et al . _ about the minimum entanglement entropy that supports the sending of @xmath4 messages , @xmath5 ; moreover , we show that the @xmath6 and @xmath7 cases of the conjecture are valid in all dimensions .
( iii ) given that @xmath1 allows the sending of @xmath8 messages and has @xmath9 as its largest schmidt coefficient , we show that the inequality @xmath10 , established by wu _
et al . _
[ phys .
rev .
a * 73 * , 042311 ( 2006 ) ] , must actually take the form @xmath11 if @xmath12 , while our constructions of local unitaries show that equality can be realized if @xmath13 or @xmath14 . |
whereas it is known from cosmological observations that cold dark matter represents a fraction @xmath2 of the total present energy density of our universe @xcite , its nature is still elusive .
dark matter can not only be detected directly in dedicated experiments searching for nuclear recoils from the scattering of dark matter particles , or produced in particle accelerators such as the lhc , but can also reveal its existence indirectly @xcite : although , apart from dilution from cosmic expansion , the density of dark matter does not change significantly after self - annihilations freeze out in the early universe , residual self - annihilation can give rise to significant fluxes of @xmath3rays , electrons , positrons , neutrinos , and even some antimatter such as anti - protons and positrons , especially in regions with large dark matter densities .
the energies of the secondary particles can reach up to the dark matter particle mass which can be of order a few hundred gev .
secondary electrons and positrons can annihilate and give rise to a 511 kev line emission , and they emit synchrotron radiation in the magnetic fields of galaxies which can be detected in the radio band .
therefore , cosmic and @xmath3ray detectors , neutrino telescopes , and even radio telescopes can be used for indirect dark matter detection as well .
one of the most promising dark matter candidates are weakly interacting massive particles ( wimp ) such as they are predicted within supersymmetric extensions of the standard model . such particles have masses @xmath4gev . if they are produced thermally , in order to reproduce the correct average dark matter density , their annihilation cross sections have to be @xmath5s whereas large values are possible in case of non - thermal production .
traditionally , indirect dark matter detection has focused on high energy emission , specifically signatures in @xmath3rays @xcite .
high energy emission of neutrinos , the particle the most difficult to detect , can be used to establish conservative constraints on the total annihilation cross section , which are of the order @xmath6s @xcite .
the fluxes of antiprotons and positrons from galactic dark matter annihilations has also been extensively used to constrain dark matter properties @xcite .
furthermore , the synchrotron radiation emitted by dark matter annihilation products close to the galactic centre has also been studied .
it has been found that if the dark matter profile close to the central black hole is a spike formed by adiabatic accretion , typical dark matter annihilation cross sections within supersymmetric scenarios can lead to intensities comparable to the radio emission observed from the galactic centre @xcite .
it has further been shown that measurements of the radio flux away from the galactic centre by the wmap experiment strongly constrain the annihilation cross sections to values @xmath7s for @xmath8gev @xcite .
a multi - wavelength analysis of dark matter annihilations from the galactic centre has recently been performed in ref .
@xcite . in the present paper
we evaluate the diffuse synchrotron emission from the electrons and positrons produced by dark matter annihilation in the cosmological distribution of dark matter halos .
we compute both its overall intensity and its angular power spectrum as well as the distribution of visible dark matter annihilation sources as a function of apparent luminosity .
we will find that comparing the resulting signals with other backgrounds and foregrounds under conservative assumptions allows to test annihilation cross sections close to the natural scale @xmath5s . in sect .
2 we provide the general setup of our calculations . in sect .
3 and 4 , we apply it to astrophysical backgrounds and the dark matter induced signal , respectively . in sect .
5 we compare the overall diffuse signal and its anisotropy to other foregrounds and discuss the resulting dark matter constraints , and in sect .
6 we conclude .
finally , an appendix presents technical details of the calculations .
we will use natural units in which @xmath9 throughout .
we consider a distribution of sources which emit a radio luminosity per frequency interval @xmath10 which depends on a parameter @xmath11 , on frequency @xmath12 and on redshift @xmath13 .
the energy flux per frequency interval and solid angle is then given by @xmath14}{4\pi d_l(z)^2}\,,\ ] ] where for abbreviation we write @xmath15 , @xmath16 is the co - moving volume density of objects per unit interval in the parameter @xmath11 , @xmath17 is the luminosity distance , the factor @xmath18 comes from redshifting the frequency interval @xmath19 , and the co - moving volume per solid angle and redshift interval is @xmath20 here , for a flat cosmological geometry , the hubble rate is @xmath21^{1/2}\,,\ ] ] @xmath22 is the co - moving distance and @xmath23 $ ] is cosmic time as function of redshift .
throughout this paper we will assume a flat , @xmath24cdm universe with the total non - relativistic matter density @xmath25 and the dark energy density @xmath26 ( all other contributions to the energy density are negligible ) in units of the critical energy density @xmath27 , where @xmath28 is newtons constant and @xmath29 with @xmath30 @xcite . in order to calculate the anisotropies we introduce an emissivity of squared power per frequency interval , @xmath31 , @xmath32 which , similar to the approach in ref .
@xcite , we split into the two parts @xmath33 and @xmath34 and which also depends on the co - moving wavenumber @xmath35 .
the first part is essentially poisson noise and corresponds to the sum over squared luminosities , @xmath36 ^ 2\,,\ ] ] where in the following we define @xmath37 as the spatial fourier transform of any function @xmath38 and where @xmath39 is the fourier transform of the spatial emission density @xmath40 of an individual source , normalized to unity , @xmath41 .
the second contribution to eq .
( [ eq : l2 ] ) is determined by the correlation between sources , @xmath42 ^ 2\,,\ ] ] where @xmath43 is the linear power spectrum of the density fluctuations @xmath44 and we have also introduced a bias factor @xmath45 of the sources with respect to the density field . the angular power spectrum @xmath46 is given by @xmath47 where @xmath48y^*_{lm}(\omega)\ ] ] in terms of the spherical harmonic functions @xmath49 and the intensity @xmath50 measured along direction @xmath51 .
for a statistically isotropic sky this results in @xmath52 ^ 2}\,.\ ] ] using @xmath53 and inserting eq .
( [ eq : dv ] ) in eqs .
( [ eq : i0 ] ) and ( [ eq : cl2 ] ) finally gives @xmath54 and @xmath55 formally , for point - like sources , the integral over redshift in eq .
( [ eq : cl22 ] ) is divergent at @xmath56 . in practice
this is regularized by the fact that the nearest source has some minimal distance and that one can subtract the most luminous point sources which are also the nearest sources .
in addition , the integral is regularized by the spatial extent of the sources , represented by the factor @xmath57 in eqs .
( [ eq : l21 ] ) and ( [ eq : l22 ] ) .
the role of these effects in practical calculations will be discussed in sect .
for astrophysical sources , @xmath11 can be identified with the radio luminosity @xmath58 at some fixed frequency @xmath59 .
( [ eq : i02 ] ) then simplifies to @xmath60 where @xmath61 is the intrinsic luminosity corresponding to the apparent point source flux @xmath62 above which we consider the source to be resolvable and thus subtractable from the diffuse background . for the multipoles eqs .
( [ eq : l21 ] ) and ( [ eq : l22 ] ) can then be written as @xmath63 ^ 2\ ] ] and @xmath64 ^ 2\,,\ ] ] respectively .
for the luminosity functions @xmath65 of normal and radio galaxies we will use the expressions given in ref .
for annihilation of dark matter with mass @xmath66 and phase space averaged annihilation cross section times velocity @xmath67 , @xmath11 can be identified with the mass @xmath68 of dark matter halos .
we then follow the approach of ref .
@xcite and write @xmath69 where we define @xmath70 as a quantity which does not depend on annihilation cross section or mass of the dark matter particles , @xmath71\,.\ ] ] in eq .
( [ eq : e_ann ] ) , @xmath72 and @xmath73 are the electron charge and mass , respectively , @xmath74 is the dark matter halo density profile , @xmath75 is the local magnetic field strength , , @xmath76 is the multiplicity per annihilation of electrons and positrons with energies larger than @xmath77 , @xmath78 is the total synchrotron emission power of one electron of energy @xmath77 in a magnetic field of strength @xmath79 , corresponding to an energy density @xmath80 , and @xmath81 is the energy loss rate at energy @xmath77 due to inverse compton scattering on a low energy photon field of energy density @xmath82 .
furthermore , we use the function @xmath83 in eq .
( [ eq : e_ann ] ) , with the critical frequeny @xmath84 in the following we use the approximation @xcite @xmath85\ ] ] such that eq .
( [ eq : e_ann ] ) can be simplified to @xmath86}{\nu^{1/2}}\,i(m)\ ] ] where @xmath87 and the critical energy @xmath88 is the inversion of eq .
( [ eq : nu_crit ] ) , @xmath89 in eq .
( [ eq : e_ann2 ] ) we neglect the magnetic field dependence of @xmath90 $ ] . for @xmath4gev , @xmath91 a few micro - gauss and @xmath92ghz , the parameters we are interested in , this is a good approximation because the critical energy @xmath93
. typical values for these parameters are @xmath94 @xcite .
this corresponds to a fraction @xmath95 of the total annihilation energy going into pairs .
the energy fraction going into pairs of energy above @xmath77 can be expressed in terms of @xmath76 as @xmath96 with the above expressions we can rewrite eq .
( [ eq : i02 ] ) as @xmath97i(m)\,.\ ] ] furthermore , we can redefine @xmath98 and @xmath99 from eqs .
( [ eq : l21 ] ) and ( [ eq : l22 ] ) by extracting constant factors and write @xmath100i(m)|u(k , m)|\right)^2\ ] ] and @xmath101i(m)b(m , z)|u(k , m)|\right)^2\,,\ ] ] where @xmath102 relates to the halo profile , @xmath103 . with these quantities
we can now write @xmath104 details about the quantities that enter these expressions are given in appendix a. eq .
( [ eq : cl_dm ] ) can also be obtained as follows : limber s equation relates the two - dimensional angular power spectrum @xmath105 to the three - dimensional power spectrum @xmath106 in the flat sky approximation @xcite : given a three - dimensional statistically random field @xmath107 , one considers the observation at @xmath108 of the projection @xmath109 with some given radial weight function @xmath110 , where @xmath111 is the co - moving distance .
if the field @xmath112 fluctuates on scales much smaller than the characteristic scale over which @xmath110 varies , then we have @xmath113 where @xmath114 is the power spectrum of @xmath115 at the co - moving wavenumber @xmath116 . neglecting the variation of the magnetic field @xmath79 within the halo regions contributing most to the annihilations , the radio intensity eq .
( [ eq : i0_dm ] ) along a given direction @xmath51 can be written as @xmath117 \frac{\left[1+\delta(z,\omega)\right]^2}{1+u_\gamma / u_b}\,,\ ] ] where @xmath118 is the average dark matter density at zero redshift , and @xmath119 is the relative overdensity . because the dominant contribution comes from the dark matter halos , where @xmath120 , we can approximate @xmath121 . assuming a constant @xmath79 and a constant optical photon field of density @xmath122 , we can write the factor @xmath123^{-1}$ ] , where @xmath124 is the cmb energy density at @xmath125 .
this factor effectively cuts off the redshift integration at @xmath126 .
since @xmath127 varies little over this redshift range , we can then further simplify eq .
( [ eq : i - syn - dm1 ] ) to @xmath128\left\langle\sigma v\right\rangle}{m_x^2 } \frac{9\rho^2_m}{64\pi\sqrt{eb\nu } } \left(\frac{m_e^3}{0.29\pi}\right)^{1/2}\int dz \frac{(1+z)^{3/2}\delta^2(z,\omega)}{h(z)\left[1+\frac{u_{\rm op}}{u_b}+\frac{u_0}{u_b}(1+z)^4\right]}\,.\ ] ] comparing this with eq .
( [ eq : limber1 ] ) , we can use @xmath129 for the random field and the weight function is @xmath130\left\langle\sigma v\right\rangle}{m_x^2 } \frac{9\rho^2_m}{64\pi\sqrt{eb\nu } } \left(\frac{m_e^3}{0.29\pi}\right)^{1/2}\frac{(1+z)^{3/2}}{1+\frac{u_{\rm op}}{u_b}+\frac{u_0}{u_b}(1+z)^4}\,.\ ] ] the power spectrum @xmath131 appearing in eq .
( [ eq : limber2 ] ) is then the fourier transform of the two - point correlation function of @xmath112 in real space .
following ref .
@xcite , @xmath132 can be written as the sum of a one - halo and a two - halo term , @xmath133 , with @xmath134 ^ 2= \int_{m_{min}}^{m_{\rm cut}(z ) } dm \frac{dn}{dm}\left(\frac{a_b{\cal f}_{\rho^2_h}(k , m , z ) } { \rho_m^2(1+z)^6}\right)^2\ ] ] @xmath135 ^ 2\nonumber\\ \fl & = & p_{\rm lin}(k)\left[\int_{m_{min}}^{m_{\rm cut}(z ) } dm \frac{dn}{dm } b(m)\left(\frac{a_b{\cal f}_{\rho^2_h}(k , m , z)}{\rho_m^2(1+z)^6}\right ) \right]^2\,,\end{aligned}\ ] ] where @xmath136 is the minimal halo mass and @xmath137 is the halo mass corresponding to the apparent point source flux @xmath62 above which we consider the source to be resolvable and thus subtractable from the diffuse background .
furthermore , @xmath138 is a boost factor which accounts for possible substructure in the halos . the average of the clumping factor appearing in eq .
( [ eq : i - syn - dm2 ] ) is given by @xmath139 where @xmath140 is the halo volume element .
a generic form for the halo mass function @xmath141 appearing in the equations above was first proposed by press & schechter ( ps ) @xcite ; a modified version of this form is given by sheth and tormen @xcite ( st ) . when comparing the results obtained from these two forms , we find differences by factors less than 2 .
thus , we adopt the ps formula throughout our paper .
current knowledge of the dark matter density distribution mostly comes from n - body simulations , and the universal dark matter profile firstly proposed is the navarro - frenk - white ( nfw ) model @xcite .
combining eq .
( [ eq : rho_nfw ] ) with eq .
( [ eq : delta_ch ] ) in appendix a , in this model the dark matter profile within each halo can be written as @xmath142 where @xmath143 is a characteristic radius , and the concentration parameter @xmath144 is defined as the ratio of @xmath143 and the virial radius @xmath145 , @xmath146 , see also appendix a for more details .
note that @xmath143 is not a free parameter , but depends on @xmath68 and @xmath144 because @xmath145 is related to @xmath68 via @xmath147 , where @xmath148 in an einstein - de sitter universe is about @xmath149 . with the above definition , @xmath144 and @xmath68
completely determine the dark matter distribution of a given halo .
the minimal halo mass is still rather uncertain .
the value @xmath150
@xcite is close to the free - streaming mass @xcite , below which there are no fluctuations in the dark matter density to form a halo .
note that the magnetic field may be much smaller than micro gauss scales in such small halos .
in contrast , the value @xmath151 roughly corresponds to the minimal mass of dwarf galaxies which are known to contain micro gauss scale magnetic fields @xcite .
we , therefore , choose @xmath151 as fiducial value in the following , noting that the dark matter signal would increase by only a factor about two for @xmath150 .
we will furthermore use @xmath152 g as fiducial value for the magnetic field .
this is a realistic value given that most annihilations occur in the densest regions where also magnetic fields are somewhat larger than typical average galactic fields .
the clumping factor is very sensitive to the concentration parameter , namely @xmath153 .
n - body simulations indicate that the concentration has a log - normal distribution @xcite with a median value of @xmath154 where the collapse redshift @xmath155 is implicitly given by the relation @xmath156 , where @xmath157 is the mass scale at which @xmath158 . how the concentration parameter depends on halo mass and redshift
is still an open question .
one can extrapolate eq .
( [ eq : con ] ) to minimal halo masses @xmath159 . when comparing the parameterization eq .
( [ eq : con ] ) with high resolution simulations @xcite we find that it gives realistic values for the minimum halo mass .
it is a conservative estimate because at @xmath160 it gives values @xmath161 for the minimum halo mass which is significantly smaller than other parameterizations @xcite .
recent studies show that dark matter halos exhibit considerable substructure @xcite .
the total mass of these substructures only account for about @xmath162 of the host halo , but they can give an extra boost factor @xmath163 for dark matter annihilation .
some studies show that if one takes into account substructure and assumes a cuspy center slope @xcite , the theoretical prediction can well explain the excess of high energy positrons and the diffuse @xmath3 ray background observed by the heat @xcite and egret @xcite experiments , respectively .
the subhalos follow a certain mass and redshift distribution which is still unknown .
therefore , to be conservative we assume the nfw halo model and simply parametrize any possible boost factor with the parameter @xmath163 .
the substructures occur on small scales and do not influence the power spectrum in the range we are interested , @xmath164 .
the cosmic microwave background ( cmb ) dominates the radio sky at frequencies above @xmath165ghz , whereas astrophysical sources such as normal galaxies and radio galaxies dominate at lower frequencies down to khz frequencies @xcite .
recently it was argued that synchrotron emission of strong intergalactic shocks can also significantly contribute to the diffuse extragalactic radio below 500 mhz @xcite . using the formuli developed in sect . 4
, we now evaluate the contribution of synchrotron emission from pairs produced by dark matter annihilation in the magnetic fields of dark matter halos .
we consider neutralinos as dark matter candidate , and for the following figures we assume a neutralino mass of @xmath166 gev and a total annihilation cross section of @xmath167/s , fixed for reproducing the correct relic density for thermal relics .
we also assume that the average total number of electrons and positrons per annihilation is @xmath94 , and that the halo substructure implies a boost factor @xmath168 .
we compare the resulting dark matter signal with astrophysical contributions to the diffuse background that can be computed from the expressions in sect .
3 . for astrophysical sources
the diffuse radio background is likely dominated by normal galaxies and radio galaxies . to estimate the contributions from these sources , we follow ref .
@xcite , which use the observed correlation between the radio and infra - red flux of galaxies .
this approach assumes that the radio emission is related to the star formation and is sensitive to the redshift evolution of the sources , but can explain the observed radio background quite well .
gev , @xmath5s , @xmath169 , @xmath170 g , @xmath171 ( brown curves ) . here , the solid brown curve is for @xmath172 , while the dashed brown curve is for @xmath173 .
also shown is the cmb background ( cyan solid curve ) as well as its subtractable part , determined by uncertainties of the absolute cmb temperature ( dotted cyan curve ) .
the galactic foreground at galactic latitude @xmath174 is shown as the green band within uncertainties.,scaledwidth=60.0% ] following the above assumptions , in fig .
[ fig : radio_bg ] , we show the different contributions to the average diffuse radio intensity . for astrophysics sources , normal galaxies
contribute more than radio galaxies .
this is because although the individual radio galaxy is brighter than a normal galaxy on average , this is overcompensated by the larger number of normal galaxies .
also shown in fig .
[ fig : radio_bg ] is a possible contribution from intergalactic shocks @xcite normalized such that its angular power spectrum is comparable to the one of the galactic foreground , see sect.5.2 .
of course , the cmb absolutely dominates the radio sky in the wide range from @xmath175ghz to a few hundred ghz @xcite , and above these frequencies galactic foregrounds such as dust emission dominates .
since the cmb is a black body radiator its contribution to the solid angle averaged radio flux can be subtracted up to the uncertainty of its average absolute temperature . currently the cmb temperature is measured to @xmath176 k @xcite .
we convert this temperature uncertainty into an intensity of cmb confusion noise .
[ fig : radio_bg ] shows that this confusion noise dominates other astrophysical backgrounds and the diffuse signal of our fiducial dark matter scenario at @xmath177ghz . at lower frequencies the dark matter signal @xmath178 tends to decrease as @xmath179 for @xmath180const .
, see eq .
( [ eq : i - syn - dm2 ] ) , whereas the background from normal galaxies tends to be flat , see fig .
[ fig : radio_bg ] .
there is thus an optimal window at frequencies @xmath92ghz where dark matter annihilation signatures can be detected and where self - absorption is negligible .
constraints on dark matter parameters can , therefore , only be established for annihilation cross sections about a factor ten higher than the fiducial cross section required for thermal dark matter . in addition , there are three diffuse foregrounds from our galaxy in the frequency range we are interested : the first is synchrotron radiation emitted by high energy electrons gyrating in the galactic magnetic field , the second is free - free emission from the thermal bremsstrahlung from hot ( @xmath181k ) electrons produced in the interstellar gas by the galactic uv radiation field , and the third foreground is dust emission which arises from the thermal re - radiation of absorbed stellar light .
[ fig : radio_bg ] shows that these foregrounds tend to dominate the astrophysical backgrounds and the dark matter signal in the fiducial scenario .
can we test the properties of dark matter more powerfully ?
the absolute cmb temperature is difficult to measure more precisely than to the current permille level , because of inevitable systematic errors .
small - scale temperature fluctuations @xmath182 have been seen by the cobe and wmap satellites because temperature differences can be measured more precisely since systematic errors cancel in measurements of temperature differences . furthermore ,
if the galactic foregrounds have a smooth directional dependence , they may pose less of a contamination when considering the anisotropy of the radio sky .
we , therefore , consider in the following the angular power spectra of the radio sky in order to see if it can provide further tests of dark matter properties . whereas the diffuse average radio flux provides only one number at a given frequency to compare with other astrophysical and cosmological backgrounds , potentially much more information
is contained in the angular power spectrum .
for example , the power spectrum as a function of angular scale tends to be different for dark matter annihilation and astrophysical sources because the contribution of the latter to the diffuse radio flux is dominated by fewer bright sources .
our goal in this section is whether this can provide dark matter signatures or constraints on mass and annihilation cross section . before calculating the angular power spectra
, we discuss their qualitative behaviors .
the angular power spectrum @xmath183 can be divided into one - halo ( @xmath184 ) and two - halo ( @xmath185 ) terms , corresponding to the two contributions eqs .
( [ eq : l21 ] ) and ( [ eq : l22 ] ) to eq .
( [ eq : l2 ] ) , and thus to eq .
( [ eq : cl22 ] ) .
the two - halo term arises from the correlation between distinct halos which is described by the linear power spectrum .
the one - halo term represents correlation within the same halo .
both one - halo and two - halo term are proportional to @xmath57 , the square of the fourier transform of the spatial emission profile . at large angular scales , @xmath186 , such that @xmath184 is essentially independent of @xmath187 .
the one - halo term is thus sometimes called poisson noise . at scales comparable to the size of the source , @xmath57 starts to become suppressed .
therefore , both one - halo and two - halo terms are expected to be suppressed for multipoles @xmath187 larger than the typical distance to the source divided by the linear source size .
the two - halo term is furthermore proportional to the linear power spectrum which is also suppressed for co - moving wavenumbers @xmath188 .
therefore , the ratio of the two - halo term to the one - halo term is suppressed for @xmath189 , where @xmath190mpc is the hubble scale .
the one - halo term eventually dominates at very small angular scales . in eq .
( [ eq : cl22 ] ) , for point - like sources , formally the one - halo term @xmath184 would diverge for @xmath191 , whereas the two - halo term eq .
( [ eq : l22 ] ) is regularized by the linear power spectrum @xmath192 , which is suppressed at large @xmath193 .
this is because the flux of nearby sources of a given luminosity diverges .
we can ignore such sources because they can be identified as individual bright sources and be removed from the background flux in actual observations .
we can remove sources with intrinsic luminosity @xmath194 , corresponding to the point - source sensitivity @xmath62 of the telescope .
alternatively , one can regularize eq .
( [ eq : cl22 ] ) by integrating from some finite minimum distance corresponding to the typical distance to the nearest source , @xmath195mpc .
furthermore , eq . ( [ eq : cl22 ] ) is also formally regularized at @xmath191 by the spatial extent of nearby sources , described by @xmath57 . for the nfw profile
, the mass of the halo within distance @xmath111 from the halo centre increases as @xmath196 up to @xmath197 , and then increases logarithmically between @xmath143 and @xmath145 since @xmath198 , see eq .
( [ eq : nfw ] ) .
therefore , the dominant contribution to the halo mass comes from @xmath199 .
similarly , for @xmath199 the annihilation signal increases as @xmath111 , but between @xmath143 and @xmath145 increases only as @xmath200 . assuming the emission traces @xmath201 for astrophysical emission processes and @xmath202 for dark matter annihilation , the fourier transforms of these dependencies then give @xmath203 for @xmath204 , with @xmath205 for astrophysical emission and @xmath206 for dark matter , see appendix a.6 for more details . since @xmath193 , and thus @xmath207|^2\propto r(z)^{2\gamma}$ ] , the one - halo term in eq .
( [ eq : cl22 ] ) diverges only for @xmath208 .
therefore , under our assumptions for the emission profile , eq .
( [ eq : cl22 ] ) is convergent even without cut - offs in either @xmath209 or the apparent luminosity . since nevertheless in particular the one - halo term is quite sensitive to nearby sources , in the following we study its dependence on @xmath62 and @xmath209 .
smaller than @xmath62 to the two - halo term at @xmath210ghz .
the red , blue and brown lines represent the contribution from radio galaxies , normal galaxies , and dark matter ( fiducial scenario with @xmath172 ) , respectively.,scaledwidth=60.0% ] fig .
[ fig : sc_2 g ] shows the dependence of the two - halo term on @xmath62 . according to eq .
( [ eq : l22 ] ) , the two - halo term scales with the square of the average flux .
since the apparent luminosity of radio galaxies can be of the order of a jansky ( @xmath211 ) , the two - halo term from radio galaxies starts to decline when we cut sources being less luminous than a critical luminosity below a jansky .
in contrast , the contribution of normal galaxies which are much less luminous than radio galaxies starts to decline only when we cut sources more luminous than @xmath212jy .
the contributions of dark matter halos to the dark matter annihilation signal is basically unaffected by any source removal , even to luminosities down to @xmath213jy .
this is easy to explain : in our fiducial scenario the largest dark halos of about @xmath214 produce only about @xmath215s at 2 ghz from dark matter annihilation , far less than the typical radio luminosity of galaxies of about @xmath216/s . as a result , removing bright sources increases the contribution of dark matter annihilation to the two - halo term relative to the contribution from astrophysical sources .
smaller than @xmath62 to the one - halo ( poisson ) term at 2 ghz .
the solid and dotted curve represent the cases of @xmath217 and @xmath218 mpc , respectively .
color keys are as in fig .
[ fig : sc_2g].,scaledwidth=60.0% ] next we discuss the one - halo term .
the one - halo term is more sensitive to the cut - offs in apparent luminosity @xmath62 and to the minimal distance @xmath209 than the two - halo term because of two reasons : first , the two - halo term eq .
( [ eq : l22 ] ) is the square of an integral of luminosities , whereas the one - halo term eq .
( [ eq : l21 ] ) is essentially poisson noise and thus proportional to an integral of squared luminosities , which makes the contribution from bright sources more important .
second , the two - halo term is further regularized by the linear power spectrum at large @xmath193 . in fig .
[ fig : ssc_2 g ] we show the cumulative contribution of sources dimmer than @xmath62 to @xmath184 .
similarly to the two - halo term shown in fig .
[ fig : sc_2 g ] , the contribution of radio galaxies and ordinary galaxies decreases rapidly below @xmath165jy and @xmath219jy , respectively , whereas the contribution of dark matter annihilation is affected less by source removal .
nevertheless , the contribution of bright sources is now much larger than for the two - halo term , as expected , and the one - halo term continues to rise with inclusion of brighter sources . on the other hand
, practically one should cut off the integral at some minimal distance @xmath220mpc within which there are essentially no bright sources .
since sources at small distance appear bright , the cut - off in luminosity and minimal distance is of course to some extent degenerate , as confirmed by fig .
[ fig : ssc_2 g ] . for radio galaxies , removal above @xmath221jy is equivalent to restricting to distances larger than 1 mpc .
for ordinary galaxies , cutting at a minimal distance @xmath222mpc is equivalent to removing sources brighter than 0.1 jy .
since observational sensitivities are considerably better than these luminosities , cutting at @xmath220mpc does , therefore , not introduce any significant uncertainties .
note that in fig .
[ fig : sc_2 g ] dark matter dominates the two - halo terms if all sources above @xmath223jy are removed , while in fig .
[ fig : ssc_2 g ] it would dominate the one - halo terms only for unrealistically small cut - off luminosities @xmath224njy . this is because the one - halo term is much more sensitive to bright sources than the two - halo term and because the dark matter contribution consists of dimmer sources than ordinary astrophysical sources .
solid lines and dotted lines represent the one - halo and two - halo terms , respectively .
we assume the astrophysical sources to be point - like .
the minimal dark matter halo mass is @xmath171 .
sources at distances below @xmath222mpc , and of apparent luminosity above @xmath225mjy were removed .
color keys are as in fig .
[ fig : sc_2g].,scaledwidth=60.0% ] , but assuming the emission profile of the astrophysical sources follows an nfw profile .
sources with luminosities above @xmath225mjy are again subtracted.,scaledwidth=60.0% ] , but subtracting sources above @xmath226jy.,scaledwidth=60.0% ] the angular power spectra of the radio background at 2 ghz produced by galaxies and by our fiducial dark matter scenario are shown in figs .
[ fig : r1-point - sim-01mj ] , [ fig : z0-nfw - sim-01mj ] , and [ fig : z0-point - sim-1uj ] for different source removal cuts . based on the above discussion
, the qualitative behavior of the one- and two - halo terms can be easily understood : in fig .
[ fig : r1-point - sim-01mj ] we assume galaxies to appear point - like and we remove sources brighter than @xmath227mjy .
the one - halo terms from these sources thus increase proportional to @xmath228^{1/2}$ ] in the above figures .
the same applies to the one - halo term of the dark matter contribution for @xmath164 , corresponding to angular scales @xmath229 . at smaller angular scales
the power spectrum is suppressed by the inner structure of the dark matter halos .
we can estimate this critical scale as follows : the one - halo term is dominated by the brightest halos which correspond to the largest and nearest halos . in our fiducial scenario ,
the annihilating dark matter in the largest halos can emit a radio flux of @xmath230 at 2 ghz , such that the minimum co - moving distance is @xmath231mpc , @xmath232 if sources brighter than @xmath227mjy are removed .
the scale @xmath143 for the corresponding @xmath233 halo is about 0.21 mpc .
this corresponds to a multipole @xmath234 .
this simple estimation is consistent with our detailed calculation shown in fig .
[ fig : r1-point - sim-01mj ] . figs .
[ fig : r1-point - sim-01mj ] , [ fig : z0-nfw - sim-01mj ] , and [ fig : z0-point - sim-1uj ] show that for radio galaxies the one - halo term is always larger than the two - halo term at all multipoles , as expected because of the high luminosity of radio galaxies . for dark matter and normal galaxies , the two - halo term dominates at small @xmath187 .
the dependence of the angular power spectrum on @xmath187 can potentially be used to discriminate the dark matter signal from astrophysical contributions : for @xmath235 , the annihilation power spectrum looks significantly flatter than the signal from normal galaxies . in other words , at large angular scales , the annihilation signal has relatively more power .
this can be understood as follows : after cutting bright sources , many more dim nearby annihilation sources than galaxies contribute .
in addition , at large redshift the synchrotron emission from dark matter annihilation is suppressed by the increased inverse compton scattering rate on the cmb , see eq .
( [ eq : i - syn - dm2-w ] ) .
the two - halo term is proportional to @xmath236 which peaks at @xmath237 , corresponding to @xmath238mpc .
the on average smaller distance to the dark matter halos then translates into relatively more power at small @xmath187 . in fig .
[ fig : z0-nfw - sim-01mj ] , we take into account the spatial extent of the radio emission of galaxies .
we assume the luminosity profiles of galaxies to be roughly proportional to the dark matter density profile which is obtained following appendix a.2 with the halo mass is obtained from the relation between mass and bolometric luminosity @xcite . as a result , for normal and radio galaxies the one - halo term starts to drop for @xmath239 and @xmath240 , respectively .
compared to the dark matter signal , the suppression thus sets in at slightly smaller @xmath187 for normal galaxies , but only at larger @xmath187 for radio galaxies . for normal galaxies
this is due to the more extended emission profile which more closely follows the density as opposed to the squared density in case of dark matter .
this is also reflected by the fourier transform of the emission profiles shown in fig .
[ fig : fourier_rho ] . for radio galaxies
this effect is overcompensated by the fact that they are much brighter such that after cutting bright nearby sources , their average distance is much larger where their angular extent appears smaller . since future radio detectors such as the square kilometer array ( ska ) @xcite
can reach point flux sensitivities of @xmath241jy , we show the power spectra of the background remaining after a corresponding luminosity cut in fig .
[ fig : z0-point - sim-1uj ] . since the two - halo term from dark matter annihilation is insensitive to such luminosity cuts whereas the contribution from galaxies decreases rapidly , as shown in fig .
[ fig : sc_2 g ] , the relative contribution of dark matter annihilation increases and gives rise to a flatter power spectrum at moderate @xmath187 .
( green shaded region ) and the cmb ( cyan curve ) .
the brown band represents the annihilation spectrum , where the upper and lower ends correspond to @xmath242 and @xmath243 , respectively , see eq .
( [ eq : f_dm ] ) , and from which halos brighter than @xmath227mjy were removed .
the black - dotted and black - solid curves represent the total signal from normal and radiogalaxies , for luminosity cuts @xmath244mjy and @xmath225mjy , respectively .
also shown is a possible contribution from intergalactic shocks @xcite , normalized such that its angular power spectrum is comparable to the galactic foreground.,scaledwidth=60.0% ] we now have to compare the cosmological background power spectra discussed so far with other potential contaminations .
[ fig : cl_sum ] compares the signals from ordinary and radio galaxies and from our fiducial dark matter scenario with the power spectra of the cmb and of the galactic foreground at high galactic latitude .
the power spectrum of the galactic foreground is not very well measured and we represent its uncertainties as a green band in fig .
[ fig : cl_sum ] . at high galactic latitude below 10 ghz local foreground fluctuations dominate over the cmb power spectrum which is why the cmb anisotropy measurements are performed above 20 ghz .
concerning annihilation signatures of dark matter with mass @xmath4gev in the angular power spectrum of the radio sky , the optimal frequency band is around 2 ghz . at higher frequencies , the synchrotron emission of electrons produced from dark matter annihilations cuts off due to eq .
( [ eq : e_crit ] ) and the cmb signal increases . at lower frequencies ,
synchrotron emission by galactic electrons dominates the power spectrum even at high galactic latitude @xcite .
around 2 ghz , galactic synchrotron emission always dominates , whereas free - free emission is a factor few smaller . also shown in fig .
[ fig : cl_sum ] is a possible signal from intergalactic shocks @xcite .
since its normalization is rather uncertain , we normalized it such that it is comparable to the average estimate of the galactic foreground .
the thermal sz effect @xcite is another characteristic contamination caused by hot ionized gas in galaxy clusters and filaments outside of clusters @xcite .
since it dominates at small angular scales , @xmath245 , and at high frequencies above 30 ghz , we can neglect this effect here . as can be seen from comparing fig .
[ fig : z0-nfw - sim-01mj ] and [ fig : z0-point - sim-1uj ] and from fig .
[ fig : cl_sum ] , future radio telescope arrays sensitive around @xmath246ghz , with their higher point flux sensitivities should allow to further reduce the contribution from galaxies , whereas for @xmath247 the dark matter contribution is hardly changed by removing still fainter sources .
this can be understood from the fact that the dark matter signal is dominated by the two - halo term which is insensitive to @xmath62 for @xmath248njy , see fig .
[ fig : sc_2 g ] .
this shows that for dark matter annihilation the distribution of @xmath249 is nearly flat for @xmath250 . at smaller @xmath187 the power spectrum
is dominated by galactic foregrounds and at larger @xmath187 the one - halo term from galaxies grows rapidly .
the most sensitive range @xmath251 should be accessible to present and future radio telescopes with their high angular resolution .
the ska will have a sensitivity of about @xmath252 in the units of the above figures .
we conclude that the power spectrum from dark matter annihilation tends to be flatter than other contributions because of an interplay of the following effects : * the astrophysical signals are dominated by fewer and much brighter sources than the dark matter annihilation signal which consists of many faint sources . for @xmath225mjy , the two - halo term dominates for @xmath253 for both the signals from galaxies and from dark matter annihilation .
in addition , the dark matter signal is significantly flatter in that angular range , i.e. it has relatively more power at small @xmath187 .
this is because the two - halo term is proportional to the linear power spectrum whose peak in wavenumber for the on average closer and dimmer dark matter annihilation sources translates into smaller @xmath187 at these luminosities . * for @xmath254 ,
the inner spatial structure of the galaxies and dark matter halos becomes important .
the inner structure tends to suppress the power spectra , but the exact angular scale at which these effects become important depends on the halo size , the profile of the emission and source luminosity cut - off . * the various components evolve differently with the universe expansion .
for example , at high redshift inverse compton scattering on the cmb tends to suppress synchrotron emission in dark matter halos , whereas astrophysical sources such as radio galaxies tend to be more active at @xmath255 .
we can now scale the dark matter signal to parameter values different from the fiducial scenario , by multiplying with the factor @xmath256 we caution that a boost factor as high as @xmath168 has not been verified in all dark matter structure simulations and that the average magnetic field @xmath79 could be significantly smaller than 10@xmath257 g if many small - scale subhalos contribute .
however , smaller values for @xmath138 and @xmath79 partially compensate in eq .
( [ eq : f_dm ] ) so that one could still obtain observable signatures as long as @xmath79 is large enough to produce emission at ghz frequencies , see eq .
( [ eq : e_crit ] ) . if we choose @xmath242 ( upper end of brown band in fig .
[ fig : cl_sum ] ) and @xmath225mjy , the annihilation spectrum dominates over other cosmological backgrounds for @xmath258 and should become distinguishable from the galactic foreground .
note that this foreground is likely further reduced close to the galactic poles . in this situation
it should thus be possible to disentangle the rather flat power spectrum of the dark matter annihilation signal from other contributions in the range of @xmath251 .
we can thus assert that radio observations are sensitive to @xmath259 with some dependence on the source luminosity cut - off @xmath62 .
note that the dark matter signal shown in fig .
[ fig : cl_sum ] does not strictly scale with @xmath260 because it depends on sources dimmer than @xmath62 , here chosen as 0.1 mjy .
however , since few brighter dark matter halos contribute , the signal scales with @xmath260 in first approximation . as function of apparent radio flux @xmath261 compared with predictions for normal galaxies ( blue curve ) , radio galaxies ( red curve ) , and annihilations from dark matter halos ( brown band , for @xmath262 ) .
green shaded region and triangles are data from ref .
@xcite.,scaledwidth=60.0% ] one can also compare observed radio source counts as a function of apparent point source flux with predictions for astrophysical sources and dark matter annihilation sources .
this is done in fig .
[ fig : counts ] for the same parameters as used in fig .
[ fig : cl_sum ] .
this establishes the constraint @xmath263 .
in contrast , fig . [
fig : cl_sum ] provides dark matter signatures for future measurements but currently does not allow to put a constraint on @xmath260 because of the uncertainties in the galactic foreground spectrum .
note that a future observational extension of the source count spectrum in fig .
[ fig : counts ] to apparent luminosities @xmath264jy will provide an additional test for dark matter which predicts a shallower source count distribution than astrophysical sources .
we have not computed the contribution from dark matter annihilations in our own galaxy to the anisotropic radio flux in the present work .
however , we know from refs .
@xcite that for our fiducial values for cross section and mass , at least the smooth halo component does not lead to fluxes higher than current observations from wmap .
the contribution from galactic substructures is probably more model dependent than our cosmological flux which apart from an overall boost factor depends only on the host halo distribution and effectively averages over a much larger ensemble of halos .
this can also be seen from ref .
@xcite where the predictions of the @xmath3ray flux from galactic dark matter annihilations varied over orders of magnitude .
many different indirect detection signatures have been investigated to constrain the parameter space for dark matter @xcite . in the present paper
we have calculated intensity and angular power spectrum of the cosmological background of synchrotron emission from the electrons and positrons produced in annihilations of cold dark matter .
the resulting radio background around @xmath265ghz and its angular power spectrum for multipoles @xmath251 has comparable or better sensitivity to dark matter annihilation cross sections than other signatures .
furthermore , a comparison of observed radio source counts with predictions for dark matter annihilation results in the constraint @xmath263 for the parameter defined in eq .
( [ eq : f_dm ] ) . under reasonable assumptions on dark matter clustering and magnetic fields in the halo environment ,
the range of annihilation cross sections corresponding to the constraint @xmath263 is comparable to constraints from synchrotron emission in an nfw profile @xcite .
galactic @xmath3ray constraints derived under similar assumptions @xcite are also comparable .
the sensitivity of our signal is considerably better than conservative limits based on annihilation into neutrinos @xcite , and comparable to limits on annihilation into @xmath3rays from diffuse cosmological emission @xcite .
sensitivities to values of order ten for the parameter @xmath266 defined in eq .
( [ eq : f_dm ] ) are interesting for non - thermal dark matter whose annihilation cross sections can be larger than our fiducial value @xmath267 , the cross section required for thermal dark matter .
an example for a non - thermal dark matter candidate with large cross section is the wino lsp occurring in supersymmetric theories with anomaly mediation @xcite .
we believe that radio observations in particular with future instruments such as ska can provide valuable information on dark matter .
we acknowledge partial support by the dfg ( germany ) under grants sfb-676 and gk 602 , and by the european union under the ilias project ( contract no .
rii3-ct-2004 - 506222 ) .
l.z . would like to thank shinichiro ando , yanchuan cai and yan qu for valuable discussions .
the mass function of the halo distribution is derived from the press - schechter formalism @xcite . in this approach
fluctuations in the linear density field with @xmath268 decouple from the local hubble expansion of the universe and collapse to form non - linear structures .
the fraction of the volume that has collapsed is predicted to be @xmath269 where @xmath270 is the co - moving radius over which the density field has been smoothed , which is related to the halo mass by @xmath271 with @xmath272 the co - moving matter density of the universe .
the number density of halos is then found to be given by @xcite @xmath273 where @xmath274e^{-\frac{a\nu^2}{2 } } \ , .
\label{eqn : fnu}\ ] ] here @xmath275 and @xmath276 is the critical density required for spherical collapse at a redshift @xmath13 in an einstein - de sitter space .
the variance in the density field smoothed with a top - hat filter of radius @xmath277 is @xmath278 where @xmath279\,,\ ] ] @xmath236 is the linear matter power spectrum , and @xmath280^{-3 } } { h_0\int_{0}^{\infty}dz^{\prime}(1+z^{\prime})[h(z^{\prime})]^{-3}}\ ] ] is the growth factor with linear perturbation theory , often also denoted by @xmath281 . in eq .
( [ equ : fnu ] ) @xmath282 , @xmath283 , and @xmath284 are constants , with the canonical press - schechter ( ps ) and sheth - tormen ( st ) mass functions corresponding to the parameters @xmath285 and @xmath286 , respectively .
the normalization @xmath282 is determined by requiring mass conservation such that @xmath287 for ps @xmath288 and for st @xmath289 .
the primordial power spectrum @xmath290 can be modified by the content and evolution of different matter components of the universe due to the perturbations that enter the horizon at different epochs .
this allows one to relate the linear power spectrum to the primordial power spectrum through a transfer function @xmath291 via @xmath292 fitting formula for an adiabatic cdm model give @xcite @xmath293^{-1/4}\ ] ] where @xmath294 and @xmath295 $ ] .
one usually uses the rms fluctuation on an @xmath296mpc scale to normalize the amplitude of the present power spectrum . from wmap 5 year data ,
we adopt @xmath297 @xcite , and @xmath298 . following ref .
@xcite , we may furthermore write the linear growth factor as @xmath299 where an approximate expressionfor @xmath300 is @xmath301 the halo mass function has to be supplemented by the dark matter density profile . for the dark matter profile within each halo we use an nfw profile @xcite , @xmath302 where @xmath143 is a characteristic radius .
within the context of the spherical collapse model , the outer extent of the cluster is taken to be the virial radius @xmath303^{\frac{1}{3 } } \ , , \ ] ] where @xmath304 is the average physical background matter density of the universe at redshift @xmath13 , and @xmath305\ ] ] is the overdensity of the halo relative to the background density @xcite .
the ratio of the virial radius to the scale radius is called the concentration parameter @xmath306 .
a combination of the definitions of virial mass and density profile gives @xmath307 together , @xmath144 and @xmath68 completely determine the dark matter distribution of a given halo .
note that an nfw profile is a conservative assumption compared to steeper profiles that have been proposed , for example the moore profile @xcite , which would consequently lead to a larger dark matter signal .
it is still uncertain how the concentration parameter depends on mass and redshift . in our paper
we consider a realistic value of @xmath308 .
extrapolating to the low mass range , we use @xcite @xmath309 where the collapse redshift @xmath155 is implicitly given by the relation @xmath156 , where @xmath157 is the mass scale at which @xmath158 . in a less conservative parametrization motivated by numerical simulations
@xmath144 would follow a log - normal distribution with standard deviation @xmath310 , @xmath311 ^ 2}{2\sigma_c^2}}}\right)\,,\ ] ] where the mean concentration parameter @xmath312 is related to the halo mass via @xcite @xmath313^{-\alpha_{c}}\,,\ ] ] where @xmath314 and @xmath315 are constants whose numerical values @xcite are typically chosen to be @xmath316 and @xmath317 . however , application of this parameterization to low - mass halos and to high redshift give values inconsistent with some simulations @xcite . a third parameterization @xcite uses @xmath318^{b(z)}\,,\ ] ] with @xmath319 , and @xmath320 .
the parametrization eqs .
( [ eqn : pc ] ) , ( [ eqn : pc2 ] ) give comparable dark matter signals , whereas the parametrization eqs .
( [ eqn : sj ] ) would lead to signals about a factor 20 higher than our conservative calculation . for the linear dark matter halo bias
@xmath321 appearing in eq .
( [ eq : l22 ] ) we adopt @xcite @xmath322 whereas for the galaxy bias we simply use unity . in order to compute the fourier transform of galaxy density profiles ,
we first need the galaxy mass . for normal galaxies we use the relation @xcite
@xmath323 where @xmath324 is the luminosity at 60 microns , and @xmath325 ghz is its frequency , and @xmath326s is the solar bolometric luminosity . for radio galaxies we simply adopt their typical mass of about @xmath327 to estimate the fourier transform of the their density profile .
( solid lines ) and @xmath328 ( dotted lines ) of @xmath201 and @xmath329 , respectively , as functions of co - moving wavenumber @xmath35 .
the vertical lines denote the scale @xmath330.,scaledwidth=60.0% ] the fourier transform of the spherically symmetric nfw profile of mass m can be written as @xmath331 and analogously for @xmath332 . for the purpose of plotting these fourier transforms , see fig .
[ fig : fourier_rho ] , it is convenient to renormalize them to unity for @xmath333 by introducing the new functions @xmath334 and @xmath335
. we then have @xmath336 , and @xmath337 for @xmath338 . for the nfw density profile , @xmath339 , where @xmath340/\left[\log(1+c)-c/(1+c)\right]^2$],and @xmath341 .
for the nfw profile , the mass of the halo within radius @xmath111 increases @xmath342 for @xmath343 , and then increase logarithmically for @xmath344 where @xmath198 .
therefore , the dominant contribution to the halo mass comes from @xmath343 .
similarly , the annihilation signal is produced mainly within @xmath343 , increasing there @xmath345 , but increases only @xmath346 for @xmath344 .
[ fig : fourier_rho ] shows that for @xmath347 we have @xmath348 , whereas for @xmath349 one has @xmath350 , and @xmath351 .
the radio intensity @xmath352 at a given frequency @xmath12 can be expressed in terms of antenna temperature @xmath353 via @xmath354 , where @xmath314 is the speed of light .
alternatively , @xmath352 can be written in terms of the thermodynamic temperature as the temperature of a blackbody with the given intensity at frequency @xmath12 , thus @xmath355 , where @xmath356 with @xmath357 the planck constant .
thus , for power law spectra @xmath358 , @xmath359 . in general , the cmb is expressed in terms of thermodynamic temperature @xmath360 , while galactic and extragalactic foregrounds are expressed in term of antenna temperature .
thermodynamic and antenna temperature are then related by @xmath361 , and their fluctuations by @xmath362 . for the cmb , @xmath363 with the cmb temperature @xmath364k @xcite . since we consider frequencies @xmath365ghz in the present paper , @xmath366 and thus @xmath367 and @xmath368 . from the definition of @xmath369 we get @xmath370 since for @xmath358 the power spectrum @xmath371 of @xmath352 at frequency @xmath12 scales as @xmath372
, we can express it in terms of the power spectrum @xmath373 of the antenna temperature @xmath369 at frequency @xmath374 via @xmath375 here , @xmath376 for synchrotron emission and @xmath377 for free - free emissions , respectively @xcite .
for the normalization and the dependence on @xmath187 , we adopted the best - fit model from observations at 2.3 ghz @xcite .
these parametrizations have been used in fig .
[ fig : cl_sum ] . rays from dark matter annihilation for our fiducial scenario , with @xmath150 .
this is consistent with fig . 1 in ref
. @xcite within about 10%.,scaledwidth=60.0% ] rays from dark matter annihilation in our fiducial scenario .
brown lines are for minimal halo mass @xmath151 , whereas red lines are for @xmath150 .
the higher and lower curves at high @xmath187 denote the one - halo and two - halo terms , respectively . note that @xmath46 is now normalized to the total intensity squared , @xmath378 .
this is consistent with fig . 6 in ref .
@xcite within a factor @xmath379.,scaledwidth=60.0% ] we can use our approach also to compute diffuse energy spectra and angular power spectra of @xmath3rays from dark matter annihilation . for neutralinos the spectrum of @xmath3rays of energy @xmath380 per annihilation can be parameterized by the simple expression @xcite @xmath381 to reasonably match the egret data @xcite , and compare with the results in ref .
@xcite , we multiply the predicted average @xmath3ray intensity spectrum by the boost factor @xmath382 due to substructure within the host halos .
note that this boost factor is much more extreme than @xmath168 assumed in the present work .
we then have @xmath383 for the weight function in eq .
( [ eq : limber1 ] ) . for our fiducial dark matter scenario
we then obtain fig .
[ fig : i_gamma ] for the solid angle averaged energy spectrum and fig .
[ fig : cl_gamma ] for the angular power spectrum at @xmath3ray energy @xmath384gev .
note that if fainter sources were included , corresponding to smaller minimal halo mass @xmath136 , the relative fluctuations normalized to @xmath378 , and especially the one - halo poisson term , would decrease .
the diffuse energy spectrum is consistent with the results in ref .
@xcite within about 10% .
in addition , the angular power spectrum at @xmath3ray energy 10 gev is also consistent with the results in ref .
@xcite within a factor two .
this serves as an important cross - check of our results with independent calculations .
10 e. komatsu _ et al . _
[ wmap collaboration ] , arxiv:0803.0547 [ astro - ph ] .
j. dunkley _ et al . _
[ wmap collaboration ] , arxiv:0803.0586 [ astro - ph ] .
g. bertone , d. hooper and j. silk , phys .
rept . * 405 * , 279 ( 2005 ) [ arxiv : hep - ph/0404175 ] .
l. bergstrom , j. edsjo and p. ullio , annihilations , phys .
lett . * 87 * , 251301 ( 2001 ) [ arxiv : astro - ph/0105048 ] .
p. ullio , l. bergstrom , j. edsjo and c. g. lacey , look , phys .
d * 66 * , 123502 ( 2002 ) [ arxiv : astro - ph/0207125 ] .
j. e. taylor and j. silk , mon . not .
soc . * 339 * , 505 ( 2003 ) [ arxiv : astro - ph/0207299 ] .
d. elsaesser and k. mannheim , background , phys .
lett . * 94 * , 171302 ( 2005 ) [ arxiv : astro - ph/0405235 ] .
s. ando and e. komatsu , phys .
d * 73 * , 023521 ( 2006 ) [ arxiv : astro - ph/0512217 ] .
s. ando , e. komatsu , t. narumoto and t. totani , phys .
d * 75 * , 063519 ( 2007 ) [ arxiv : astro - ph/0612467 ] .
d. hooper , g. zaharijas , d. p. finkbeiner and g. dobler , wmap haze , arxiv:0709.3114 [ astro - ph ] .
a. cuoco , j. brandbyge , s. hannestad , t. haugboelle and g. miele , phys .
d * 77 * , 123518 ( 2008 ) [ arxiv:0710.4136 [ astro - ph ] ] . s. ando , e. komatsu , t. narumoto and t. totani , mon . not .
soc . * 376 * , 1635 ( 2007 ) [ arxiv : astro - ph/0610155 ] .
g. d. mack , t. d. jacques , j. f. beacom , n. f. bell and h. yuksel , arxiv:0803.0157 [ astro - ph ] . h. yuksel , s. horiuchi , j. f. beacom and s. ando , phys . rev .
d * 76 * , 123506 ( 2007 ) [ arxiv:0707.0196 [ astro - ph ] ] .
l. bergstrom , j. edsjo , m. gustafsson and p. salati , jcap * 0605 * , 006 ( 2006 ) [ arxiv : astro - ph/0602632 ] .
t. bringmann and p. salati , phys .
d * 75 * , 083006 ( 2007 ) [ arxiv : astro - ph/0612514 ] .
t. delahaye , r. lineros , f. donato , n. fornengo and p. salati , phys .
d * 77 * , 063527 ( 2008 ) [ arxiv:0712.2312 [ astro - ph ] ] .
g. bertone , g. sigl and j. silk , mon . not .
soc . * 326 * , 799 ( 2001 ) [ arxiv : astro - ph/0101134 ] .
g. bertone , g. sigl and j. silk , mon . not .
* 337 * , 98 ( 2002 ) [ arxiv : astro - ph/0203488 ] .
d. hooper , phys .
d * 77 * , 123523 ( 2008 ) [ arxiv:0801.4378 [ hep - ph ] ] .
p. grajek , g. kane , d. j. phalen , a. pierce and s. watson , arxiv:0807.1508 [ hep - ph ] .
m. regis and p. ullio , arxiv:0802.0234 [ hep - ph ] .
r. j. protheroe and p. l. biermann , astropart .
* 6 * , 45 ( 1996 ) [ erratum - ibid . * 7 * , 181 ( 1997 ) ] [ arxiv : astro - ph/9605119 ] .
rybicki g.b . ,
lightman a.p .
, radiative processes in astrophysics , 1979 , john wiley & sons .
w. h. press and p. schechter , astrophys .
j. * 187 * , 425 ( 1974 ) .
r. k. sheth and g. tormen , mon . not .
soc . * 308 * , 119 ( 1999 ) [ arxiv : astro - ph/9901122 ] .
j. f. navarro , c. s. frenk and s. d. m. white , astrophys . j. * 462 * , 563 ( 1996 ) [ arxiv : astro - ph/9508025 ] ; j. diemand , b. moore and j. stadel , nature * 433 * ( 2005 ) 389 [ arxiv : astro - ph/0501589 ]
. s. hofmann , d. j. schwarz and h. stoecker , phys .
d * 64 * , 083507 ( 2001 ) .
x. l. chen , m. kamionkowski and x. m. zhang , phys .
d * 64 * , 021302 ( 2001 ) .
a. m. green , s. hofmann and d. j. schwarz , jcap * 0508 * , 003 ( 2005 ) .
kronberg , rept .
* 57 * , 325 ( 1994 ) .
j. s. bullock _ et al .
_ , mon . not .
soc . * 321 * , 559 ( 2001 ) [ arxiv : astro - ph/9908159 ] .
u. seljak , mon . not .
318 * , 203 ( 2000 ) [ arxiv : astro - ph/0001493 ] .
a. cooray and r. k. sheth , phys .
rept . * 372 * , 1 ( 2002 ) [ arxiv : astro - ph/0206508 ] .
g. tormen , a. diaferio and d. syer , mon . not .
299 * , 728 ( 1998 ) .
a. a. klypin , s. gottlober and a. v. kravtsov , astrophys . j. * 516 * , 530 ( 1999 ) . b. moore , s. ghigna , f. governato , g. lake , t. quinn , j. stadel and p. tozzi , astrophys . j. * 524 * ( 1999 ) l19
. s. ghigna , b. moore , f. governato , g. lake , t. quinn and j. stadel , astrophys .
j. * 544 * , 616 ( 2000 ) .
v. springel , s. d. m. white , g. tormen and g. kauffmann , mon . not .
soc . * 328 * , 726 ( 2001 ) .
a. r. zentner and j. s. bullock , astrophys .
j. * 598 * , 49 ( 2003 ) . g. de lucia _ et al .
_ , mon . not .
soc . * 348 * , 333 ( 2004 ) .
a. helmi , s. d. m. white and v. springel , phys .
d * 66 * , 063502 ( 2002 ) .
l. gao , s. d. m. white , a. jenkins , f. stoehr and v. springel , mon . not .
* 355 * ( 2004 ) 819
. l. shaw , j. weller , j. p. ostriker and p. bode , astrophys . j. * 646 * , 815 ( 2006 ) .
j. diemand , m. kuhlen , p. madau , m. zemp , b. moore , d. potter and j. stadel , arxiv:0805.1244 [ astro - ph ] .
x. j. bi , nucl .
b * 741 * , 83 ( 2006 ) [ arxiv : astro - ph/0510714 ] .
q. yuan and x. j. bi , jcap * 0705 * , 001 ( 2007 ) [ arxiv : astro - ph/0611872 ] . j. lavalle , q. yuan , d. maurin and x. j. bi , arxiv:0709.3634 [ astro - ph ]
. s. w. barwick _ et al .
_ [ heat collaboration ] , astrophys .
j. * 482 * , l191 ( 1997 ) [ arxiv : astro - ph/9703192 ] . s. coutu _ et al .
_ , _ prepared for 27th international cosmic ray conference ( icrc 2001 ) , hamburg , germany , 7 - 15 aug 2001
_ a. w. strong , i. v. moskalenko and o. reimer , astrophys .
j. * 613 * , 956 ( 2004 ) [ arxiv : astro - ph/0405441 ] .
p. sreekumar _ et al . _
[ egret collaboration ] , astrophys .
j. * 494 * , 523 ( 1998 ) [ arxiv : astro - ph/9709257 ] .
e. waxman and a. loeb , astrophys . j. * 545 * , l11 ( 2000 ) [ arxiv : astro - ph/0007049 ] .
u. keshet , e. waxman and a. loeb , astrophys .
j. * 617 * , 281 ( 2004 ) [ arxiv : astro - ph/0402320 ] . g. f. smoot , arxiv : astro - ph/9902201 . j. c. mather , d. j. fixsen , r. a. shafer , c. mosier and d. t. wilkinson , astrophys .
j. * 512 * , 511 ( 1999 ) [ arxiv : astro - ph/9810373 ] .
l. spinoglio , m. a. malkan , b. rush , l. carrasco and e. recillas - cruz , astrophys .
j. * 453 * , 616 ( 1995 ) [ arxiv : astro - ph/9506139 ] . c. a. jackson , new astron .
rev . * 48 * ,
1187 ( 2004 ) [ arxiv : astro - ph/0409180 ] . m. tegmark and g. efstathiou , arxiv : astro - ph/9507009 . m. tegmark , d. j. eisenstein , w. hu and a. de oliveira - costa , astrophys .
j. * 530 * , 133 ( 2000 ) [ arxiv : astro - ph/9905257 ] .
g. giardino , a. j. banday , p. fosalba , k. m. gorski , j. l. jonas , w. omullane and j. tauber , arxiv : astro - ph/0103233 .
l. la porta , c. burigana , w. reich and p. reich , arxiv:0801.0547 [ unknown ] .
r. a. sunyaev and y. b. zeldovich , astrophys .
space sci .
* 7 * , 3 ( 1970 ) .
f. m. , cen , r. , ostriker , j. p. , astrophys .
j. * 442 * , 1 ( 1995 ) j. m. siegal - gaskins , arxiv:0807.1328 [ astro - ph ] .
x. l. chen and m. kamionkowski , phys .
d * 70 * , 043502 ( 2004 ) [ arxiv : astro - ph/0310473 ] .
l. zhang , x. l. chen , y. a. lei and z. g. si , phys .
d * 74 * , 103519 ( 2006 ) [ arxiv : astro - ph/0603425 ] .
l. zhang , x. chen , m. kamionkowski , z. g. si and z. zheng , phys . rev .
d * 76 * , 061301 ( 2007 ) [ arxiv:0704.2444 [ astro - ph ] ] .
z. myers and a. nusser , arxiv:0710.0135 [ astro - ph ] . n. padmanabhan and d. p. finkbeiner , phys
. rev .
d * 72 * , 023508 ( 2005 ) [ arxiv : astro - ph/0503486 ] .
s. dodelson , d. hooper and p. d. serpico , phys .
d * 77 * , 063512 ( 2008 ) [ arxiv:0711.4621 [ astro - ph ] ] .
l. randall and r. sundrum , nucl .
b * 557 * , 79 ( 1999 ) [ arxiv : hep - th/9810155 ] .
g. f. giudice , m. a. luty , h. murayama and r. rattazzi , jhep * 9812 * , 027 ( 1998 ) [ arxiv : hep - ph/9810442 ] .
j. m. bardeen , j. r. bond , n. kaiser and a. s. szalay , astrophys . j. * 304 * ( 1986 ) 15 .
s. m. carroll , w. h. press and e. l. turner , ann .
astrophys . * 30 * ( 1992 ) 499 .
j. p. henry , astrophys .
j. * 534 * ( 2000 ) 565 [ arxiv : astro - ph/0002365 ] .
b. moore , t. quinn , f. governato , j. stadel and g. lake , mon . not .
soc . * 310 * , 1147 ( 1999 ) [ arxiv : astro - ph/9903164 ] .
h. j. mo , y. p. jing and s. d. m. white , mon . not .
soc . * 282 * , 1096 ( 1996 ) [ arxiv : astro - ph/9602052 ] . | we calculate intensity and angular power spectrum of the cosmological background of synchrotron emission from cold dark matter annihilations into electron positron pairs .
we compare this background with intensity and anisotropy of astrophysical and cosmological radio backgrounds , such as from normal galaxies , radio - galaxies , galaxy cluster accretion shocks , the cosmic microwave background and with galactic foregrounds . under modest assumptions for the dark matter clustering
we find that around 2 ghz average intensity and fluctuations of the radio background at sub - degree scales allows to probe dark matter masses @xmath0gev and annihilation cross sections not far from the natural values @xmath1 required to reproduce the correct relic density of thermal dark matter .
the angular power spectrum of the signal from dark matter annihilation tends to be flatter than that from astrophysical radio backgrounds .
furthermore , radio source counts have comparable constraining power .
such signatures are interesting especially for future radio detectors such as ska . |
massive stars inject a significant amount of energy into the surrounding medium in various forms : photoionization / heating by stellar radiation and mechanical energy through stellar winds and supernova ( sn ) explosions , some of which later transforms into gas thermal energy and radiation .
for example , ob stars deposit stellar radiation energy into the interstellar medium ( ism ) with the luminosity of @xmath8 and the h lymann - limit photon luminosity ( @xmath9ev ) of @xmath10 @xcite . so the radiative feedback of an early type star during the main sequence ( @xmath11 yrs ) may amount up to @xmath12ergs , which is comparable to typical sn explosion energy .
these stars also inject mechanical energy into the ism via stellar winds during the main sequence ( ms ) and the red supergiant ( rsg ) or luminous blue variable ( lbv ) phase , and also wolf - rayet stage in the case of very massive stars with @xmath13 @xcite .
the ms winds have a typical wind speed of @xmath14 and a mass loss rate of @xmath15 , while the rsg winds have much lower speed , @xmath16 , but higher mass loss rate , @xmath17 @xcite . owing to much longer duration
the ms wind generates a bigger bubble , while the rsg wind produces a much smaller but denser shell inside the ms bubble created earlier .
the mechanical energy luminosity of the ms winds ranges @xmath18 @xcite .
massive stars with mass @xmath19 end their lives as core - collapse sne and deposit typically @xmath20 ergs in the form of kinetic energy of the ejecta . initially the sn ejecta expands freely into the wind - blown cavity and later interacts with the shell produced previously by the winds , generating various shocks , transmitted and reflected shocks .
most of the explosion energy is thermalized via such shocks .
after the shock - heated gas cools adiabatically down to @xmath4 k or after the forward ejecta shock catches up with the wind bubble shell , the supernova remnant ( snr ) begins to lose a significant fraction of thermal energy via radiative cooling .
the final energy budget deposited into the surrounding medium in the forms of thermal , kinetic , and radiative energy depends on the detailed history of radiative cooling and heat transport by thermal conduction , which in turn depends on the size and structure of the preexisting wind bubble as well as on the ambient ism .
for example , if a sn explodes in a dense environment , the explosion energy would be lost mostly by radiative cooling and the impact to the ism should be minimal . according to @xcite , in the case of sn explosions inside wind - blown bubbles in the typical ism with @xmath21 , the remaining kinetic energy fraction is @xmath22 at the end of snr evolution , while the remaining thermal energy fraction , @xmath23 .
feedback from massive stars should play a significant role in the formation and evolution of galaxies .
especially , feedback from the first - generation sne affects the formation and evolution of proto - galaxies and the large scale structure in the early universe ( * ? ? ?
* ; * ? ? ?
* and references therein ) .
although sn feedback is an important ingredient for galaxy formation , it occurs typically on length and time scales much smaller than grid spacings and time steps of numerical simulations .
so such subgrid physics is often treated through phenomenological prescriptions in numerical simulations for galaxy formation and large scale structure formation @xcite .
either ` thermal ' feedback @xcite or ` kinetic ' feedback @xcite with a feedback efficiency parameter @xmath24 , which is the fraction of sn explosion energy deposited in the form thermal or kinetic energy , were tried with varying degrees of success . for a thermal feedback model
, the radiative cooling must be turned off artificially for a period of time in order to prevent the dense gas from re - radiating the deposited energy immediately .
@xcite showed that both thermal and kinetic feedback models can alleviate the so - called `` catastrophic cooling '' problem of galaxy formation . thus in order to construct a physical subgrid prescription for sn feedback in numerical studies of galaxy formation it is necessary to estimate quantitatively the feedback efficiency parameters , @xmath25 , @xmath26 , and @xmath27 for thermal , kinetic and radiation energy , respectively .
massive stars often form in an association , so multiple explosions of core - collapse sne can create a large superbubble @xcite .
multiple sn explosions in the proto - galactic environment , in particular , may have played a significant role in the formation of halo globular clusters . in a self - enrichment model of metal - poor globular clusters ,
proto - globular cluster clouds are assumed to be enriched by multiple sn explosions from massive first - generation stars contained within the clouds and later second - generation stars form in the cooled shell driven by the sn explosions ( * ? ? ? * and references therein ) .
however , some have argued that the binding energy of a proto - globular cluster cloud may not be large enough to survive a few sne , and that the homogeneity of metalicity within a globular cluster indicates that pre - enrichment must have occurred within larger systems @xcite .
according to numerical studies of early structure formation , first objects might form at a redshift of @xmath28 with a typical mass of @xmath29 and first - generation stars enrich the protogalactic clouds before they merge into larger clouds of @xmath30 via gravitational clustering @xcite .
this favors the model that combines the pre - enrichment by first stars before the formation of protogalactic halos and the self - enrichment by next generation sne within proto - globular cluster clouds . in this scenario
the stability of the cooled shell and further star formation critically depends on how efficiently the sn energy is radiated away during the snr evolution .
thus in the present work we attempt to estimate feedback effects of core - collapse sne , including multiple detonation , through numerical hydrodynamic simulations .
we first calculate the creation of a bubble by stellar winds from multiple massive stars ( @xmath31 ) during the precedent ms stage .
the end results of the ms wind stage are adopted as the initial states of the sn explosion simulations .
we then make quantitative estimates of the fractions of explosion energy transferred into the ism in different forms , thermal , kinetic , and radiation energy .
diffuse radiation emitted by the shock - heated gas , especially , ionizing radiation for hydrogen and helium atoms are calculated . in 2 ,
the basic features of our numerical simulations , including radiative cooling and thermal conduction , are described .
we then present the numerical results of wind bubble models in 3 , and multiple supernova models in 4 .
a brief summary is given in 5 .
we solve the hydrodynamic conservation equations of the ideal gas in one - dimensional spherical coordinate system , including radiative cooling and thermal conduction : @xmath32 @xmath33 @xmath34 = - \lambda - { 1 \over r^2 } { \partial \over \partial r}(r^2s),\ ] ] where @xmath35 is the radial velocity , @xmath36}+ u^2/2 $ ] is the total energy of the gas per unit mass , @xmath37 is the energy loss rate per unit volume , and @xmath38 is the radial heat flux .
a ratio of specific heats is assumed to be @xmath39 .
the gas temperature is determined by @xmath40 , where @xmath41 is boltzmann s constant , @xmath42 is the mass of hydrogen nucleus and @xmath43 is the mean molecular weight .
we use a grid - based , eulerian hydrodynamics code in the spherical symmetric geometry , based on the total variation diminishing ( tvd ) scheme @xcite .
the tvd scheme solves a hyperbolic system of gasdynamical conservation equations with a second - order accuracy .
it has been widely used for a variety of astrophysical problems @xcite .
we define the cooling rate as @xmath44 where @xmath45 is the number density of hydrogen nuclei . for the cooling rate coefficient , @xmath46
, we adopt the non - equilibrium radiative cooling rate for an optically thin gas that is calculated by following the non - equilibrium collisional ionization of the gas cooling from @xmath47k to @xmath48 k under the isobaric condition @xcite .
we consider cases with the solar metalicity ( @xmath49 ) and with 1/10 of the solar metalicity ( @xmath50 ) .
although we calculate diffuse radiation emitted by the shock - heated gas , we do not explicitly follow the radiative transfer of diffuse radiation and photoionization of the gas .
instead we assume @xmath51 for @xmath52k and set the minimum temperature at @xmath53k .
this is in effect equivalent to assuming thermal equilibrium at @xmath48k between radiative cooling and photoionization heating . after completing the hydrodynamic part , updating hydrodynamical quantities , radiative cooling
is applied to the thermal energy per unit volume , @xmath54 , as a separate step @xcite : @xmath55 where @xmath56 is the cooling time scale and @xmath57 is the integration time step .
this is designed to follow approximately the thermal history of the gas that cools radiatively and approaches the specified minimum temperature in a time scale much shorter than the dynamical time scale @xcite . in some astrophysical plasmas such as solar corona and stellar winds
thermal conduction can be important in heat transport .
we include the so - called classical spitzer conduction for a fully ionized gas @xcite : @xmath58 @xmath59 where @xmath60 and @xmath61 is the electron number density .
this classical formula is based on the assumption that the mean free path of the electrons is short compared to temperature scale height .
however , when temperature scale height is comparable or shorter than the electron mean free path , thermal conduction reaches the saturation limit and heat flux is no longer equal to @xmath62 . in the saturation limit
, heat flux is given approximately by @xmath63 where @xmath64 is the electron mass @xcite .
so we set @xmath65 $ ] as local heat flux . with the classical formula
the thermal conduction term becomes a diffusion term , @xmath66 .
this type of partial differential equation can be solved _
implicitly _ by cranck - nicholson scheme @xcite . for the saturated formula ,
however , this can not be applied .
so we solve the thermal conduction term _ explicitly _ and the integration time step is chosen as @xmath67 where @xmath68 is the sound speed . in grid - based eulerian simulations a shock transition
is typically resolved over several grid zones with a strong temperature gradient , although the physical shock thickness may be much smaller than the size of one grid zone .
so un - physically rapid radiative cooling and heat transport can occur across the numerical shock transition zones . to avoid such numerical artifacts , we identify zones that undergo shocking process and turn off radiative cooling and heat transport by thermal conduction inside those zones .
we neglect magnetic fields and dust in our calculations .
if there exist dynamically significant , regular magnetic fields in the ambient medium , the assumed spherical symmetry would brake down . also coherent tangential magnetic fields would substantially reduce the radial thermal conduction flux .
dust may play some important roles through radiation pressure on dust and dust cooling in the cooled bubble shell in a more realistic situation .
proper treatments of such processes require radiative transfer , chemistry and dust formation / destruction , which are beyond the scope of this study .
however , if the bubble shell stays photoionized at @xmath48 k by diffuse radiation , as we assume here , then the effects of dust are likely to be minimal .
@xcite , for example , included dust cooling in their simulations of multiple sne and found that the dust cooling effects were only minor .
we first consider the development of a wind bubble to study the pre - supernova circumstellar environment .
massive early - type stars emit a steady stellar wind with constant terminal velocity @xmath69 and mass - loss rate @xmath70 .
we assume that the progenitor star has an initial mass @xmath71 and spends @xmath72 in the main sequence phase .
the ms wind is characterized with a mass - loss rate , @xmath73 , and wind velocity , @xmath74 @xcite .
here we consider only the ms wind , since the rsg wind produces a small shell with the radius of a few pc inside the bubble created by the ms wind .
the rsg wind shell would be swept up quickly by the sn ejecta during the early free expansion stage , so interactions of the ejecta with this shell can be ignored .
we calculate the evolution of a bubble blown by collective stellar winds from multiple massive stars in an association .
the luminosity of the collective wind is the key physical parameter that determines the properties and evolution of the bubble and it is given by @xmath75 we select the number of stars , @xmath76 and 50 , because a typical ob association contains 10 - 100 early - type stars ( see table 1 ) . in a realistic model , one needs to specify the initial mass function of the association to obtain a total wind luminosity . in our simplified model
, however , @xmath77 , is parameterized with @xmath78 .
so a larger value of @xmath77 can be regarded as a higher wind luminosity either from a more massive star or from multiple stars .
we consider @xmath79 as the fiducial density of the ambient ism ( wf models ) , where @xmath80 .
two additional values are considered : @xmath81 ( wl01 model ) and @xmath82 ( wh01 model ) .
the ambient medium is assumed to be photoionized at @xmath83k .
so its sound speed is @xmath84 and the pressure is @xmath85 . with wz01 and wz10 models , we consider metal poor cases with @xmath50 .
the wind is realized numerically by injecting the gas flow with @xmath69 at the inner boundary ( @xmath86 ) of the spherical simulation grid .
the wind density at the inner boundary is given by @xmath87 $ ] .
we use @xmath88 3000 uniform spherical grid zones in 150 pc radius with a spatial grid spacing of @xmath89 pc .
numerical convergence is often an important issue for hydrodynamic simulations with radiative cooling , because the cooling rate has a strong density dependence ( @xmath90 ) .
so in order to test numerical convergence of the simulations wf01 model is calculated with several different grid spacings .
1 shows the structure of the bubble in the simulations with 3 different grid spacings . the contact discontinuity ( cd ) between the shocked wind and the shocked
ism is around 23 pc , while the location of the conduction front ( cf ) is different in the three simulations .
the size of the hot cavity ( @xmath91 20pc ) is smaller than the so - called field length @xmath92^{1/2 } \approx 56 $ ] pc , where @xmath93k and @xmath94 are used @xcite .
so heat transport by thermal conduction is dominant over radiative cooling inside the cavity .
heat transport from the hot shocked wind gas to the warm shocked ism reduces the temperature of the shocked wind gas , triggering radiative cooling and compression .
these processes are not followed accurately enough to be converged in the simulations , so the density and temperature distributions are significantly different inside the cavity among these simulations .
however , the pressure distribution is very similar in the three calculations and so the dynamical evolution of the wind bubble shell seems to be converged .
radial positions of the cd and outer compression wave seem to be converged within a few percent , while the total energy inside the simulation volume seems to be converged within @xmath91 10 % .
this fast convergence is achieved , because the density compression factor at the shell is not very large in the simulations ( @xmath95 ) .
we also find that the differences in the density and temperature structures inside the cavity among the simulations with different grid spacings do not affect significantly the evolution of sn explosions that will be described in 4.2 .
a typical bubble has four distinct regions from the central star outward : 1 ) freely expanding wind with the density profile , @xmath96 , 2 ) a wind termination shock facing inward and the shocked wind region , 3 ) a contact discontinuity ( cd ) between the shocked wind region and the shocked ism region , 4 ) a forward shock propagating into the ambient ism @xcite .
while the termination shock is strong , a compression wave may form , instead of the outer forward shock , depending on the ratio of wind ram pressure to the ambient pressure .
2 shows the gas density and temperature profiles at @xmath97 years in the models listed in table 1 .
although the outermost structure is initially developed as a weak shock for all models , it becomes a compression wave with the propagation speed , @xmath98 in some models , as the bubble expands out into the ambient medium . for wf01 model , for example
, the pressure increases only by a factor of 1.5 across the compression wave . even for wf50 model the outer shock speeds
is only @xmath99 .
also the shock transition is much wider than that typically found in adiabatic simulations using the same tvd code . in wf50 model
the inner termination shock is at 10 pc , the conduction front at @xmath100 pc , the cd at 41 pc , and the outer shock wave at @xmath101 pc . without thermal conduction
, the hot cavity inside the cd would be almost isothermal with @xmath102 k. but heat transport by thermal conduction enhances radiative cooling near the conduction front , leading to a strong temperature gradient inside the cavity .
near the cd , the gas density changes by more than two orders of magnitude , so the gas cools quickly down to @xmath53k and forms a shell .
the thickness of the shell is rather thick ( @xmath103 ) .
the continuous injection of wind energy seems to be radiated away efficiently near the conduction front . for a spherical shell with inner radius , @xmath104 , outer radius , @xmath105 , and uniform density ,
the shell density can be derived from the condition that the shell mass is equal to the swept - up mass : @xmath106 so the density enhancement factor is small for thick shells found in our simulations . for wf01 model , for example ,
@xmath107 and @xmath108 . from fig .
2 we can see that the size of bubble increases with @xmath78 and decreases with @xmath109 .
@xcite showed that the bubble would evolve self - similarly and the outer radius of the shell would grow as @xmath110 , if radiative cooling and the ambient pressure were negligible .
they also derived a slightly modified time dependence of @xmath111 by numerical integrations of simplified conservation equations where an approximate cooling rate and an ambient pressure with @xmath1128000k were adopted . in our simulations
the ambient pressure is not negligible and the outward moving front is not a shock , so the self - similarity seems to be broken .
3 shows the evolution of @xmath105 and swept - up shell mass , @xmath113 , for all the models in table 1 . for our models with @xmath114 and @xmath115 ,
the radius and mass of wind - blown bubble shell can be approximated by @xmath116 and @xmath117 these relations are slightly different from what @xcite derived through simplified numerical integrations described above .
right panels of fig . 3 show how @xmath105 and @xmath118 depend on @xmath109 or the metalicity @xmath119 .
these quantities do not scale with the ambient density in a simple way , because the radiative energy loss depends not only on the gas density but also on the detailed history of how heat is transported from the hot cavity to the warm shell via thermal conduction .
it seems that for @xmath120 the swept - up mass is nearly independent of the ism density , while @xmath121 . for high density medium ( @xmath122 ) ,
the cooling time scale becomes much shorter than the dynamical time , and so the numerical simulation becomes increasingly difficult .
for wh01 model with @xmath123 , @xmath118 is about 3 times that of wf01 models .
however , this trend can not be generalized to higher density cases because of nonlinear behavior of heat transport and radiative cooling .
for wz01 with lower metalicity ( @xmath124 ) the cooling is less efficient and so the results are similar to those of wl01 model with lower ambient density .
bottom panels of fig .
3 show the fraction of energy lost due to radiative cooling . the radiation energy is calculated as the energy lost from the simulation volume , @xmath125 , where @xmath126 is the total energy inside the simulation volume and @xmath127 is its initial value . here
the thermal energy @xmath128 excludes the initial thermal energy of the ambient gas .
the energy loss due to radiative cooling is significant : about 60 % for wl01 and wlz01 models and about 60 - 80 % for wh01 and wf01-wf50 models . for wf models
the fraction of @xmath129 is reduced somewhat for higher @xmath77 ( larger @xmath78 ) , because higher power wind creates hotter cavities .
next we calculate the evolution of multiple supernovae exploded inside the preexisting wind bubble created by the ms winds .
so we take the final structures of the wind bubble simulations at @xmath130yrs , described in the previous section , as the states of the circumstellar medium prior to sn explosions .
each sn explosion is characterized by the explosion energy , @xmath131erg , and the ejecta mass , @xmath132 .
then the ejecta can be approximated by a uniform density core ( @xmath133 ) and an outer envelope with a steep power - law density profile ( @xmath134 ) , expanding freely with the radial velocity , @xmath135 @xcite . for a multiple explosion where @xmath136 supernovae detonate simultaneously , total explosion energy @xmath137 ,
the density profile of the ejecta is defined by @xmath138 @xmath139 where @xmath140}^{5/2 } } \over { { [ 10(n-5)e_{ej}]}^{3/2}}}},\ ] ] and @xmath141}^{1/2}}.\ ] ] we adopt @xmath142 , @xmath143 and @xmath144 .
we note that overall results of our calculations do not depend sensitively on the specific values of @xmath145 and @xmath146 adopted .
as shown below , the evolution of a superbubble is controlled mainly by the total explosion energy @xmath0 rather than @xmath146 .
also a snr should lose its memory of ejecta mass after the swept - up mass becomes much greater than @xmath145 .
table 2 shows the multiple sn explosion models considered here .
the number of sne is given by @xmath147 and varies from 1 to @xmath78 , since @xmath78 is the total number of possible sne inside a given bubble . as in the wind bubble stage , we consider three values of the ism density : @xmath79 as a fiducial value ( snf models ) , @xmath148 ( snl01 ) and @xmath82 ( snh01 ) .
since sn explosions can occur sequentially with a time interval , we also consider snf02a model in which a second sn detonates @xmath149 years after the first sn .
this will be compared with snf02b model in which two sne explode simultaneously . again
snz01 and snz10 models have a lower metalicity , @xmath50 .
we use the same grid setup as the wind simulations : @xmath88 3000 uniform grid zones in 150 pc radius . again
two additional simulations are carried out for snf01 model with smaller grid spacings ( @xmath88 6000 and 12000 ) to test numerical convergence .
radial position of the outer shock is converged within 0.5 % , while the total energy is converged within 5 % .
the simulations are terminated at @xmath150 years . fig .
4 shows the radial profile of the remnant for the single sn case ( snf01 model ) .
initially , the sn profile described by eqs .
( 14)-(17 ) is superimposed onto the final output of the corresponding wind bubble model ( wf01 ) : the wind termination shock is at 3 pc and the hot cavity extends out to 25 pc ( solid lines in fig .
the compression wave has weakened and the shell thickness is large ( @xmath151 ) , so the density of swept up shell , @xmath152 , is slightly higher than the ambient density ( see eq . [ 11 ] ) .
the sn blast shock wave passes through the pressure barrier produced by the wind termination shock and travels quickly through the hot cavity .
the forward shock first interacts with the bubble shell at @xmath153 years , pushing out the inner shell boundary and generating two secondary waves , i.e. , a transmitted shock propagating outward into the bubble shell and a weak reflected compression wave moving inward into the ejecta . a reverse shock forms in the sn ejecta around @xmath154 years and is reflected at the center around @xmath155 years .
afterward it becomes a traveling compression wave , because the shock heated gas inside the cavity is hot .
when this wave runs into the shell at @xmath156 years , the shell is heated for the second time . by the termination time
( @xmath4 years ) the shell is pushed out by the snr to @xmath15767 pc with inner radius @xmath15854 pc and the swept up shell gas has @xmath159 k and @xmath160 , not much higher than the ambient medium . on the other hand ,
the cavity is filled with the hot gas of @xmath161 k and @xmath162 .
the outer shock weakens with expansion and slows down to @xmath163 .
5 compares the two different models of double explosions , sequential ( snf02a ) and simultaneous ( snf02b ) explosion models . in snf02a model ,
the second sn explosion is loaded @xmath149 years after the first sn explosion .
three snapshots are shown : 1 ) at @xmath164 years the outer shock in snf02b model has expanded further out than the shock in snf02a model .
2 ) at @xmath165 years , a little after loading of the second sn in snf02a model , the second sn shock expands out to 35 pc , still well inside the shell at 42 - 46 pc generated by the first sn .
3 ) at @xmath166 years the sn - driven shell has expanded _ slightly _ further out ( @xmath9150 pc ) in snf02b model than in snf02a model .
however , the difference is small enough , so we say the bubble size is similar for the two models . the left panels of fig .
6 also show the comparison of the two double explosion models .
the outer shock radius and shell mass of the two models are very similar . here
@xmath128 and @xmath167 are total thermal and kinetic energy inside the simulation volume , respectively , excluding the initial thermal energy of the ambient gas .
the time evolution of @xmath128 and @xmath168 shows that the initial kinetic energy of the sn ejecta is transferred to the thermal energy gradually in the beginning as the blast wave expands , but rather abruptly at @xmath169 years when the blast wave runs into the preexisting wind - blown shell .
the thermalization process happens again at later times in snf02a model , when the second sn detonates at @xmath149 years and when its blast wave arrives at the first sn shell at @xmath170 years .
interestingly , the latter event occurs at the similar time when the aforementioned compression wave hits the shell in snf02b model , causing the second period of active radiative cooling .
most of thermal energy is lost through radiative cooling . here
the radiation energy is calculated by @xmath171 .
we note that the initial total energy , @xmath127 , includes both the sn energy and the energy of the preexisting wind bubble , so @xmath172 is greater than one . at the termination time of our simulations , the final energy budget in the two models are similar : 6% in thermal energy , 7% in kinetic energy , and 90% lost through radiation energy . figs .
5 - 6 imply that , for given total explosion energy , the different fashion of multiple explosion , either sequential or simultaneous detonation , does not make significant differences in the size and mass of the final bubble shell . the right panels of fig .
6 compare the models with different ambient density and different metalicity . at lower density or with lower metal abundance ,
the bubble size is bigger and @xmath129 is smaller , compared to higher density models , but the shell mass is similar for all the models considered here .
simultaneous explosions with 1 , 3 , 5 , and 10 sne inside a wind bubble generated by 10 massive stars are considered as snf10a - d models .
initial conditions are taken from wf10 model simulation : the cavity is bounded by the cd at 56 pc and the outer compression wave is at 82 pc .
the upper two panels of fig . 7 show the bubble structure at @xmath4 years for snf10a - d models .
more energetic explosion pushes the shell further out and produces a faster shock , leading to a larger and denser shell .
the lower panels of fig . 7 show the profiles of snrs in a uniform ism ( @xmath173 ) for comparison .
it shows that sn - driven bubbles are larger and hotter , if sne explode inside a preexisting bubble .
for example , the size of hot bubble for snf10a - d models is 33 - 55 % larger than that for sne inside a uniform ism , resulting in 2.4 - 3.7 times bigger volume of hot gas .
also the cavity temperature is 2 - 3 times higher in snf10a - d models .
this implies that the volume filling factor and the temperature of the hot bubbles generated by core - collapse sne would be underestimated by a factor of 2 - 4 , if the preexisting bubble is not properly accounted for .
the left panels of fig .
8 show the evolution of the bubble properties for snf10a - d models , while the right panels compare the snz10a - d models with lower metalicity .
the radial position of the outermost shock increases as @xmath174 before the arrival of the blast wave at the shell , and afterward it increases very slowly , @xmath175 .
the size of sn - driven bubbles depends on both the sn explosion energy and the size of pre - existing wind bubbles . according to the numerical results shown in fig .
8 , @xmath176 and @xmath177 depends rather weakly on @xmath0 , @xmath2 and @xmath3 .
again the evolution of @xmath128 , @xmath167 , and @xmath129 indicates the complex thermalization history via shocks .
the first major episode of radiative cooling occurs when the sn blast wave first hits the wind bubble at @xmath178 years . when the compression wave generated by the reverse shock runs into the shell at @xmath179 years , the shell is heated and compressed , leading to the second episode of rapid cooling . in snf10d model
this behavior is most distinct . for snf10a - d models ,
about 80 - 90 % of @xmath0 is lost via radiative cooling within @xmath4 years , while only 10 % remains as thermal energy and @xmath510 % as kinetic energy . for snz10a - d models with lower metal abundance
the radiative cooling is less efficient compared to snf10a - d models , so about 70 % is lost , 20 - 40 % remains as thermal energy and 10 % as kinetic energy .
since the shell has a significant kinetic energy at that time , it would continue to expand slowly .
exploration of later evolutionary stage requires more realistic treatments of radiative transfer , photoionization / heating , and non - equilibrium radiative cooling below @xmath48 k , which is beyond the scope of this study .
in the self - enrichment model of globular cluster formation , the shell of gas swept up by the blast wave from multiple sne must cool and be confined gravitationally within the parent proto - globular cloud in order for the enrichment process to be effective @xcite .
thus the success of this model lies on how efficiently the sn explosion energy is radiated away to form gravitationally bound system . according to fig .
8 ( snz10 models ) , in an average ism with @xmath115 and @xmath85 , massive stars in an association with @xmath180 would generate a bubble shell with @xmath181 pc and @xmath182 .
incidentally this mass is similar to the characteristic mass of globular clusters . if 10 % of explosion energy is left as the kinetic energy of the shell , then @xmath183 ergs . with the gas temperature of @xmath83 k ,
the total thermal energy of the photoionized shell is @xmath184 ergs .
but the gravitational energy , @xmath185 ergs , is too small to confine the shell gravitationally .
thus the self - enrichment model would not work inside such low density clouds , but may operate in much higher density environment . if we blindly take the simple scaling relations shown in figs .
6 and 8 , @xmath186 , but @xmath187 , then the gravitational energy of the shell would increase with the ambient density as @xmath188 . in a much denser cloud higher ambient pressure would also help the shell to be confined , possibly resulting in a gravitationally bound system .
as shown in the previous section , most of mechanical sn energy is transformed into thermal energy of the bubble shell and then lost via diffuse radiation .
so radiative feedback is a significant part of sn feedback into the surrounding ism .
in order to study the photoionization / heating feedback of sn explosions we calculate the diffuse radiation emitted by the bubble .
first , we generate both continuum and line emissivities , @xmath189 ( @xmath190 ) and @xmath191 ( @xmath192 ) , for @xmath193 as a function of gas temperature for photon energy @xmath194=[1.36\times10^{-4},10
^ 8]$ ] , using cloudy90 code ( ferland 1998 ) .
collisional ionization equilibrium is assumed and no background radiation field is adopted .
the emissivity tables are generated for two values of metalicity , @xmath49 and @xmath50 .
volume integrated spectrum of the diffuse radiation emitted by the bubble during the integration time @xmath195 is calculated by @xmath196 n_h^2(t , r ) dr,\ ] ] where @xmath197 is given in units of @xmath198 and line emissivities divided by the width of a given photon energy bin are added to the corresponding continuum emissivity . in practice ,
only @xmath199 is calculated during simulations and then @xmath200 $ ] is calculated in a post - processing step .
the upper panel of fig .
9 shows the evolution of @xmath201 for a single sn model .
most of hard x - ray photons with @xmath202 kev are emitted early before the blast wave runs into the wind bubble shell ( @xmath203 years ) .
afterward , the fluxes of lower energy photons increase with time , as the sn energy is deposited as heat into the bubble shell and radiated away mainly in uv and optical ranges .
the lower panel of fig .
9 compares @xmath197 for snf01 and snz01 models integrated to the final simulation time .
in snz01 model metal lines are weaker , but h and he recombination continuum radiation is stronger , because the temperature is higher due to lower cooling rate .
we also calculate the integrated energies and photon numbers of diffuse radiation that can ionize hydrogen and helium atoms as follows : @xmath204 @xmath205 where @xmath206 24.6 , and 54.4 ev is the ionization energy for h , he , and he@xmath207 , respectively .
these quantities are summarized in table 3 . for snf01 model ,
the total number of h ionizing photons is @xmath208 and the total energy carried away with those photons is @xmath209 .
the total number of h ionizing photons scales roughly with the sn explosion energy as @xmath210 .
diffuse radiation is enhanced overall in the models where a denser and hotter shell is produced . comparing the results of snl01 , snf01 and snh01 models in table 3
, we can see that more diffuse photons are produced at denser environments .
a sn with higher explosion energy generates a faster blast wave and produces a denser and hotter shell , emitting harder diffuse radiation .
the results of snf10a - d models show such trends . on the other hand ,
the comparison of snf02a and snf02b models indicate that the simultaneous double explosion generates more ionizing photons than the sequential double explosion , although the total radiation energy ( @xmath129 ) is similar in the two models .
snz models with lower metalicity produce less ionizing radiation and more non - ionizing radiation , compared to the models with solar metalicity .
10 shows the evolution of @xmath211 for snf10a - d and snz10a - d models .
most of ionizing photons are emitted during @xmath212 years when the sn shock propagates into the bubble shell . for models with @xmath213 ionizing photons
are emitted actively again when the reflected compression wave runs into the shell around @xmath214 years ( see also fig . 8) .
we study multiple supernova ( sn ) explosions inside a preexisting cavity blown up by stellar winds from massive progenitor stars through numerical hydrodynamic simulations .
calculations are performed with a grid - based eulerian code in one dimensional spherical symmetry , including radiative cooling and thermal conduction .
heat transport by thermal conduction is important inside the hot cavity , while radiative cooling is dominant in the bubble shell .
we consider models with a wide range of physical parameters : the ambient density , @xmath215 , the metalicity , @xmath49 and @xmath50 , the number of massive stars , @xmath31 , and the number of supernovae , @xmath216 . in order to study the radiative feedback effects
, we also calculate the ionizing diffuse radiation emitted by the sn - driven bubble .
we first calculate the development of a bubble driven by collective winds from @xmath78 massive stars with @xmath217 .
characteristic structures of wind bubbles are developed initially : an inward facing termination shock , a conduction front near the contact discontinuity , and a forward expanding shock .
the outermost shock is not strong and becomes a compression wave , because the wind ram pressure decreases as the bubble expands . at the end of the main sequence ( @xmath218 years ) ,
the outer boundary of the bubble shell ( radius of the compression wave ) is @xmath219 , and the inner boundary is @xmath220 for @xmath115 .
so the shell is rather thick and the enhancement factor of the shell density is small , @xmath221 .
the mass of the bubble shell is @xmath222 .
the evolution of a supernova remnant inside a wind bubble is much more complex than that in a uniform medium .
it depends on the structure and size of the preexisting bubble as well as the total explosion energy , @xmath223 ergs .
the forward shock travels quickly through the cavity and runs into the bubble shell , resulting in a transmitted forward shock . on the other hand ,
the reverse shock in the ejecta is reflected at the center and becomes a compression wave that travels outward .
when this wave runs into the shell , the shell is heated for the second time , followed by the second period of active radiative cooling .
the wave bounces off the shell and moves inward with a much reduced amplitude . at the termination time of the simulations ( @xmath150 years ) ,
the outer shock radius is @xmath224 pc for a single sn inside a bubble blown by one massive star ( snf01 model ) , while @xmath225 pc for a single sn inside a bubble blown by 10 massive stars ( snf10a model ) . in comparison ,
the shock radius for a single sn in a uniform ism of the same density at the same time is about 60 pc .
the radii of the hot cavity and the bubble shell can be bigger up to 50 % and the cavity temperature is higher by 2 - 3 times for sn explosions in a wind bubble considered here , compared to the sn explosions inside a uniform ism . for multiple sn explosions inside a bubble blown by 10 massive stars , the size and mass of the shell increase rather weakly with the total explosion energy as @xmath2 and a mass of @xmath3 . for @xmath226 , for example
, the sn - driven shell would have typically @xmath227 pc and @xmath228 for the average ism .
kinetic energy of the sn is transfered to thermal energy via various shocks .
about 20 - 60 % of the explosion energy is radiated away after the sn shock hits the wind bubble shell , and then 20 - 40 % is lost after the compression wave crushes into the shell later .
although the final energy budget at the termination time of our simulations depends on both the wind bubble structure and @xmath0 , in most cases with solar metalicity , @xmath229 , @xmath230 and @xmath231 .
ionizing photons of diffuse radiation account for up to @xmath232 , while the total number of ionizing photons scales as @xmath233 . for the cases with @xmath50 , the radiative loss
is reduced a little bit , @xmath234 , but the fraction of ionizing radiation is reduced a factor of 2 - 3 , compared to the models with solar metalicity .
so a larger fraction of energy is lost via non - ionizing radiation at metal - poor environments .
the authors would like to thank j. kim , d. ryu and an anonymous referee for helpful comments on the paper .
this work was supported by the korea research foundation grant funded by korea government ( moehrd , basic research promotion fund ) ( r04 - 2006 - 000 - 100590 ) and by kosef through the astrophysical research center for the structure and evolution of cosmos ( arcsec ) .
the work of hk was also supported in part by korea foundation for international cooperation of science & technology ( kicos ) through the cavendish - kaist research cooperation center .
years in wf01 model calculated with @xmath88 3000 ( solid lines ) , 6000 ( dotted ) , and 12000 ( dashed ) radial grid zones .
the density is given in units of @xmath236 , and the pressure in units of @xmath237 .
, title="fig : " ] -7.0 cm years in wf01 ( solid lines ) , wf10 ( dotted ) , wf30 ( dashed ) , and wf50 ( long dashed ) models . _ lower panels : _ gas density and temperature profile at @xmath97 yr for wf01 ( solid line ) , wl01 ( dotted ) , wh01 ( dashed ) , and wz01 ( long dashed ) models .
the density is given in units of @xmath238 . ]
( solid lines ) , @xmath239 years ( dotted ) , @xmath240 years ( dashed ) , @xmath241 years ( long dashed ) , and @xmath242 years ( dot - dashed ) .
the normalization constants are : @xmath243 , @xmath244 , and @xmath245 . ]
_ lower panels : _ supernova explosions inside a uniform ism of @xmath115 .
the explosion energy is @xmath248 ergs , where @xmath136 = 1 ( solid lines ) , 3 ( dotted ) , 5 ( dashed ) , and 10 ( long dashed ) . here @xmath243 . ]
years ( red ) , @xmath249 years ( blue ) , and @xmath4 years ( green ) .
_ lower panel : _ integrated spectrum of diffuse radiation emitted by the bubble in snf01 model with @xmath250 ( red ) and snz01 with @xmath251 ( blue ) models at @xmath97 years . ] ( left panels ) and snz10 models with @xmath251 ( right panels )
the explosion energy is @xmath248 ergs , where @xmath136 = 1 ( solid lines ) , 3 ( dotted ) , 5 ( dashed ) , and 10 ( long dashed ) . here @xmath243 . ] | we study the evolution of multiple supernova ( sn ) explosions inside a pre - exiting cavity blown by winds from massive progenitor stars .
hydrodynamic simulations in one - dimensional spherical geometry , including radiative cooling and thermal conduction , are carried out to follow first the development of the wind - blown bubble during the main sequence and then the evolution of the sn - driven bubble .
we find the size and mass of the sn - driven bubble shell depend on the structure of the pre - existing wind bubble as well as the sn explosion energy @xmath0 ( @xmath1 ) .
the hot cavity inside the bubble is 2 - 3 times bigger in volume and hotter than that of a bubble created by sne exploded in a uniform interstellar medium ( ism ) . for an association with 10 massive stars in the average ism
, the sn - driven shell has an outer radius of @xmath2 and a mass of @xmath3 at @xmath4 years after the explosion . by that time
most of the explosion energy is lost via radiative cooling , while @xmath5 10% remains as kinetic energy and @xmath610% as thermal energy .
we also calculate the total integrated spectrum of diffuse radiation emitted by the shock - heated gas of the sn bubble .
total number of h lymann - limit photons scales roughly as @xmath7 and those photons carry away 20 - 55 % of the explosion energy . for the models with 0.1 solar metalicity ,
the radiative energy loss is smaller and the fraction of non - ionizing photons is larger , compared to those with solar metalicity .
we conclude the photoionization / heating by diffuse radiation is the most dominant form of feedback from sn explosions into the surrounding medium .
stars : early type , stars : winds , supernova remnants , galaxies : ism 97.10.me , 98.38.mz , 98.20.af |
turbomachinery bladed disks operate in severe environments in terms of aerodynamic loads and may experience high vibratory stresses due to resonance or flutter . their ability to withstand these stress levels , which may cause high cycle fatigue , mainly depends on the external inputs of damping . in bladed disk assemblies , friction in interfaces is the most widely used source of external damping ; this includes shroud contact in blades or platform dampers .
these devices were widely studied in the past and numerous methods of design and analysis are found in the literature .
the interest reader can find examples based on lumped parameter or equivalent models in works of @xcite , @xcite , @xcite or @xcite and , more recently , extended to more complex structures in works by @xcite , @xcite or @xcite among others .
the contact interface could be one dimensional , two dimensional or three dimensional ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . beside ,
numerous experimental works were presented both in contact kinematic description ( see @xcite for example ) and dynamic response prediction ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
however , in blisks ( integrally bladed disks where disk and blades are a single piece ) , inherent damping is very low and mechanical joints no longer exist leading to the disappearance of nearly every sources of energy dissipation .
consequently , new methods and devices need to be found for the damping of blisks . among others ,
friction damping using a ring is a solution which is found , experimentally , to be efficient on single piece rotating structures ( such as labyrinth seals for example ) .
these rings are located in dedicated grooves underside wheels and contact is permanently maintained by centrifugal load due to engine rotation ; friction and slipping occur during the differential motion of the two bodies in contact . as opposed to under - platform dampers in bladed disk assemblies , these devices were seldom studied ; @xcite have presented methods of analysis and optimal design of split ring dampers for seals were the motion is studied using a quasi - static beam - like description ( strength of materials ) . this quasi - static description lead to a good understanding of the phenomenon and give some qualitative results .
however , a nonlinear dynamics analysis is required to raise satisfying design rules . in this paper , methods for predicting nonlinear steady - state response of blisks with friction ring dampers
are presented along with some qualitative results based on a representative lumped parameter modelling .
the dynamical analysis is performed using the multi - harmonic balance method ( _ hbm _ ) .
this method is found to be efficient and was widely used , in the past , to study friction dampers in bladed disks assemblies or shroud contacts .
some examples have been presented by @xcite , @xcite and @xcite for mono - harmonic vibrations and by @xcite , @xcite , @xcite , and @xcite for multi - harmonic vibrations .
results from numerical simulations ( using this nonlinear analysis ) are then presented and described .
several parametric studies are presented in terms damper characteristics , contact characteristics , excitation configuration and mistuning are presented and some design guidelines are raised .
lumped parameter models have been widely used to study friction . beside from being
simpler to study and compute ( particularly for parametric studies ) , they can provide some interesting and valuable qualitative results if correctly designed . in turbomachinery applications ,
several models have been proposed . among others , we can mention @xcite or @xcite for friction damping applications and @xcite and @xcite for mistuning applications . in this paper , we present a new lumped parameter model .
it consists in an elementary sector of a blisk and its associated friction ring element and it is represented in figure [ fig : fig01 ] . as this system is rotationally periodic ( or cyclic ) , its modes are such that the deflection of a given elementary sector ( described in figure [ fig : fig01 ] ) is the same as , but with a constant phase difference from , the preceding ( or following ) sector
. if @xmath5 is the number of elementary sectors of the structure , the phase angle is @xmath6 where @xmath7 is called the _ nodal diameter number_. there are @xmath5 possible values of @xmath7 ( from @xmath8 to @xmath9 ) but only @xmath10 values ( from @xmath8 to @xmath11 ) , if @xmath5 is even or @xmath12 values ( from @xmath8 to @xmath13 ) , if @xmath5 is odd , represent different mode shapes ; the others being orthogonal to them with the same eigenfrequency .
consequently , except for @xmath14 or @xmath15 if @xmath5 is even , the natural frequencies are repeated .
further information on this subject can be found in the literature ( e.g. * ? ? ?
* ; * ? ? ?
as an example , figure [ fig : fig02 ] displays the natural frequencies versus nodal diameter number of the ( blisk ) lumped parameter model with @xmath16 sectors .
the blisk lumped - parameter model was designed so that it stands as much as possible for a real blisk , particularly in terms of coupling between the blades and the disk .
there are three families of modes and regions of various modal density .
this modal density is directly linked to the coupling between the blades and the disk in the particular mode .
a general rule , which for that matter applies to the present example , is that the blade / disk coupling can be important is low frequency / low nodal diameter and high frequency / high nodal diameter regions whereas in low frequency / high nodal diameter or high frequency / low nodal diameter regions this coupling is generally weak . for example , in the two first families the coupling between the blades and the disk is weak for high nodal diameters whereas in small nodal diameters the coupling is stronger .
also note an important coupling appearing in the @xmath17 and @xmath18 modes with nodal diameter number 2 . recalling that , in operating conditions , the external forcing generally acts on the blades this coupling parameter is an important factor regarding the efficiency of the friction ring damping since it determines the disk _ participation _ in the global motion .
the energy dissipation due to friction occurs when the relative displacement between the disk and the ring is enough ; this is achieved when the blisk is excited in areas of strong coupling where both the blades and the disk move .
the numerical values of the lumped - parameter are shown in table [ tab : tab01 ] .
.numerical values of the model [ cols="^,^,^,^,^,^,^,^",options="header " , ] [ tab : tab01 ] finally , concerning the ring , its mass and stiffness are chosen so that its lumped - parameter model represents a circular beam in extension motion and its stiffness @xmath19 represents the longitudinal rigidity of the ring .
an examination of the local behaviour of the friction area can help understanding the friction dissipation mechanism .
with reference to figure [ fig : fig03a ] , both the disk and the ring experience bending motion ( bending stress @xmath20 ) and are held in contact ( slipping or not ) due to the centrifugal load @xmath21 .
the friction forces in the interface ( @xmath22 ) between the two elements will generate tension or compression stresses ( @xmath23 ) mainly in the ring which ensure the continuity of stresses at the interface as shown in figure [ fig : fig03a ] . as a consequence the friction load
is directly linked to the tension stresses and , assuming a linear elastic deformation , to the axial rigidity of the ring .
figure [ fig : fig03b ] shows a physical model of the friction ring in its groove . in order to be assembled in these groove , underside the wheel of the disk , the ring needs to be split ; the influence of this split on the dynamics of the system will be discussed later .
the system of equations of motion can be represented as : @xmath24 where @xmath25^t$ ] is the physical displacements vector which gathers displacements of the disk , the blades and the ring degrees - of - freedom .
the matrices @xmath26 , @xmath27 and @xmath28 are the structural matrices of mass , damping and stiffness respectively .
the vector @xmath29 is an external distribution of rotating forces of engine order type , so that , considering a bladed disk with n sector , the @xmath30 sector s forcing term is : @xmath31 } , \label{eq : eoforce}\ ] ] where @xmath32 is engine order .
the vector of nonlinear forces @xmath33 due to friction is expressed , according to the coulomb law , as : @xmath34 where @xmath35 is the relative displacements vector between the ring and the disk , @xmath36 is the constant friction coefficient , and @xmath21 the normal contact load , which can be expressed for rotating structures as : @xmath37 where @xmath38 is the ring s mass , @xmath39 is the distance from the axe of rotation and the ring and @xmath40 is the engine rotation speed . due to the high contact load resulting from the engine rotation ,
is assumed that the ring and blisk remain in contact at all times .
the aim of the nonlinear analysis performed here is to calculate the steady - state response of the system under a periodic external forcing . among the variety of method available to achieve this goal
, the harmonic balance method @xcite presents several advantages in terms of efficiency and accuracy with respect to its computational costs .
the idea is to express the time - dependent variables of the nonlinear equation ( [ eq : mvt ] ) in terms of fourier series ( truncated in practice ) ; the response @xmath41 is then : @xmath42 the new unknowns of the problem are the fourier components @xmath43 and @xmath44 which are balanced ( using a galerkin procedure ) and the equation of motion can be rewritten using this formalism for each terms of equation ( [ eq : mvt ] ) in the frequency domain as : @xmath45 with : @xmath46^t,\\ { \mathbf{f}}_{nl}&=\left[{\mathbf{f}}_{nl}^0,{\mathbf{f}}_{nl}^{1,c},{\mathbf{f}}_{nl}^{1,s},\dots,{\mathbf{f}}_{nl}^{n_h , s}\right]^t,\\ { \mathbf{f}}_{ext}&=\left[{\mathbf{f}}_{ext}^0,{\mathbf{f}}_{ext}^{1,c},{\mathbf{f}}_{ext}^{1,s},\dots,{\mathbf{f}}_{ext}^{n_h , s}\right]^t .
\end{aligned}\ ] ] the matrix @xmath47 is block - diagonal , @xmath48 , and : @xmath49 the problem ( [ eq : mvthbm ] ) is nonlinear and is usually solved iteratively using a newton - like method .
one of the issue of the _ hbm _ is that the fourier components of the nonlinear forces @xmath50 can not , in most cases , be derived straightforward . in particular , in problems involving friction , the nonlinear forces are only known ( using equation ( [ eq : forcecoulomb ] ) ) in the time domain and the expansion in the frequency domain is not obvious .
when few harmonics ( no more than @xmath51 in practice ) are retained an analytical derivation of these frequency domain components is possible ( e.g. * ? ? ?
* ; * ? ? ?
* ) . to overcome this problem when more harmonics are considered , the use of an _ alternating frequency time _ method ( e.g. *
* ) can be a solution .
this method will be detailed in the next section .
the aim is , in the newton procedure , to derive the fourier components of the nonlinear forces as a function of the displacements fourier components . in computational applications ,
the idea of the _ aft _ method is to use discrete fourier transformation ( dft ) to derive the fourier components of the nonlinear forces for given displacements in the frequency domain .
let s start from @xmath43 and @xmath44 , the harmonic components of the response , predicted by a given iteration of the newton - like method . using , an idft ( _ inverse discrete fourier transform _ ) procedure , one can express the associated displacements @xmath41 and velocities @xmath52 in the time - domain . then using a nonlinear operator the nonlinear force in time - domain
is derived ( here this step is achieved using equation ( [ eq : forcecoulomb ] ) as described in the next section ) . finally , using a dft ( _ discrete fourier transform _ )
algorithm one expresses the harmonic components of the nonlinear force in the frequency - domain .
the following diagram illustrates this procedure : @xmath53 at a given iteration of the newton scheme and during the _ aft _ procedure , the friction forces need to calculated , in the time domain , in accordance with the velocity and displacement given ( by the newton method ) .
this is generally done using the coulomb law or , more usually , using a regularized coulomb law and several strategies can be used .
some authors have proposed a regularized expression of the sign function ( e.g. * ? ? ? * ) or arctangent or hyperbolic tangent function ( e.g. * ? ? ?
* ) , some others have used the theory of distributions ( e.g. * ? ? ?
* ) , and some have used an iterative procedure in the time domain ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
for the latter strategy , it is convenient to add a tangent stiffness to each friction element ; doing so , the coulomb law can be used straightforward to compute at each time step the equilibrium of the contact point . in previous studies on under - platform dampers ,
this stiffness was included in the damper model .
however , in general contact problem , this tangent stiffness appears as an equivalent penalty stiffness and can then account for the elasto - plastic shear deformations of the contact asperities of the bodies in contact .
figure [ fig : fig04 ] shows a friction element defined in relative displacements , with a tangent stiffness : * @xmath54 is the relative between two points in contact , @xmath55 in our example .
* @xmath56 is the relative displacement of the contact point .
the nonlinear forces vector given by equation ( [ eq : forcecoulomb ] ) becomes : @xmath57 in the time domain of the _ aft _ procedure , velocities and displacements are given in one period of the motion .
these are entries for the computation of the associated nonlinear forces in one period of the motion .
this computation has to be performed time - iteratively ( like a numerical integration ) because of the history dependency of the friction law .
for example , given the relative displacements and velocities at time @xmath58 , @xmath59 and @xmath60 and the displacement of the contact point @xmath61 , the friction force can be predicted ( using equation ( [ eq : force_frottement_tempo ] ) ) for time @xmath58 assuming a sticking state ( @xmath62 ) for the contact point : @xmath63 the coulomb law is then used to correct the friction force and the displacement of the contact point : @xmath64 using this iterative scheme , the periodic friction force can be derived .
in this section , several aspects of friction ring damping are investigated through numerical simulations and discussed .
all nonlinear numerical simulations were performed using the multi - harmonic balance procedure and a third order lagrange predictor was used for the continuation of the frequency responses .
a @xmath65 order lagrange predictor , see @xcite for example , requires the last @xmath66 points to predict the next one @xmath67 and an interpolation is build on the arc - length @xmath68 ( which is here the excitation frequency ) as : @xmath69 \scriptstyle j=0\\[-0.4 cm ] \scriptstyle j\neq 1 \end{array}}^d \dfrac{s^{(p+1)}-s^{(p - d+j)}}{s^{(p - d+i)}-s^{(p - d+j)}}.\ ] ] the step @xmath70 is adjusted _ a posteriori _ as a function of the ( newton - like ) corrector s performance ( number of iterations ) at the previous step .
all frequency response plots represent the blade response ( degree - of - freedom @xmath71 in figure [ fig : fig01 ] ) , with , in dashed line , the linear response ( without ring ) and in solid lines , the nonlinear responses ( with friction ring ) .
parametric studies on the type of excitation ( or blade / disk coupling as mentioned earlier ) , on contact parameters , damper characteristics are presented and the efficiency of this damping strategy is discussed . then the influence of ring s split discussed . and finally , the impact mistuning and friction ring damping is investigated . in this first part , the influences of the ring damper mass and of the excitation configuration are investigated .
the rotation speed is kept constant is these simulations ; however , note that as the ring s mass changes , the contact normal load also change with respect to equation ( [ eq : normalload ] ) . [ [ strongly - bladedisk - coupling ] ] strongly blade / disk coupling + + + + + + + + + + + + + + + + + + + + + + + + + + + + with reference to the frequency / nodal diameter map of figure [ fig : fig02 ] , we first focus on a small engine order excitation ( @xmath72 for example in equation ( [ eq : eoforce ] ) ) with frequency range around the first modal family ( similar behaviour would appears in other modal family however ) .
several frequency responses are plotted in figure [ fig : fig05 ] for different damper mass . from these response
, we can first notice that the damping is efficient in this configuration when the ring is sufficiently thick and that an optimal value appears ( @xmath73 ) . at this optimal value ,
the reduction of vibration amplitude is nearly 4 .
also note the progressive frequency shift of the resonance peaks as the ring s mass increase , this is due to two facts combined .
first , considering the associated linear problem , as the ring s thickness increases the global system becomes stiffer which account for the frequency shift to the higher frequency . and
second , the nonlinear effect of friction has usually a softening effect on system s dynamic and , with respect to figure [ fig : fig05 ] , as the ring s mass decreases the contact normal load also decreases ( assuming a fixed rotation speed ) and more and more slipping occurs which account for the frequency shift to the lower frequency as the damper s mass decreases . [ [ weakly - bladedisk - coupling ] ] weakly blade / disk coupling + + + + + + + + + + + + + + + + + + + + + + + + + + a weaker blade / disk coupling can be found in higher nodal diameter in figure [ fig : fig02 ] . as an example , let s consider a 6 engine order excitation on the second modal family and the associated frequency responses plotted in figure [ fig : fig06 ] . in this example
, the friction ring damper is clearly less efficient ; even if an optimal mass appears ( around 5 g ) , the reduction of vibration s level is not as important as in strong coupling configuration . in effect , in the configuration , there is not enough motion in the disk as the blades are excited and as a consequence no relative motion between the disk and the ring can occur . in this section ,
we focus on the influence of the normal contact load ( which is , in our application , related to the engine rotation speed ) on the friction ring damping efficiency .
to do so , the damper mass is fixed as well as the excitation parameters ( engine order 2 ) and several simulations were performed for different values of the engine rotation speed . in figure [ fig : fig07 ] , the results are depicted and the first remarks is that the existence of an optimal rotation speed / normal load is clear around 3000 rpm .
moreover , one can see that that in a quite large vicinity of this optimal value ( from 2000 to 6000 rpm ) , the friction ring damping is still efficient .
this is a key feature since a turbomachinery often operates at different rotation speeds ( take - off , cruise , landing , ) . as a consequence , the rings
can , for example , be optimally designed for the cruise speed and be also efficient at take - off or landing speeds .
as mentioned earlier , in order to be assembled the ring need to be split . however , considering the model described in figure [ fig : fig01 ] , a split in the ring would necessary spread over one elementary sector and generate an exaggerated gap compared to the technological reality . as a consequence
, we built a refined model , presented in figure [ fig : fig08 ] , in which the disk mass , @xmath74 , is refined into three masses , @xmath75 .
this allows the split to spread only over a portion of one sector .
the influence of the ring s split was then investigated by comparing the dynamic response of the system with or without a split .
first , the ring is assumed to be fully stuck ( no sliding occurs ) and the system is then linear .
the frequency response plots resulting from these simulations are displayed in figure [ fig : iflucoup ] .
we focused on the first resonance peak of the 2 engine order excitation and comparing the blade s resonant responses with ( figure [ fig : fig09b ] ) or without ( figure [ fig : fig09a ] ) ring s split , it appears that each blades have a different response and that the maximum level is more important when the ring is split .
this symmetry breaking , which corresponds to the separation of the initial twin modes ( two - fold degeneracy ) , is similar to the consequence of mistuning and can lead to localization phenomena ( see @xcite for example ) .
when the nonlinear effect are introduced ( figure [ fig : influcoupnl ] ) , the same phenomena of symmetry breaking and increasing of resonant responses appear . however , the response amplification is limited by the nonlinear effect and then smaller than in the linear case .
these results show the influence of the split in the ring damper on the global dynamics of a bladed disks
. however , these observations have to be put in perspective with respect to the influence of the inherent mistuning ( discussed next ) of a real bladed disks which may generally be more important . to conclude on the numerical results , the influence of mistuning was investigated .
mistuning refers to small variations in the properties of a system originally designed to be cyclically symmetric .
these variations are mainly due to manufacturing tolerances or material inhomogeneity and are inevitable .
mistuning can lead to significant changes in dynamical behaviour with respect to the tuned case as presented by @xcite ; vibratory amplitudes in forced response are increased and impact on high cycle fatigue can be considerable .
these effects are even more important in integrally bladed disks since their `` natural '' damping is far smaller than for bladed disks assemblies . the influence of mistuning on friction damping has been studied earlier by @xcite or @xcite and the performance of the damping technology discussed here as applied to blisks may be influenced by this phenomenon . considering the lumped parameter model of figure [ fig : fig01 ] , mistuning
was introduced as random perturbations in the blades stiffness @xmath76 and @xmath77 as it is often the case since the disk is assumed to be symmetric .
these perturbations lead to changes in the natural frequencies of the blades ( cantilevered ) .
a frequency response is shown in figure [ fig : fig11 ] , the mistuning standard deviation is about 1.2% of the cantilevered blade s nominal natural frequency .
the resonant responses of the blades are displayed for the tuned and the mistuned case with or without ring .
first , we can notice the typical feature of mistuned responses which is the break of symmetry which leads to a separation of the resonance peak ( with or without rim damping ) .
then , the vibratory amplitudes are increased in the mistuned case with respect to the tuned case ; however , the damping is still efficient in presence of mistuning .
the results of these numerical simulations clearly highlight the possibilities and limits of the friction ring damping strategy .
in essence , the dissipation mechanism of friction is efficient only if some relative motion between the two bodies ( the disk and the ring ) in contact occurs . as opposed to under - platform dampers , the rings are not directly in contact with the blades to which the external forcing applies .
as a consequence , the coupling between the blades and the disk which conditions the energy transfer from the blades to the disk is an essential criterion .
it has been demonstrated that when the coupling is weak ( that is when little energy get transferred to the disk ) , the ring damping is less efficient . on the other hand ,
when this coupling is sufficiently strong the rings are efficient .
finally , it was shown that the contact parameters ( normal loads or friction coefficient ) are also significant in the friction damping performance .
the qualitative results given by these simulations are quite similar to those observed with under - platform damper ; the alternation between slipping and sticking states results in frequency shift which offsets the resonant frequency from its initial value ( _ i.e. _ non damped ) , and an optimum can be reached for a given excitation configuration in the ring s mass .
a strategy of damping for integrally bladed disks ( blisks ) using a friction split ring was investigated .
a procedure of analysis using a multi - harmonic balance method and an _ aft _ method was described and applied for calculating the steady - state nonlinear response of a system under periodic excitation .
simulations on the lumped - parameter model presented lead to qualitative results in terms of energy dissipation with respect to several configurations .
some guidelines on the use of this technology , its limits and performances , were highlighted .
we show that the coupling ratio between the disk and the blades was a critical parameter of the damping power of such rings .
thanks go to snecma for their technical and financial supports .
guillen , j. , pierre , c. , 1999 .
an efficient , hybrid , frequency - time domain method for the dynamics of large - scale dry - friction damped structural systems . in : iutam symposium on unilateral multibody contacts .
wei , s. t. , pierre , c. , 1988 .
localization phenomena in mistuned assemblies with cyclic symmetry part 1 : free vibrations .
journal of vibration , acoustics , stress , and reliability in design 110 , 429437 .
wei , s. t. , pierre , c. , 1988 .
localization phenomena in mistuned assemblies with cyclic symmetry part 2 : forced vibrations .
journal of vibration , acoustics , stress , and reliability in design 110 , 439449 .
yang , b. , menq , c. , 1998 .
characterization of contact kinematics and application to the design of wedge dampers in turbomachinery blading : part 2 prediction of forced response and experimental verification .
journal of engineering for gas turbines and power 120 , 418423 . | a damping strategy for blisks ( integrally bladed disks ) of turbomachinery involving a friction ring is investigated .
these rings , located in grooves underside the wheel of the blisks , are held in contact by centrifugal loads and the energy is dissipated when relative motions between the ring and the disk occur .
a representative lumped parameter model of the system is introduced and the steady - state nonlinear response is derived using a multi - harmonic balance method combined with an _ aft _ procedure where the friction force is calculated in the time domain .
numerical simulations are presented for several damper characteristics and several excitation configurations . from these results ,
the performance of this damping strategy is discussed and some design guidelines are given .
journal home page : http://www.sciencedirect.com/science/journal/09977538 + qualitative analysis of forced response of blisks with friction ring dampers + european journal of mechanics - a / solids , volume 26 , issue 4 , july - august 2007 , pages 676 - 687 + d. laxalde , f. thouverez , j .- j .
sinou and j .- p .
lombard * qualitative analysis of forced response of blisks with friction ring dampers * + d. laxalde@xmath0 , f. thouverez@xmath1 , j .- j .
sinou@xmath1 , j
.- p .
lombard@xmath2 @xmath3 laboratoire de tribologie et dynamique des systmes ( umr cnrs 5513 ) + cole centrale de lyon , 36 avenue guy de collongue , 69134 ecully cedex , france + @xmath4 snecma
safran group , 77550 moissy - cramayel , france [ [ keywords ] ] keywords : + + + + + + + + + bladed - disk , friction damping , nonlinear dynamics , harmonic balance method |
a symplectic realization of a poisson manifold @xmath1 is a poisson map from a symplectic manifold @xmath2 to @xmath1 , which is also a surjective submersion .
symplectic realizations are very natural objects in poisson geometry from the point of view of integration and quantization theory of poisson manifolds .
namely , following the standard `` quantization dictionary '' that associates a quantum algebra @xmath3 of observables with a poisson manifold @xmath1 ( deformation quantization ) and a hilbert space @xmath4 of states with a symplectic manifold @xmath2 ( geometric quantization ) , a symplectic realization should quantize to a representation of @xmath3 on @xmath4 , which is the object encoding the symmetries at the quantum level ( see @xcite ) . from the perspective of global integration of poisson manifolds ,
symplectic realizations naturally arise as source maps of symplectic groupoids @xmath5 , which are the objects that integrate poisson manifolds .
symplectic realizations contain a lot of information about integrability .
for instance , crainic and fernandes showed in @xcite that a poisson manifold is globally integrable if and only if it admits a complete symplectic realization .
however , in the global integration case , the symplectic realization do not carry all the information about the integrating symplectic groupoid ( some extra data is needed to define a global inverse map for instance ) .
the situation is different for the local / formal integration of poisson manifolds by local / formal symplectic groupoids .
recall that a local symplectic groupoid is , roughly , the structure obtained by restricting a global symplectic groupoid to a neighborhood of its unit space , while a formal symplectic groupoid is the structure obtained by taking @xmath6-jets of the symplectic groupoid structure maps ( or their pullbacks as in @xcite ) at the unit space . in the local / formal case , all the structure maps of the local / formal symplectic groupoid , including the inverse map , can be recovered from the source map alone ( @xcite ) , i.e. , from a local / formal symplectic realization .
another simplification in this case is that the domain of the symplectic realization can always be taken to be the cotangent bundle of the poisson manifold , or , more precisely , a local / formal neighborhood of its zero section . given a poisson manifold @xmath1 , the problem of constructing a local / formal symplectic realization can essentially be tackled in two ways , both of which start by considering a deformation @xmath7 of the zero poisson structure on @xmath8 by a parameter @xmath9 . for the zero poisson structure ,
the canonical bundle projection @xmath10 is a symplectic realization , where the symplectic form on the cotangent bundle is the canonical one @xmath11 .
the first way is to try to deform @xmath11 into another symplectic form @xmath12 such that the bundle projection @xmath13 remains a symplectic realization from @xmath14 ( or from a neighborhood of its zero section ) to @xmath15 .
this approach was the original one of weinstein in @xcite , who gave an integral formula for @xmath12 and proved the result in the case @xmath16 .
the result for a general manifold has been proven recently by crainic and marcut in @xcite . the second way is the approach of karasev in @xcite that keeps the canonical symplectic form @xmath11 fixed and deforms the bundle projection into a symplectic realization @xmath17 . the formal expansion of @xmath18 in the case @xmath16 has been shown in @xcite to coincide with the formal symplectic realization that one can extract ( see @xcite ) from the tree - level part of the kontsevich star - product given in @xcite . in this paper , we show , in the case @xmath16 , the explicit relationship between these two constructions ( proposition [ prop : we - kar ] and theorem [ thm : realizations ] ) , and we give a very explicit formula for the formal symplectic realization in terms of topological rooted trees ( theorem [ thm : formula ] and proposition [ prop : weights ] ) .
in section [ sec : general - results ] , we review the construction of @xmath12 as considered in @xcite , and we give a canonical way to obtain a karasev - like realization @xmath18 of any poisson manifold out of the deformed symplectic form @xmath12 by putting it in global darboux form . in section [ sec : comparison ] ,
we review the formal symplectic realization @xmath19 where @xmath20 is the generating function ( extracted from the kontsevich star - product ) of the formal symplectic groupoid constructed in @xcite that integrates the poisson manifold @xmath21 .
we prove that it coincides with the weinstein realization for @xmath22 when suitably put in darboux form and that the coefficients admit a recursive definition ( theorem [ thm : realizations ] ) . in section [ sec : formula ] , we show ( theorem [ thm : formula ] ) that @xmath23 is given by the formula @xmath24}\e^{|t|}\left(\frac{w_{t}}{|\sym(t)|}\right)d_{t}^{i}l_{v},\ ] ] where @xmath25 $ ] is the set of topological rooted trees , @xmath26 is the kontsevich weight of the kontsevich graph associated to the topological rooted tree @xmath27 as in figure [ fig : konttree ] , and @xmath28 is the elementary differential ( see definition [ def : elemdiff ] ) associated with @xmath27 and the vector field @xmath29 .
building on kathotia s results in @xcite , we show in proposition [ prop : weights ] that these weights are easily computable : @xmath30 , where @xmath31 is a recursive formula @xmath32 associated with the rooted tree @xmath33_{\bullet}$ ] made out of the subtrees @xmath34 by grafting their root to a new additional root , and where @xmath35 when @xmath27 is the rooted tree with only one vertex .
we thank alberto s. cattaneo for useful remarks and suggestions .
b.d . acknowledges partial support fapesp grants 2010/15069 - 8 and 2010/19365 - 0 , as well as the hospitality of the uc berkeley mathematics department and so paulo university icmc . a.c .
also thanks icmc - usp and fapesp for support and hospitality during the elaboration of this paper .
in this section , we review the symplectic realization @xmath36 considered originally by weinstein in @xcite and , more recently , by crainic and marcut in @xcite . we explain how it can be constructed out of a poisson spray , and we show how to put this realization in `` global darboux form '' to obtain a karasev - like realization @xmath37 .
we end up by further analyzing the case @xmath16 , for which we also consider a formal version of the constructions .
let @xmath1 be a poisson manifold and @xmath38 its cotangent bundle
. throughout the paper @xmath39 will denote a poisson spray for @xmath40 , namely , a vector field on @xmath41 satisfying : 1 .
@xmath42 for all @xmath43 2 .
@xmath44 , for @xmath45 being the diffeomorphism obtained by fiberwise multiplication by @xmath46 notice that @xmath47 is thus a spray for the re - scaled poisson structure @xmath48 .
it follows from the definition that the flow @xmath49 of @xmath50 fixes the zero section and , thus , that there exists an open neighborhood @xmath51 of the zero section such that @xmath49 is defined for @xmath52 $ ] satisfying @xmath53 and such that there is a neighborhood of the zero section on which the flow is defined on @xmath54 $ ] . ] .
[ ex : spray from nabla ] a linear connection @xmath55 on @xmath38 defines a poisson spray by setting , for @xmath43 , @xmath56 for each @xmath57 $ ] , we choose an open neighborhood @xmath58 of the zero section such that @xmath49 is defined for @xmath59 $ ] .
we can take @xmath60 for all @xmath57 $ ] .
following @xcite , we consider the following differential forms on @xmath61 @xmath62 where @xmath63 is the liouville one form on @xmath41 and @xmath64 the associated symplectic form . ] . as explained in @xcite ,
when evaluated at points of the zero section @xmath65 @xmath66 is non degenerate for all @xmath67 , where @xmath68 is the natural projection .
it thus follows that we can choose possibly smaller neighborhoods @xmath69 where this 2-form is non - degenerate .
then , @xmath70 is a ( exact ) symplectic form on this @xmath71 .
[ thm : omega v realization ] the bundle projection @xmath72 is a symplectic realization .
this was proven in @xcite .
consider the path construction of the ( local ) symplectic groupoid @xmath73 integrating @xmath74 in terms of ( small ) cotangent paths modulo cotangent homotopies as in @xcite and @xcite .
for each initial condition , close enough to the zero section in @xmath41 , the flow of @xmath75 produces a particular cotangent path @xmath54\to t^{*}m$ ] , defining in this way an exponential map @xmath76 .
the symplectic form above arises as @xmath77 .
the content of the above theorem can then be understood as follows : the symplectomorphism @xmath78 takes the source map of @xmath79 , which is a symplectic realization ( as for any symplectic groupoid ) , to the projection @xmath13 .
[ lem : change nabla]_(change _ _ of spray)_. let @xmath1 be a poisson manifold and consider two poisson sprays @xmath80 and @xmath81 . then , uniqueness of symplectic realizations ( see @xcite ) together with theorem [ thm : omega v realization ] imply the existence of a symplectomorphism @xmath82 such that @xmath83@xmath84 .
notice that the family of symplectic forms @xmath70 are exact for all @xmath67 and that @xmath85 .
one can thus try to apply a moser - type argument to find a flow that transforms this family back into @xmath11 .
[ def : psi_v global darboux ] we say that a symplectic realization @xmath86 is in _ darboux form _ if @xmath87 is an open neighborhood of the zero section and @xmath88 .
given a realization @xmath89 , a _ global darboux frame _ for it consists of a diffeomorphism @xmath90 such that @xmath91 .
when a global darboux frame for @xmath92 exists , the composition of poisson maps @xmath93 yields an _ induced realization in darboux form_. in @xcite , a symplectic realization is called _ strict _ when the realization map admits a global lagrangian section . by the lagrangian embedding theorem
, strict realizations always admit global darboux frames . for completeness ,
we now show that , for each @xmath67 , there is a natural global darboux frame for @xmath70 attached to the given data @xmath94 .
[ cor : alpha for any m]let @xmath1 be a poisson manifold and @xmath50 a poisson spray .
consider the ( @xmath67-dependent ) vector field @xmath95 on @xmath96 defined by @xmath97 then , there exists an open neighborhood @xmath98 of the zero section such that the flow @xmath99 of @xmath95 is defined on @xmath100 for @xmath57 $ ] and @xmath101 for all @xmath57 $ ] .
notice that the forms @xmath70 and @xmath102 are well defined on @xmath96 for all @xmath57 $ ] and that they define a smooth @xmath67-family of differential forms on @xmath96 .
the vector field @xmath95 defined in eq .
satisfies @xmath103 for all @xmath67 since , when evaluated at points of the zero section , @xmath104 .
then , there exists an open neighborhood @xmath105 on which the flow @xmath99 is defined for @xmath57 $ ] .
finally , a standard computation shows that @xmath106 the statement then follows from the initial condition @xmath85 . as a consequence
, there exists a natural family of induced symplectic realizations in darboux form @xmath107 for @xmath57 $ ] associated to the spray @xmath50 . in the particular case
@xmath16 below , we shall see that there exists a simpler alternative construction leading to a different darboux realization which was introduced earlier by karasev ( @xcite ) . in this paragraph , we analyse further the case @xmath108 together with a `` flat poisson spray ''
@xmath50 defined from the flat connection @xmath109 on @xmath110 as in example [ ex : spray from nabla ] .
we shall also consider its `` opposite '' @xmath111 ( which will help us later on to take care of some sign issues ) , so that they are given by @xmath112 in this case , the symplectic realization realization @xmath113 obtained from the poisson spray as outlined in section [ sub : symplectic - realizations - from - sprays ] coincides with the original construction by weinstein in @xcite .
we shall thus refer to it as the _ weinstein realization_. ( sign convention ) .
here , we will consider the `` opposite '' symplectic realization @xmath114 obtained from @xmath111 in the same way .
observe that @xmath115 . in @xmath22
there is a simple way of obtaining global darboux frames for @xmath116 as follows .
since we are on flat space , we can define an @xmath67-family of maps @xmath117 from @xmath71 to @xmath110 by its action on coordinates @xmath118 it follows that @xmath119 since the flat spray @xmath111 only acts on the @xmath120 coordinates .
notice that when @xmath40 is an analytic bivector on @xmath121 then @xmath122 is an analytic function of @xmath67 . on the other hand , since @xmath123 and @xmath124 , the inverse @xmath125 exists in a neighborhood @xmath126 of the zero section .
it is then an immediate consequence of eq . that @xmath127 is a global darboux frame for @xmath128 .
the induced darboux - form realization is then given by the map @xmath129 it is not hard to see that the realization @xmath130 coincides with the one given by karasev in @xcite ( after rescaling the momenta using @xmath131 ) and we shall refer to it as the _ karasev realization_. we have then shown : [ prop : we - kar ] for @xmath16 , the map @xmath127 defined above is a global darboux frame for the weinstein realization defined by @xmath116 .
moreover , the induced darboux - form realization coincides with the karasev realization @xmath130 .
the components @xmath132 of the karasev realization @xmath130 can be thus obtained by solving the following equation : @xmath133 notice that the karasev realization @xmath130 does not coincide with the realization @xmath134 coming from the general construction of prop . [
cor : alpha for any m ] ( @xmath122 is not a flow ) .
we now turn to the formal setup by expanding @xmath135 in formal power series in @xmath9 .
( conventions ) .
we shall adopt the following conventions when working with formal power series in @xmath136 $ ] .
let @xmath137 with @xmath138 and @xmath139 $ ] .
for @xmath140 we shall denote ( multivariable taylor expansion ) @xmath141(w_{0}(x , p),p)\label{eq : formal composition}\ ] ] first of all notice that the formal flow of @xmath111 can be written as @xmath142 where @xmath143 is the formal flow of @xmath111 ( seen as a @xmath144-dependent vector field on @xmath22 ) defined in components by the following formal power series @xmath145 the map @xmath146 defined in eq .
is given by the formal expansion
@xmath147 this means that the formal version of the karasev realization @xmath130 is defined by the ( unique ) functions @xmath148 $ ] solving eq . with @xmath146 given by the formal expansion in eq . .
we have thus shown [ thm : formal canonical flat realization ] , @xmath149 is satisfied at all orders in @xmath67 . moreover , the formal darboux frame @xmath150 takes the formal weinstein realization @xmath151 to the formal darboux realization given by @xmath152 .
in this section , we briefly review the construction of @xcite that produces a formal symplectic realization out of the ( tree - level part of ) kontsevich s quantization formula , and we show that it coincides with the one obtained in theorem [ thm : formal canonical flat realization ] . in @xcite
, it was shown that the formal power series @xmath153 with @xmath154 and @xmath155 is a formal generating function for the formal symplectic groupoid @xmath156 ( where the cotangent bundle is endowed with its canonical symplectic structure @xmath157 ) integrating @xmath158 . in the above formula ,
* @xmath159 is the set of * kontsevich trees * of type @xmath160 ; * @xmath161 is a real number called the * kontsevich weight * of @xmath162 ; * @xmath163 is the symbol of the * kontsevich operator * @xmath164 .
we define these objects in the following section and refer the reader to @xcite for more details .
the formal generating function above depends on the poisson structure through the symbols of the kontsevich operators . in @xcite and @xcite ,
the scaling of the poisson structure is different : @xmath165 is used instead of @xmath166 , and thus the corresponding formal symplectic groupoid integrates the poisson structure @xmath167 instead of @xmath7 .
we also use kathotia s convention ( @xcite ) for the kontsevich weights ; namely , @xmath168 where @xmath169 is the weight actually defined by kontsevich in @xcite .
as shown in @xcite , the formal source and target maps can be extracted from the formal generating function as follows : @xmath170 therefore , we obtain a formal symplectic realization from @xmath171 to @xmath158 of the form @xmath172 as it will become clear after we define the kontsevich operators in the next section , the factor @xmath173 in the formula above comes from the recalling of the poisson structure : namely , @xmath174 a * kontsevich graph * of type @xmath175 is a graph @xmath176 whose vertex set is partitioned in two sets of vertices @xmath177 , the * aerial vertices * @xmath178 and the * terrestrial vertices * @xmath179 such that * all edges start from the set @xmath180 , * loops are not allowed , * there are exactly two edges going out of a given vertex @xmath181 , * the two edges going out of @xmath181 are ordered , the first one being denoted by @xmath182 and the second one by @xmath183 .
an * aerial edge * is an edge whose end vertex is aerial , and a * terrestrial edge * is an edge whose end vertex is terrestrial .
we denote by @xmath184 the set of kontsevich graphs of type @xmath175 .
figure [ fig : k-2 - 3 ] illustrates a graphical way to represent kontsevich graphs . given a poisson structure @xmath40 and a kontsevich graph @xmath185
, one can associate a @xmath186-differential operator @xmath164 on @xmath0 in the following way : for @xmath187 , we define @xmath188\prod_{i\in v_{\gamma}^{g}}\big(\prod_{\substack{e\in e_{\gamma}\\ e=(*,i ) } } \partial_{i(e)}\big)f_{i}.\ ] ] the symbol @xmath189 of @xmath190 is defined by @xmath191 the following figure illustrates this . a kontsevich graph @xmath162 of type @xmath160 .
the terrestrial vertices are on an imaginary line ( the dotted line ) and the aerial vertices are placed above them .
the first arrow stemming out of an arial vertex has a solid black head , while the second one has a hollow white one . ]
consider the kontsevich graph @xmath162 as in figure [ fig : k-2 - 3 ] .
the three aerial vertices @xmath192 and @xmath193 correspond to three copies of the poisson structure , say @xmath194 , @xmath195 and @xmath196 .
the two terrestrial vertices @xmath197 and @xmath198 correspond to two functions smooth functions on @xmath22 , say @xmath199 and @xmath200 .
the arrows correspond to derivatives .
therefore , the kontsevich operators associated with this graph is @xmath201 where we use the einstein summation convention of repeated indices .
the corresponding symbol is obtained by replacing @xmath202 by @xmath203 and @xmath204 by @xmath205 : namely , @xmath206 the order of the arrow is important because flipping , for example , the order of the first aerial vertex order would introduce a sign , since @xmath207 .
@xmath159 is a subset of @xmath208 that we now define : let @xmath209 be a kontsevich graph .
the * interior * of @xmath162 is the graph @xmath210 obtained from @xmath162 by removing all terrestrial vertices and terrestrial edges .
a kontsevich graph is a * kontsevich tree * if its interior is a tree in the usual sense ( i.e. it has no cycles ) .
we denote by @xmath211 the set of kontsevich s trees of type @xmath175 .
we give now a very rough presentation of the kontsevich weights , which follows kathotia s conventions in @xcite , instead of the original ones of kontsevich in @xcite .
let @xmath212 be the upper half plane and @xmath213 be the @xmath120-coordinate line in @xmath212 .
given two points @xmath144 and @xmath13 in @xmath212 with @xmath214 , we define the * angle function * @xmath215 , which depends on whether @xmath216 or not ( @xmath217 if @xmath216 and @xmath218 if @xmath219 ) , as depicted in figure [ fig : angles ] ( we refer the reader to @xcite for more details on the angles ) : ] observe that , for fixed @xmath13 and @xmath220 as in figure [ fig : angles ] , the angles @xmath221 and @xmath222 determine completely the position of the point @xmath144 . a * configuration * @xmath223 with @xmath224 of a kontsevich s graph @xmath225 is an identification @xmath226 of its aerial vertices @xmath227 with the points of the @xmath228-tuple @xmath229 and of its terrestrial vertices @xmath197 and @xmath198 with respectively @xmath230 and @xmath231 .
we denote by @xmath232 the manifold of all the configurations of @xmath162 ; its compactification @xmath233 is a manifold with corners ( see @xcite for details ) . given @xmath225
, we have @xmath228 pairs of angle - functions on @xmath233 , which we denote by @xmath234 .
they are determined by the configuration of the aerial vertices of @xmath162 : namely , @xmath235 is the angle - function as above , where @xmath236 is the vertex target of @xmath237 ( the @xmath238 edge of @xmath239 ) .
this allows us to define the angle - form on @xmath233 associated with @xmath162 : @xmath240 since @xmath241 is a @xmath242-form and @xmath233 is a compact real @xmath242-manifold , we can integrate @xmath241 , which give us the kontsevich weight of @xmath162 : @xmath243 following kathotia , we do not include the original factor @xmath244 in the definition of the weights , and the factor @xmath245 is taken into account by the rescaling of the angles ( we use @xmath246 instead of @xmath247 ) . as mentioned earlier , the realization @xmath248 of theorem [ thm : formal canonical flat realization ] is a formal version of the one introduced by karasev in @xcite which , in turn , was shown in @xcite to coincide with the realization @xmath23 coming from kontsevich trees . here
we repeat the argument for completeness ( c.f .
theorem [ thm : realizations ] ) and also show that the coefficients in the @xmath67-expansion of @xmath152 admit a recursive definition ( c.f .
equation below ) .
recall that a formal symplectic realization from @xmath171 to @xmath158 is a formal power series of the form @xmath249 where the @xmath250 s are smooth functions on @xmath251 , formally satisfying ( i.e. at each order ) the partial differential equation @xmath252 [ lem : uniqueness flat realizations ] all formal symplectic realizations from @xmath171 to @xmath253 starting with the same first order term @xmath254 and such that , for @xmath255 , we have @xmath256 coincide .
if we develop the right hand side of in taylor s series and read off the @xmath257 order in @xmath9 for each @xmath258 , we obtain an infinite system of recursive equations of the form @xmath259 where @xmath260 is an expression involving only lower orders of @xmath261 .
now suppose we have another formal symplectic realization @xmath262 satisfying and such that @xmath263 .
we show by induction on @xmath239 that @xmath264 for all @xmath239 . by assumption , it is true for @xmath265 .
suppose that it is true all @xmath266 .
set @xmath267 . then tells us that @xmath268 this implies , by the stokes theorem , that there is a function @xmath269 such that @xmath270
now , using and , we have that @xmath271 which implies that @xmath272 does not depend on @xmath144 , and thus @xmath273 .
now , using definition equation and formula for @xmath146 , one gets that equation for @xmath274 is equivalent to : @xmath275(x , p).\ ] ] reading this equation at order @xmath276 in @xmath67 , one gets @xmath277(x , p)+\end{gathered}\ ] ] @xmath278 this is a recursive formula for the terms @xmath279 in @xmath152 .
[ thm : realizations]let @xmath280 be the formal symplectic realization of theorem [ thm : formal canonical flat realization ] .
this realization coincides with the kontsevich realization @xmath23 and , moreover , the coefficients in the formal expansion of these coinciding realizations can be recursively computed using formula .
the last statement follows from the first one and from the computation above .
it remains to prove the first statement , which is done by applying lemma [ lem : uniqueness flat realizations ] as follows . as shown in @xcite
, @xmath23 satisfies the hypothesis of this lemma . as for @xmath248 ,
let us write @xmath281 using equation , for @xmath282 , we obtain @xmath283 .
thus , the first term of @xmath248 is @xmath284 , which is the same as for the kontsevich realization . by inspection of the recursive formula above
, we see that each term of @xmath285 has a factor of the form @xmath286 which is zero by the antisymmetry of the poisson structure .
this shows that holds .
an induction on @xmath276 using the same equation immediately yields .
for a linear poisson structure @xmath287 on @xmath22 ( i.e. the one living on the dual @xmath288 of a lie algebra @xmath289 ) , the kontsevich trees that have an aerial vertex with more than one incoming edge have vanishing symbols : i.e. @xmath290 .
thus , the kontsevich trees that have a non - vanishing contribution to the generating function are of the form since , for two of these graphs @xmath162 and @xmath292 with @xmath228 arial vertices , we have that @xmath293 , and since there are @xmath294 of these graphs , we obtain the following explicit formula for the corresponding formal symplectic realization in the case of linear poisson structures : @xmath295 where @xmath296 for @xmath297 and @xmath298 is identified with a linear function on @xmath299 , and where we used a well - known result of kathotia in @xcite about the kontsevich weights of these types of kontsevich s graphs . in this section , we generalize formula in the case of non - linear poisson structures and we express the realization in terms of a sum over topological rooted trees . we show that the only kontsevich trees satisfying in the non - linear case are exactly the ones that kathotia in @xcite calls `` w - computable . ''
building on his results , we set up a recursive formula for the corresponding kontsevich s weights in terms of topological rooted trees .
it is interesting to notice that a direct computation of the formal karasev realization @xmath152 using the recursion in the linear poisson case yields the r.h.s . of formula .
thus , in this case , the conclusion above about the kontsevich weights can also be obtained as a corollary of the fact that @xmath300 ( c.f .
theorem [ thm : realizations ] ) . on the other hand , directly solving the recursion for non - linear poisson structures is more complicated and a general solution will be presented in the next subsections ( c.f .
theorem [ thm : formula ] ) .
a * graph * is the data @xmath176 of a finite set of vertices @xmath301 together with a set of edges @xmath302 , which is a subset of @xmath303 .
the number of vertices is called the * degree * of the graph and is denoted by @xmath304 .
we think of @xmath305 as an arrow that starts at the vertex @xmath306 and ends at @xmath307 .
two graphs are * isomorphic * if there is a bijection between their vertices that respects theirs edges .
the set @xmath308 of all isomorphic graphs @xmath83to a given graph @xmath162 is called a * topological graph*. a * symmetry * of a graph is an automorphism of the graph ( i.e. a relabeling of its vertices that leaves the graph unchanged ) .
the group of symmetries of a given graph @xmath162 will be denoted by @xmath309 .
note that the number of symmetries of all graphs sharing the same underlying topological graph is equal ; we define the * symmetry coefficient * @xmath310 of a topological graph @xmath311 to be the number of elements in @xmath309 , where @xmath312 .
a * rooted tree * is a graph that ( 1 ) contains no cycle , ( 2 ) has a distinguished vertex called the * root * , ( 3 ) whose set of edges is oriented toward the root .
we will denote the set of rooted trees by @xmath313 and the set of topological rooted trees by @xmath25 $ ] .
the set of topological rooted trees can be described recursively as follows : the single vertex graph @xmath314 is in @xmath25 $ ] and if @xmath315 $ ] then so is @xmath316_{\bullet},\ ] ] where the bracket is to be thought as grafting the roots of @xmath317 to a new root , which is symbolized by the subscript @xmath314 in the bracket @xmath318 $ ] .
for instance , the formal expressions @xmath319_{\bullet},\qquad[[\bullet]_{\bullet}]_{\bullet},\qquad[\bullet,\bullet]_{\bullet},\ ] ] correspond , from left to right , to the topological trees depicted in figure [ fig : the - rooted - tree-1 ] . also observe that the total number of `` @xmath314 '' in these formal expressions corresponds to the degree of the rooted tree ( i.e. the total number of vertices ) .
since we are dealing with topological rooted trees , the ordering in @xmath320_{\bullet}$ ] is not important ( for instance , we do not distinguish between @xmath321_{\bullet}]_{\bullet}$ ] and @xmath322_{\bullet},\bullet]_{\bullet}$ ] ) .
[ def : elemdiff ] let @xmath323 be a vector field on @xmath22 .
we define the * elementary differential * of @xmath324 recursively as follows : for the single vertex tree , we define @xmath325 , and for @xmath33 $ ] in @xmath25 $ ] , we define @xmath326 where we used the einstein summation convention . for a ( non - topological )
rooted tree @xmath327 , we define @xmath328 , where @xmath329 is the topological tree underlying @xmath27 .
we now want to find for @xmath23 a formula generalizing formula as a sum over topological rooted trees . for this
, we start by characterizing the kontsevich trees whose contributions in the expression of @xmath23 is not zero .
if @xmath333 satisfies ( 1 ) and ( 2 ) , it is clear that it also satisfies : namely , there will then be only one power of @xmath334 in @xmath163 , since there is only one edge landing on @xmath198 .
the converse is more subtle .
suppose we have @xmath333 such that holds .
then ( 2 ) is obvious , otherwise there would be more that one power of @xmath334 occurring in @xmath163 , and the differentiation by @xmath335 would not kill them all . setting @xmath336 would then yields zero . the hard part is to prove ( 1 ) , i.e. , that the interior of @xmath162 is a rooted tree . in other words ,
we need to show that @xmath210 has only one root and that its edges are all oriented toward this root .
let us first find the root .
since @xmath162 has only one edge landing on @xmath198 , there is a distinguished vertex ; namely , the one this edge stems from : let us call this vertex @xmath337 and consider it as our candidate for the root
. to see this , suppose it is not the case , i.e. that the second edge of @xmath337 is aerial .
the total number of aerial edges must be @xmath338 , since @xmath210 is a tree with @xmath228 vertices .
the total number of edges stemming out from @xmath339 is @xmath340 . among those
, @xmath341 must be aerial .
so @xmath339 must have @xmath342 terrestrial edges landing on @xmath197 .
this means that at least one vertex in @xmath339 has two arrows landing on @xmath197 , in which case @xmath290 , because of the antisymmetry of the poisson structure .
this contradicts the fact that @xmath332 .
since the two edges of @xmath337 are terrestrial , @xmath337 has only incoming edges .
let us define @xmath343 as being the set of aerial vertices ( including @xmath337 itself ) that are connected to @xmath337 by an oriented path pointing toward @xmath337 .
if @xmath344 , we are done , and @xmath210 is a rooted tree .
suppose it is not the case : this means that there is a vertex @xmath345 with an edge pointing out from @xmath343 into a vertex @xmath346 .
now the two edges of @xmath347 can not be both terrestrial ( otherwise , by ( 2 ) , they should both land on @xmath197 , which implies that @xmath290 ) .
so , one of these edges should point into a vertex @xmath348 ( @xmath347 can not point back into @xmath343 , otherwise we have a cycle ) .
the situation is illustrated in the following figure : we repeat the same argument for @xmath349 , and we see that we obtain a sequence @xmath350 of vertices in @xmath351
. this process must terminate because ( 1 ) @xmath352 is finite , ( 2 ) the vertices in the sequence must be different from each other and can not point back into @xmath343 ( otherwise we create a cycle ) .
consider the last vertex @xmath353 of this sequence .
then @xmath353 can not have any aerial edge ( otherwise we introduce a cycle ) , and thus @xmath353 must have two terrestrial edges landing on @xmath197 .
but then , this implies that @xmath290 by antisymmetry of the poisson structure .
the fact that both @xmath162 and @xmath292 have the same underlying topological rooted tree as interior means that they differ by ( 1 ) a relabeling of their aerial vertices , ( 2 ) by a flip of their edge order ( since all of the edges land on @xmath197 except for the second edge of the root ) .
as explained in @xcite , the first operation leaves both the weights and the bidifferential operators unchanged .
the second operation introduces the same sign @xmath357 in both the weight and the operator ; since we are taking products of the two , these signs cancel out .
let @xmath327 be a rooted tree .
we define @xmath358 to be the kontsevich tree
whose interior is @xmath27 and such that all of its terrestrial edges land on @xmath197 except for the root of @xmath27 , which has its first edge landing on @xmath197 and its second landing on @xmath198 .
all the edges landing on @xmath197 are considered to be the first ones .
kontsevich trees having isomorphic tree as interior differ only by a relabeling of their aerial vertices , which is an operation that leaves both the corresponding weights and bidifferential operators unchanged .
we can prove this by induction using the graphical representation for the kontsevich graphs .
graphically , @xmath369 can be represented as a graph in the following way : start with @xmath162 , detach its single edge landing on @xmath198 from its target , let it point to nothing as illustrated in figure [ fig : deriv ] below ( we will say that this edge points `` in the air '' ) , and label it with @xmath370 ; this procedure implements the differentiation by @xmath371 .
then remove @xmath198 , which represents graphically setting @xmath334 to zero .
let us denote by @xmath372 the type of graphs obtained this way .
what we need to show now is that @xmath374 with @xmath363 $ ] can be also described as a graph in @xmath372 ( up to a relabeling of the aerial vertices and a number of flips of the edge order ) .
more precisely , we need to show that , in terms of the graphical representation , @xmath374 is obtained from the kontsevich tree @xmath375 with @xmath376 ( up to a relabeling of the aerial vertices ) by removing @xmath198 and letting the single edge landing on @xmath198 pointing in the air with label @xmath370 .
suppose that @xmath377 .
then @xmath378 is the kontsevich tree with one aerial vertex with its first edge landing on @xmath197 and its second edge landing on @xmath198 .
then , @xmath379 is now the graph with one aerial vertex , a first edge landing on @xmath197 and the second edge pointing in the air with label @xmath370 .
the term represented by this graph is @xmath380 which is exactly @xmath381 .
now suppose that for @xmath363 $ ] with @xmath382 , we have that the graph representing @xmath383 is in @xmath372 .
take now @xmath384 with @xmath385 .
recursively , we have that @xmath386 $ ] with @xmath387 for @xmath388 and @xmath389 since , by induction , we have that the @xmath390 s are represented ( up to relabeling of the aerial vertices ) by the graph @xmath391 with @xmath392 to which we have removed @xmath198 and let the corresponding terrestrial edge point in the air with label @xmath393 . now the operation can be represented graphically as grafting all the edges pointing in the air in the graphical representation of the @xmath394 s to a new root vertex ( corresponding to @xmath395 ) , which itself has its first edge landing on @xmath197 and its second edge landing on @xmath198 .
the following figure illustrates the process : what we obtain is a representation of @xmath374 in terms of a graph in @xmath372 ; namely , the graph obtained from @xmath375 by removing @xmath198 and letting the corresponding edge point in the air with the label @xmath370 .
this is a direct computation : from and proposition [ prop : trees ] , we have @xmath397}\sum_{\underset{\gamma_{i}\in t}{\gamma\in\tilde{t}_{n,2}}}\frac{w_{\gamma}}{2^{n}}d_{t}l_{v},\end{aligned}\ ] ] where we used proposition [ prop : ismorphic trees ] and proposition [ prop : terms ] for the second equality .
we conclude the proof by observing that the number of kontsevich tree in @xmath330 whose interior is in the class of @xmath27 is @xmath398 namely , two trees @xmath399 having isomorphic interior differ by a relabeling of their aerial vertices and a flip of the edge order .
there are @xmath400 such flips ( where @xmath228 is the number of aerial vertices ) and @xmath401 different relabelings of the aerial vertices .
theorem [ thm : formula ] gives a very explicit description of the formal symplectic groupoid described in @xcite , which integrates the poisson manifold @xmath402 .
recall that the graph of its product is given by @xmath403 , its inverse by @xmath404 , and the unit map by @xmath405 .
now we also have an explicit formula for the source map , given by theorem [ thm : formula ] , and for the target map @xmath406}\e^{|t|}\left((-1)^{|t|}\frac{w_{t}}{|\sym(t)|}\right)d_{t}^{i}l_{v},\label{eq : kk - target}\ ] ] since @xmath407 and @xmath28 is homogeneous in @xmath144 of degree @xmath408 .
we proceed by induction on the tree degree , following the strategy outlined by kathotia in @xcite . for one vertex , we have @xmath413 , which is the kontsevich weight of @xmath378 .
suppose this is true for topological rooted trees with @xmath228 vertices .
let @xmath414 $ ] such that @xmath415 for @xmath416 .
consider the graph @xmath417 made out of @xmath375 by letting the root second edge point in the air toward an additional fixed vertex @xmath229 at distance @xmath231 and angle @xmath221 from @xmath197 as in the following figure : let @xmath418 be the angle - form defined by the angle subtended by the aerial vertices of @xmath417 minus the additional vertex @xmath229 .
we define @xmath419 where the integration is over the configuration space of the @xmath408 aerial vertices of @xmath375 .
clearly , @xmath420 .
denote by @xmath421 and @xmath422 the angle subtended by the root @xmath337 of @xmath27 in @xmath417 at @xmath197 and @xmath229 as in figure [ fig : formual ] . by fubini
, we have that @xmath423 where @xmath424 is the angle - form sustended by the aerial vertices of @xmath417 minus @xmath229 . because there is no aerial arrow between the aerial vertices of @xmath425 and @xmath426
if @xmath427 , we have that @xmath428 and by fubini , we obtain @xmath429 in @xcite , kathotia computed that @xmath430 , which allows us to conclude that @xmath431 and @xmath432 .
the iterated integrals for the weights are very easy to compute ; for the first rooted trees , we have @xmath433_{\bullet}}(\theta ) & = & \frac{\theta^{2}}{2}-\frac{\theta}{2}+\frac{1}{12},\\ i_{[\bullet,\bullet]_{\bullet}}(\theta ) & = & -\frac{\theta^{3}}{3}+\frac{\theta^{2}}{2}-\frac{\theta}{4}+\frac{1}{24},\\ i_{[[\bullet]_{\bullet}]_{\bullet}}(\theta ) & = & -\frac{\theta^{3}}{6}+\frac{\theta^{2}}{4}-\frac{\theta}{12},\end{aligned}\ ] ] which give the first kontsevich weights ( with kathotia normalization factors ) : @xmath434_{\bullet}}=\frac{1}{12},\qquad w_{[\bullet,\bullet]_{\bullet}}=\frac{1}{24},\qquad w_{[[\bullet]_{\bullet}]_{\bullet}}=0,\ ] ] which , in turn , give the first terms of the formal symplectic realization : @xmath435 a. coste , p. dazord , and a. weinstein , _ groupoides symplectiques _
, publications du departement de mathematiques , nouvelle serie _ a _ , vol . * 2 * , univ .
claude - bernard , lyon , ( 1987 ) . | we study the relationship between several constructions of symplectic realizations of a given poisson manifold .
we show that , for an arbitrary poisson structure on @xmath0 , the realization given by tree - level part of kontsevich star - product coincides with the one given by a formal version of the original construction of weinstein , when suitably put in global darboux form .
we give an explicit formula for this realization in terms of rooted trees involving only computable kontsevich s weights , for which we set up a simple iterated integral , building on results of kathotia . |
model selection is central to statistics , and the most popular statistical techniques for model selection are akaike information criterion ( aic ) @xcite , bayesian information criterion ( bic ) @xcite and minimum description length ( mdl ) @xcite . the basic idea behind mdl principle is to equate compression with finding regularities in the data .
since learning often involves finding regularities in the data , hence learning can be equated with compression as well .
hence , in mdl , we try to find the model that yields the maximum compression for the given observations .
the first mdl code introduced was the two - part code @xcite .
it was shown that the the two - part code generalizes maximum entropy principle @xcite . however , in the past two decades , the normalized maximum likelihood ( nml ) code , which is a version of mdl , has gained popularity among statisticians .
this is particularly because of its minimax properties , as stated in @xcite , which the earlier versions of mdl did not possess .
efficient methods for computing nml codelengths for mixture models have been proposed in @xcite . in both these papers
, the aim of model selection is to decide the optimum number of clusters in a clustering problem . in a maximum entropy approach to density estimation one has to decide , _ a priori _ , on the amount of ` information ' ( for example , number of moments ) that should be used from the data to fix the model .
the minimax entropy principle @xcite states that for given sets of features for the data , one should choose the set that minimizes the maximum entropy .
however , it can easily be shown that if there are two feature subsets @xmath0 and @xmath1 and @xmath2 , the minimax entropy principle will always prefer @xmath1 over @xmath0 . hence , though the minimax entropy principle is a good technique for choosing among sets of features with same cardinality , it can not decide when the sets of features have varying cardinality . in this paper
, we study the nml codelength of maximum entropy models . towards this end , we formulate the problem of selecting a maximum entropy model given various feature subsets and their moments , as a model selection problem .
we derive nml codelength for maximum entropy models and show that our approach is a generalization of minimax entropy principle .
we also compute the nml codelength for discriminative maximum entropy models .
we apply our approach to gene selection problem for leukemia data set and compare it with minimax entropy method .
let @xmath3 be a family of probability distributions on the sample space @xmath4 .
@xmath5 denotes the sample space of all data samples of size @xmath6 .
@xmath7 denotes an element in @xmath5 , where @xmath8 is a vector in @xmath4 .
a two - part code encodes the data sample @xmath9 by first encoding a distribution @xmath10 and then the data @xmath11 . the best hypothesis to explain data
@xmath11 is then the one that minimizes the sum @xmath12 @xcite .
according to mdl principle @xmath13 \label{crude } , \ ] ] where @xmath14 is the codelength of the data , @xmath15 is the codelength of distribution and @xmath16 is the codelength of data given the distribution .
@xmath16 is a measure of the error of data @xmath11 with respect to distribution @xmath17 .
hence , if @xmath17 approximates @xmath11 well enough , @xmath16 should be small and vice versa . by kraft s inequality
, there exists a codelength function @xmath18 on @xmath5 given by @xmath19 .
we an use @xmath20 as @xmath16 directly , since it is the unique minimizer of expected codelength , when @xmath17 is indeed the true distribution .
the normalized maximum likelihood code is one of the several ways for constructing @xmath15 . to formalize the definition of normalized maximum likelihood , we need the following notion of ` regret ' @xcite
. let @xmath3 be a model on @xmath5 and let @xmath21 be a probability distribution on @xmath5 .
the regret of @xmath21 with respect to @xmath3 for data sample @xmath7 is defined as @xmath22 \enspace .\ ] ] the regret is nothing but the extra number of bits needed in encoding @xmath7 using @xmath21 instead of the optimal distribution for @xmath7 in @xmath3 .
the worst case regret denoted by @xmath23 is defined as the maximum regret over all sequences in @xmath5 @xmath24\right ] \enspace .\ ] ] our aim is to find the distribution @xmath21 that minimizes the maximum regret . to this end
, we define the complexity of a model @xmath25 as @xmath26 where @xmath27 denote the maximum likelihood estimate ( mle ) of the parameter @xmath28 for the model @xmath3 for the data sample @xmath29 . in the above equation and subsequent sections , the integral
is defined subject to existence .
also , we assume that mle is well defined for the model @xmath3 .
furthermore , the error of a model is defined as @xmath30 \enspace .
\label{eq : error_def}\ ] ] the following result is due to @xcite . if the complexity of a model is finite , then the minimax regret is uniquely achieved by the normalized maximum likelihood distribution given by @xmath31 the corresponding codelength @xmath32 also known as the stochastic complexity of the data sample @xmath7 is given by @xmath33
let @xmath34 be a random variable taking values in @xmath4 .
let @xmath35 be a set of functions of @xmath34 . the resultant linear family @xmath36 is given by the set of all probability distributions that satisfy the constraints @xmath37 where @xmath38 is the empirical estimate of @xmath39 for the data @xmath7 .
the resulting maximum entropy model @xmath40 contains @xmath41 such that @xmath42 where @xmath43 . here
, @xmath44 is the normalizing constant . given a set of maximum entropy models characterized by their function set @xmath45 , we use nml code to choose the model that best describes the data .
the nml codelength of data @xmath7 for a given model @xmath3 is composed of two parts : ( i ) the error codelength and ( ii ) the complexity of the model .
error codelength of data sequence @xmath7 for the maximum entropy model @xmath46 is n times the maximum entropy of the corresponding linear family @xmath47 @xmath48 where @xmath49 is the maximum entropy distribution of @xmath36 given by . first , we compute the error codelength of the data @xmath7 for the model @xmath50 . by definition ,
@xmath51 using definition of @xmath50 from equation , we get @xmath52\notag \\ = & \inf_{\lambda } \left [ n\lambda_0 + \sum_{k=1}^m \lambda_k \sum_{i=1}^n \phi_k(\mathbf{x}^{(i ) } ) \right]\notag \\ = & \inf_{\lambda}\left [ n\lambda_0 + \sum_{k=1}^m \lambda_k \left(n\bar{\phi}_k(\mathbf{x}^n)\right ) \right ] \notag \\
= & n\left[\inf_{\lambda } \left ( \lambda_0 + \sum_{k=1}^m \lambda_k \bar{\phi}_k(\mathbf{x}^n)\right ) \right ] \label{eq : error_lambda } \enspace , \end{aligned}\ ] ] where @xmath38 is the sample estimate of @xmath39 for the data @xmath7 and @xmath53 . using lagrange multipliers , it is easy to see that the maximum entropy distribution for the linear family @xmath47 has the form @xmath54 for some @xmath55 .
since maximum entropy distribution always exists for a linear family , the parameters @xmath56 can be obtained by maximizing the log likelihood function .
@xmath57 \notag \\ = & { \operatornamewithlimits{argmin}}_{\lambda } \left [ \lambda_0 + \sum_{k=1}^m \lambda_k \bar{\phi}_k(\mathbf{x}^n ) \right ] \label{eq : lambda_star } \enspace .
\end{aligned}\ ] ] where we remove the negative sign to change @xmath58 to @xmath59 .
the notation @xmath60 is used to denote the empirical mean of @xmath61 for the data @xmath7 .
the corresponding entropy is given by @xmath62 where the last equality follows from the definition of @xmath47 in equation . by combining equations and and using the fact that maximum entropy distribution always exists for a linear family
, we get @xmath63 \label{eq : entropy_lambda } \enspace .\ ] ] by combining equations and , we get @xmath64 for fixed @xmath7 , the error depends on the function set @xmath65 through the above equation .
as the no . of functions in the function set
is increased , the size of the linear family decreases .
hence , the entropy of the maximum entropy distribution also decreases , since we are restricted to search for the maximum entropy distribution in a smaller space .
hence , error of the model decreases .
complexity of the maximum entropy model @xmath46 is given by @xmath66 where @xmath67 is the maximum entropy distribution of @xmath68 .
we have @xmath69 where we have used the definition of error in and its relationship with entropy in to get the result . by similar arguments as above , it is easy to see that the complexity of the model increases by increase in the number of constraints . by using and
we get the desired result .
hence , the nml codelength ( also known as stochastic complexity ) of @xmath7 for the model @xmath70 is given by @xmath71 in this section , we show that the presented nml formulation for maximum entropy is a generalization of the minimax entropy principle @xcite , where this principle has been used for feature selection in texture modeling .
let @xmath72 be sets of functions from @xmath4 to the set of real numbers .
corresponding to each set @xmath73 , there exists a maximum entropy model @xmath74 and vice - versa . the mdl principle states that given a set of models for the data
, one should choose the model that minimizes the codelength of the data .
here , the codelength that we are interested in is the nml codelength ( also known as stochastic complexity ) .
since , there exists a one - one relationship between the maximum entropy models and the function sets @xmath73 , the model selection problem can be reframed as @xmath75 \enspace .\ ] ] if we assume that all our models have the same complexity , then the second term in r.h.s can be ignored . since n , the size of data is a constant , the model selection problem becomes @xmath76 this is the classical minimax entropy principle given in @xcite .
hence , the minimax entropy principle is a special case of the mdl principle where the complexity of all the models are assumed to be the same and the models assumed are the maximum entropy models .
discriminative methods for classification , model the conditional probability distribution @xmath77 , where @xmath78 is the class label for data @xmath79 .
maximum entropy based discriminative classification tries to find the probability distribution with the maximum conditional entropy @xmath80 subject to some constraints @xcite , where @xmath81 is the class variable .
initially , information is extracted from the data in the form of empirical means of functions .
these empirical values are then equated to their expected values , thereby forming a set of constraints .
the classification model is constructed by finding the maximum entropy distribution subject to these sets of constraints .
we use mdl to decide the amount of information to extract from the data in the form of functions of features .
a straightforward application of this technique is feature selection .
the maximum entropy discriminative model @xmath46 , where @xmath82 is the set of all probability distributions of the form @xmath83 let us denote the denominator in above equation as @xmath84 .
since , we are not interested in modelling @xmath85 , we use the empirical distribution @xmath86to approximate @xmath85 @xcite .
the empirical distribution @xmath86 is given by @xmath87 and @xmath88 otherwise .
hence , the constraints become @xmath89 as discussed in @xcite , the sender - receiver model assumed here is as follows .
both sender and receiver have the data @xmath7 .
the sender is interested in sending the class labels @xmath90 .
if he sends the class labels without compression , he needs to send @xmath91 bits .
if , however , he uses the data to compute a probability distribution over the class labels , and then compress @xmath90 using that distribution , he may get a shorter codelength for @xmath90 .
his goal is to minimize this codelength , such that the receiver can recover the class labels from this code .
the error codelength of @xmath92 for the conditional model @xmath46 is equal to n times the maximum conditional entropy of the model @xmath93 .
error of the conditional model is given by @xmath94 \label{eq : error_lambda_disc}\!\!\ !
\enspace .
\end{aligned}\ ] ] where we have used similar reasoning as in to get the last statement .
also , the maximum conditional entropy distribution can be obtained by maximizing the corresponding log - likelihood function .
hence , @xmath95 correspondingly , @xmath96 the corresponding conditional entropy is given by @xmath97 \nonumber \\
= & \frac{1}{n}\!\ ! \left[\sum_{k=1}^m \lambda_k^*\!\ ! \left(\sum_{i=1}^n
\phi_k(\mathbf{x}^{(i ) } , c^{(i)})\right ) \right .
\!\!\!+\!\!\ ! \left .
\sum_{i=1}^n \log{z_\lambda(\mathbf{x}^{(i)})}\right ] \label{eq : entropy_lambda_disc } \!\!\!\enspace .
\end{aligned}\ ] ] here the first equality follows from the definition of conditional entropy as used in @xcite .
we use the definition of @xmath98 to convert the integral to a summation . by using and the fact that @xmath99 must sum up to 1 ,
we obtain the fourth equality . using and
, we obtain @xmath100 \label{eq : entropy_lambda_disc1 } \!\!\!\!\enspace .\ ] ] replacing the above equation in , we get the desired result . the complexity of the conditional model @xmath46 is given by @xmath101
we use gene selection as an example to illustrate discriminative model selection for maximum entropy models .
the dataset used is leukemia dataset available publicly at http://www.genome.wi.mit.edu .
the dataset was also used in @xcite to illustrate nml model selection for discrete regression .
the data set consists of two classes : acute myeloid leukemia ( aml ) and acute lymphoblastic leukemia ( all ) .
there are 38 training samples and 34 independent test samples in the data .
the data consists of 7129 genes .
the genes are preprocessed as recommended in @xcite . assuming the sender - receiver model discussed above
, the sender needs 38 bits or 26.34 nats in order to send the class labels of training data to receiver .
if the nml code is used , the sender needs 24.99 nats .
since the sender and receiver both contain the microarray data , the sender can use the microarray data to compress the class labels much more than can be obtained wihout the microarray data . specifically , we are interested in finding the genes which gives the best compression , or the minimum nml codelength .
[ ht ] genes , title="fig:",scaledwidth=45.0%,scaledwidth=20.0% ] [ ht ] genes , title="fig:",scaledwidth=45.0%,scaledwidth=20.0% ] [ ht ] [ ht ] for the purpose of our algorithm , we quantize the genes to various levels . we claim that quantizing a gene reduces the risk of overfitting of the model to data .
to support our claim , we have also plotted the change in accuracy with quantization level in figure [ accuracy ] for the top 25 genes . as can be seen from the graph , increasing quantizaton level from @xmath102 to @xmath103 results in a decrease in accuracy .
we have also plotted a graph for change in average nml codelength with quantization level for the top 25 genes in figure [ nml_codelength ] .
an interesting observation is that the minima of nml codelength coincides exactly with the maxima of accuracy .
a similar trend was obsrved when the number of genes were changed . hence , we quantize each gene to 5 levels .
other than the advantages of quantization mentioned above , quantization is also necessary for the current problem as the problem of calculating complexity can become intractable even for moderate n. the constraints that we use are moment constraints , that is @xmath104 .
we vary the value of m from 1 to 7 to get a sequence of maximum entropy models .
the nml codelength of the class labels is calculated for each such model .
the model that results in the minimum nml codelength is selected for each gene .
it was observed that for most genes , the nml codelength decreased sharply when m was increased from 1 to 2 .
the change in values of nml codelength was less noticeable for @xmath105 .
the variation of nml codelength of class labels for a typical gene are shown in figure [ nml_vs_m ] . in order to make the changes in nml codelength more visible
, we skip the nml codelength for m=1 .
our approach for ranking genes is as follows . for each gene , we select the value of m that gives the minimum nml codelength .
we then sort the genes in increasing order of their minimum nml codelengths .
the minimum codelength achieved is 8.35 nats , which is much smaller than 24.99 nats achieved without using the microarray data .
since compression is equated with finding regularity according to minimum description length principle , hence , it can be stated that the topmost gene is able to discover a lot of regularity in the data .
finally , we use mdl to build a classifier .
the amount of information to use for each gene is decided , by using mdl to fix the number of moments .
mdl is used to rank the features .
then , we use class conditional independence among features to build a maximum entropy classifier .
the number of genes used for the classifier are varied from 1 to 130 .
the resultant graph is compared with other maximum entropy classifiers in figure [ classifier ] , where the amount of information used per gene is the same for all genes .
finding appropriate feature functions and the number of moments is important to any maximum entropy method . in this paper
, we pose this problem as a model selection problem and develop an mdl based method to solve this problem .
we showed that this approach generalizes minimax entropy principle of @xcite .
we derived nml codelength in this respect , and extended it to discriminative maximum entropy model selection .
we tested our proposed method for gene selection problem to decide on the quantization level and number of moments for each gene .
finally , we selected the genes based on the codelength of the class labels and compared the simulation results with minimax entropy method .
the bottleneck for using mdl for model selection in discriminative classification is the computation of complexity .
more efficient approximations to calculate the complexity need to be developed to employ this approach for problems involving larger data sets .
p. kontkanen , p. myllymki , w. buntine , j. rissanen , and h. tirri , `` an mdl framework for data clustering , '' in _ advances in minimum description length _ , p. g. grnwald , i. j. myung , and m. a. pitt , eds .
mit press , cambridge , ma , 2005 . s. hirai and k. yamanishi , `` efficient computation of normalized maximum likelihood coding for gaussian mixtures with its applications to optimal clustering , '' in _ information theory proceedings ( isit ) , 2011 ieee international symposium on_. ieee , 2011 , pp . 10311035 . | in this paper , we treat the problem of selecting a maximum entropy model given various feature subsets and their moments , as a model selection problem , and present a minimum description length ( mdl ) formulation to solve this problem . for this , we derive normalized maximum likelihood ( nml ) codelength for these models .
furthermore , we show that the minimax entropy method is a special case of maximum entropy model selection , where one assumes that complexity of all the models are equal .
we extend our approach to discriminative maximum entropy models .
we apply our approach to gene selection problem to select the number of moments for each gene for fixing the model . |
experiments were performed on 200 nm thick @xmath0gete(111 ) films grown by molecular beam epitaxy on baf@xmath5(111 ) substrates @xcite . a protective amorphous te - cap used to avoid surface oxidation and degradation was removed in the ultrahigh vacuum chamber by annealing the thin film for 30 min at 250@xmath32 .
the surface quality checked by leed shows the 1@xmath331 diffraction pattern of the @xmath0gete(111 ) surface seen in fig .
the vuv - arpes / sarpes experiments were performed at the cophee end - station of the sis beamline @xcite , swiss light source ( sls ) , using @xmath28-polarized photons in the energy range 2025 ev , and an omicron ea 125 hemispherical energy analyzer equipped with two orthogonally mounted classic mott detectors .
the whole set - up allows the simultaneous measurements of all three spatial components of the spin - polarization vector for each point of the band structure .
the geometry of the experimental set - up is schematically shown in fig .
the spin - resolved mdcs were measured by rotating the sample azimuth @xmath34 such that @xmath15 or @xmath17 were aligned perpendicular to the scattering plane .
direct and immediate visualization of the spin splitting as a function of electron momentum was measured by exploiting the manipulator tilt angular degrees of freedom [ see @xmath24 in fig .
the spinintegrated fermi surfaces were measured with the channeltron detectors of the sarpes set - up by sweeping the manipulator tilt angle ( @xmath24 ) for selected polar angles @xmath35 as seen in the experimental setup in fig .
the angular and resolution - broadened energy range in spin - resolved mode were 1.5@xmath36and 60 mev , respectively . in spin - integrated mode the resolutions were set to 0.5@xmath36and 20 mev . to enhance the bulk sensitivity compared to vuv - arpes ,
experiments were also performed in the soft - x - ray energy range at the adress beamline @xcite .
the experimental results in the main text were reproduced on three samples with individual annealing preparation over different experimental runs .
all data were taken at 20 k. the ab - initio calculations are based on density functional theory ( dft ) as implemented within the multiple scattering theory ( sprkkr ) @xcite .
soc has been naturally included by use of a fully relativistic four - component scheme .
as s first step of our investigations we performed self - consistent calculations ( scf ) for 3d bulk as well as 2d semi - infinite surface of @xmath0gete(111 ) within the screened kkr formalism @xcite .
the corresponding ground state band structures are presented in terms of bloch spectral functions ( bsf ) .
the converged scf potentials served as an input quantities for our spectroscopic investigations .
the arpes calculations were performed in the framework of the fully relativistic one - step model of photoemission @xcite in its spin - density matrix formulation @xcite , which accounts properly for the complete spin - polarization vector , in particular for rashba systems like gete . together with a realistic model for the surface barrier potential , one - step calculations based on a semi - infinite half - space configuration were decisive to substantiate the @xmath0gete(111 ) surface - generated spectral features on both qualitative and quantitative levels .
constructive discussions with s. picozzi , r. calarco and r. bertacco are gratefully acknowledged .
this work was supported by the swiss national science foundation project no .
we acknowledge the financial support from german funding agencies dfg ( spp1666 ) and the german ministry bmbf ( 05k13wma ) is also gratefully acknowledged ( h.v.,.h.e.,j.b . and j.m . ) .
j.m . acknowledges the centem project , reg.no .
cz.1.05/2.1.00/03.0088 , co - funded by the erdf as part of the ministry of education , youth and sports op rdi program .
d.k . and v.h .
acknowledge the financial support of czech science foundation ( project 14 - 08124s ) | * in systems with broken inversion symmetry spin - orbit coupling ( soc ) yields a rashba - type spin splitting of electronic states , manifested in a kdependent splitting of the bands .
while most research had previously focused on 2d electron systems @xcite , recently a three - dimensional ( 3d ) form of such rashba - effect was found in a series of bismuth tellurohalides bite__x _ _ ( _ _ x__=i , br , or cl ) @xcite .
whereas these materials exhibit a very large spin - splitting they lack an important property concerning functionalisation , namely the possibility to switch or tune the spin texture .
this limitation can be overcome in a new class of functional materials displaying rashba - splitting coupled to ferroelectricity : the ferroelectric rashba semiconductors ( fers ) @xcite .
using spin- and angle - resolved photoemission spectroscopy ( sarpes ) we show that @xmath0gete(111 ) forms a prime member of this class , displaying a complex spintexture for the rashbasplit surface and bulk bands arising from the intrinsic inversion symmetry breaking caused by the ferroelectric polarization of the bulk ( fe ) .
apart from pure surface and bulk states we find resonant states ( sbr ) whose wavefunctions entangle the spinors from the bulk and surface contributions . at the fermi level their hybridization results in unconventional spin topologies with cochiral helicities and concomitant gap opening . the @xmath0gete(111 )
surface states ( ss ) and sbr states make the semiconductor surface conducting . at the same time
our sarpes data confirm that gete is a narrow - gap semiconductor with a well resolved dirac point in the 3d - valence band and a giant spinsplitting @xmath1 @xmath2 of the bulk states .
these results show that the @xmath0gete(111 ) surface / bulk electronic states are endowed with spin properties that are theoretically challenging to anticipate . as the helicity of the spins in rashba bands
is connected to the direction of the fe polarization , this work paves the way to allelectric nonvolatile control of spin - transport properties in semiconductors .
* semiconductor compounds such as lead chalcogenides , snte or gete have been extensively studied by band - structure calculations for which relativistic effects and soc were included @xcite .
emphasis was primarily put on the band - edge properties and the fundamental gap @xcite rather than the spin properties .
however , recent theoretical work has predicted that in topologically trivial @xmath0gete(111 ) the bulk bands are expected to be fully spin - polarized with the spin chirality directly coupled to the fe polarization @xcite .
a further development of this concept into applications requires first of all an experimental verification and , deeper theoretical considerations .
gete(111 ) and experimental geometry . * * ( a ) * brillouine zone ; * ( b ) * rheed and * ( c ) * leed pattern of the bulk and surface , respectively ; * ( d ) * te - terminated semi - infinite surface structural model with highlighted distorted rhombohedral cell .
orange arrow shows the displacement of the ge - atom responsible for the polar fe ordering . *
( e ) * experimental geometry with schematic drawing of our sarpes experiment . *
( f ) * spin - resolved initial state bloch spectral functions calculated for bulk brillouine zone along zu and za directions . ( inset )
@xmath3 quiver plot along the green arrow .
calculated spin - integrated * ( g ) * and spin - resolved * ( h ) * bsf from a semi - infinite surface . *
( i ) * logarithmic false color map of 1sm determinant demonstrating bulk , surface and sbr character of calculated arpes spectral intensities @xcite .
the green arrow indicates the @xmath4 momentum span where bulk rashba splitting is best observed.,scaledwidth=90.0% ] experiments were performed on 200 nm thick @xmath0gete(111 ) monocrystalline films grown by molecular beam epitaxy on baf@xmath5(111 ) substrates @xcite , with a te - terminated surface prepared _ in situ _ [ see methods and supplemental information ( si ) .1 ] .
fig .
1(b - d ) show the surface structural model together with results of the sample characterization by _ insitu _ low - energy electron diffraction ( leed ) and reflection high - energy electron diffraction ( rheed ) .
the lack of surface reconstruction in leed and the presence of kikuchi lines in rheed indicate superior quality @xmath0gete(111 ) surface with spontaneous fe moments @xcite .
below t@xmath6 = 700 k there is a spontaneous paraelectric to fe phase transition in which the cubic rocksalt phase changes to a polar rhombohedral structure , forming a fephase induced by elongation of the unit cell along the @xmath7 direction ( see the orange arrow in fig .
1d ) .
a concomitant shift in the te - sublattice is inducing an almost perfect collective feorder , up to @xmath892% oriented along the @xmath7 direction perpendicular to the sample surface ( si .2 ) .
the resulting symmetry breaking inside the @xmath0gete(111 ) triple - layers is responsible for the bulk rashba - splitting . of importance for the surfacesensitive uvsarpes on unpoled samples
is that the spontaneous polarization of fe domains persist in the subsurface region , as this ensures a net spin polarization also for the bulk states ( si .3 ) . due to the small probing depth of arpes , measured energy- , angle- and spin - resolved spectra can hardly be compared directly with a calculated bandstructure because photoemission intensities do not relate directly to dispersions of bare bands .
the task to describe the surface electronic structure differs from that in the bulk in several ways . due to
the broken crystal symmetry the bulk spin - resolved band - structure calculations is different compared to a semi - infinite surface [ fig .
1(f - g ) ] , mainly due to the presence of surface states ( ss ) .
the star - like feature in a constant energy map ( cem ) from bulk initial states orients along @xmath9 , whereas surface - like initial states orient along @xmath10 .
furthermore , for certain binding energies and momenta , the canted spin of the bulk states is opposite to what is predicted for surface states with rather tangential helicity .
thus , mixing the spectral weights from surface and bulk states inevitably result in a competition of spin textures between bulk and surface states which in our case facilitate their discrimination in a spin - resolved experiment . since arpes data critically depend on photoemission final states , to build a bridge between sarpes and theoretical bandstructure calculations we use so - called onestep photoemission model ( 1sm , see methods ) . in this most rigorous interpretation besides bulk states also surface evanescent ( ss ) and surface - bulk - resonance ( sbr ) states are considered . to better understand their spectral weights
, we employ the 1sm determinant criterion @xcite in fig .
1(i ) in order to qualitatively discern their contributions in the measured valence states . within the determinant logarithmic scale
small values correspond to surface states with the maximum of their wavefunctions between the first atomic layer and the surface potential .
states which embed in the bulk continuum have values close to one and are shown by black shades . on the other hand the sbr states , being partially reflected from bulk states , are given as red shades .
the ss / sbr and sbr / bulk transitions appear gradually , but as a whole the sbr is the primary spectral weight dispersing across the gap , forming metallic states at @xmath11 , and degenerating with bulk rashba bands at lower binding energies ( @xmath12 ) .
dispersion and comparison with semi - infinite surface calculations . *
( * a * ) uvarpes data taken along the @xmath13 direction for selected photon energies point out the z - point around 22 ev with the dirac point ( dp ) located around -0.25 ev @xmath12 .
the dashed lines highlight the dispersion of the dp along with the top of the valence band , indicating a bulk gap of few tens mev .
yellow arrows indicate the 2d - like dispersion of the sbr inside the gap ( section ) , in contrast to the 3d - like dispersion of the bulk - like band in section .
( * b * ) cems for selected @xmath12 inside the gap measured at @xmath14 of 22 ev and 475 ev in ( * c * ) . *
( d ) * arpes 1sm calculation calculated for @xmath14 of 22 ev along @xmath15 showing the influence of matrixelement effects on the spectral weight , which expose the surface states in the band - map section . *
( e ) * mdcs showing the 2d - like dispersion of sbr states at @xmath11 . * ( f ) * comparison of the semi - infinite and 10-layer slab calculations and corresponding second - derivative arpes data in panel * ( g ) * highlights the sbr buildup in the valence band [ yellow dashed lines in * ( f ) * and white dots in * ( g)*].,scaledwidth=100.0% ] we first concentrate on the localization of the z - point in * k*-space .
the relevant region of @xmath16momenta manifesting the rashbasplitting is expected along the @xmath13 direction , denoted in red in fig .
1(a ) .
the corresponding band maps shown in fig .
2(a ) have pronounced spectral intensity of the surface states at @xmath11 and the dirac point ( dp ) , which we used to relate experimental data to band - structure calculations .
the dp has minimum of -0.25 ev @xmath12 around a @xmath14 of 22 ev , constituting the z - point along the @xmath13 direction .
the vbm at the z - point is @xmath8 - 70 mev @xmath12 in @xmath15 and @xmath8 - 50 mev @xmath12 in @xmath17 mirror plane , as seen in fig .
2(a , f ) . what is the origin of the metallic states at @xmath11 ?
we note that sbr materialize as a superposition of bulk and surface spectral weight contributions , readily identified in corresponding theoretical cems in fig .
1(f - g ) .
we note that by mixing ss with bulk states the ss typically lose their well - defined surface character at @xmath11 by forming a sbr .
our results suggest that hybridization between the ss and bulk states is in - fact `` mediated '' by sbr .
recently such hybridization mechanism was also confirmed in sb@xmath5te@xmath18 topological insulators @xcite .
an important observation is that the sbr fill the gap from the valence band maximum ( vbm ) all the way up to @xmath11 , with progressively increased bulk - like spectral weight toward the vbm [ fig .
2(b ) ] .
the in - gap states reaching @xmath11 display rather 2d than 3d character in @xmath16 , as seen in the band - map section in fig .
2(a ) .
the metallic states filling the gap couple more to surface than bulk states for two reasons . first , consistently with the 1sm predictions in fig .
1(i ) , the spectral weight at @xmath11 for momenta
@xmath19 @xmath200.3 @xmath21 belongs to the sbr states and not to pure surface states .
this corroborates their momentum distribution dispersing toward the pure surface states near , as seen by the momentum distribution curves ( mdc ) in fig .
2(e ) and fig .
3b .
second , the sbr states filling the gap clearly detach from the main 3d - valence band dispersive trend inside the band - map section . .
* ( * a , b * ) ( top ) valence band arpes data with populations of spin - up and spin - down electrons near @xmath11 , as a function of electron momentum along @xmath17 and @xmath15(mid panels ) [ with corresponding spin polarizations @xmath22 and @xmath23 at @xmath11 ( bottom panels ) .
( * c * ) expected spin - resolved fermi surface map ( fsm ) compared with experimental data in * ( d)*. spin - resolved theoretical predictions from semi - infinite surface in fig .
1(h ) ( gray arrows ) , are overlaid with smoothed @xmath3 data from * a , b * ( light - green arrows ) .
dark - green arrows mark the results from the 3d vectorial spin analysis based on simultaneously fitted mdc peaks [ thin lines in * ( d ) * , see si ] .
the difference in spin - texture between the simplified model sketched in panel * ( c ) * and the experimental result is due to spin hybridization of the sbr wavefunctions indicated by dashed arrows in panel * ( e ) * ( see text).,scaledwidth=100.0% ] to further understand our arpes data it is useful to compare them with the semi - infinite and finite slab of 10-layers . as seen in fig .
2(f ) , it is evident that the sbr - states pile up as a superposition of bulk - projected states giving rise to enhanced spectral intensity .
such sbr spectral intensity buildup is highlighted with dashed lines and is dominant in the valence band arpes .
it refers to sbr states with strong coupling to the bulk continuum .
consequently , in agreement with the 1sm calculations , whole portions of pure bulkvb states , denoted with red and blue lines , are overlaid in arpes with the sbr states making their identification difficult by means of ground state bulk band - structure calculations @xcite . on the other hand the comparison between the 1sm arpes calculation for in fig .
2(d ) with the corresponding experimental data in panel ( a ) is rather convincing .
we discuss next our spin - resolved measurements near @xmath11 .
spin - resolved populations of spin - up and spin - down electrons were measured along the @xmath24 direction perpendicular to the scattering plane [ fig .
1(e ) ] .
we use a well - established quantitative 3d - vectorial analysis @xcite for simultaneous peak fitting of spin - integrated mdcs and spin polarizations .
the polarization curves are modeled until the best fit is reached by simultaneously fitting the mdc intensity and the polarizations @xmath22 , @xmath23 and @xmath25 ( si ) .
of importance are the mdc - peaks 1 - 2 - 3 labeled in fig .
3(a , b ) , the visibility of which is comprehensively described within the simultaneous fitting .
the resulting spin - texture at @xmath11 is seen in fig .
3(c ) .
thick arrows compare the 3d - fits with the inplane polarization predictions ( @xmath26 ) from the semi - infinite surface ( quiver plot in gray color ) . as a whole
the sarpes data at @xmath11 suggest that the sbr metallic states filling the gap have unconventional spin texture with parallel spin orientations different from the simple rashba model .
our conjecture is that such spin topology in the momentum distribution can be reconciled via soc - induced hybridization of the sbr - states .
highlighted in orange are regions where the bulk and surface contributions to the sbr wavefunctions hybridize .
the resulting gap opening in momentum distribution , sketched in fig .
3(e ) , is inducing a small spectral weight shift which realize the cochiral spin texture ( si ) .
similar interband gap opening and spin reorientation has been observed in 2d rashba systems @xcite . as a matter of fact , the in - gap states inside the band - map section in fig .
2(d ) display rather 2d than 3d character . to further examine our arpes data
we note that our experimental geometry predicts in 1sm calculations a suppression of spectral weight in @xmath15 below the dp for positive momenta .
in fact , on the experimental side the bulk contributions to the sbr function wash out ; uncovering the surface states inside the non - dispersive band - map section [ fig.2 ( a ) ] .
this photoemissiontransition matrixelement effect provides a direct evidence that the bulk and surface states indeed overlap in the initial states . as seen in figure 2 , a side effect of the vanished spectral weight intensity is that the cems appear with pronounced 3-fold rather than 6-fold symmetry @xcite
. 475 ev ) at -0.4 ev @xmath12 .
spin - resolved mdcs are measured along @xmath15 ( vertical ) and @xmath17 ( horizontal ) measured in equivalent z - point at @xmath2722 ev . the spectral weight of bulk rashba states for positive momenta vanishes as seen in fig .
3(a ) , exposing the surface states ( ss ) denoted with violet arrows .
the ss helicity ( violet arrows ) is opposite to pure rashba bulk states ( arrows with red and blue hues ) .
yellow frames highlight momenta where the 1sm predicts pure bulk states ( yellow arrows in fig .
1(f , i ) .
the measured inplane @xmath3 polarization for both bulk bands aligns with canted spins consistently with the bulk rashba initial states , showing a giant spinsplitting . along @xmath17
the semi - infinite surface quiver is displayed in gray arrows .
it is compared to measured ( light - green quiver ) and 3d - fitted ( dark - green ) spin textures . * ( b ) * hexagonal warping in the out - of - plane spin polarization @xmath25 measured along the paths ( a ) and ( b ) .
( inset ) corresponding 1sm @xmath25 calculation.,scaledwidth=100.0% ] we now look at the influence of the opposite helicity of ss and sbr states on the measured spin texture .
it turns out that an ideal locus of * k*-space momenta for measuring pure bulkrashba properties is around -0.5 ev @xmath12 in @xmath15 .
the black shades in the photoemission determinant along the green arrow in fig .
1(i ) predict bulk states with distinct canted helicity for inplane @xmath26 polarization [ @xmath22@xmath82@xmath23 , as seen in the inset in fig .
1(f ) ] . the same canted helicity is perfectly reflected in the measured polarization in fig .
4(a ) .
such experimental confirmation traces back the magnitude and direction of the spin polarization vector to the states from which these electrons originate .
this is non - trivial , since especially for systems with large soc , the total angular momentum is relevant for photoemission process rather than only the spin quantum number @xcite .
since the degree of spin polarization can depend strongly on the @xmath14 @xcite , the same states were probed with different photon energies ( si ) . because the same canted spin arrangement was confirmed in all cases ,
we conclude that the measured spin signals trace back to initial state bulk rashba spinors , making this the first experimental confirmation of the spin texture of a fers .
finally we comment on the out - of - plane polarization .
the band anisotropy between @xmath10 and @xmath9 induce rather strong hexagonal warping effect responsible for the out - of - plane @xmath25 spin polarization component @xcite .
sarpes data in data seen in fig .
4(b ) confirm the hexagonal warping measured in two equivalent @xmath15 directions denoted ( a ) and ( b ) .
as seen in the inset 1sm calculations , the measured warping is in excellent agreement with measured data . in a broader perspective , the unveiled @xmath0gete(111 ) bulk spin structure has far reaching consequences both for fundamental physics and applications .
based on our results @xmath0gete(111 ) is the prime candidate in spin - orbit - driven rashba physics in 3d * k*-space . resolving the spin structure at @xmath11 and inside
the narrow gap is relevant in terms of potential device - oriented applications .
namely , gete appears to have always @xmath28-type metallic transport properties due to ge - vacancies which lead naturally to a large density of holes in the valence band states @xcite .
nevertheless it manifests macroscopic , robust and reversible ferroelectricity promising for technological applications .
our findings confirm that a bulk rashba effect and ferroelectricity coexist and are coupled .
we emphasize that the giant spinsplitting @xmath1 @xmath2 is formed by bands which interconnect the rashba - split conduction- and valence band continuum .
we anticipate that the @xmath0gete(111 ) rashba physics is likely to extend to ge@xmath29mn@xmath30te multiferroic rashba semiconductors because the rhombohedral distortion persists up to @xmath31 @xcite .
thus the quest for semiconductors with combined functional properties might extend to systems where magnetism , ferroelectricity and rashba physics would offer unique property combinations for potential device applications . |
with the discovery of the top quark by the cdf and d@xmath9 collaborations @xcite , it has become natural to consider its properties , such as its couplings to the other particles of the standard model ( sm ) .
the top quark is singled out by its large mass ( @xmath10 175 gev ) , the same order as the electro - weak symmetry breaking ( ewsb ) scale , @xmath11 gev .
thus it seems reasonable that the top may prove valuable in probing the mechanism of mass generation in the standard model of particle physics .
indeed , if there is some type of new physics associated with the generation of mass , it may be more apparent in the top quark sector than for any of the other , lighter , known fermions , in accordance with the mass hierarchy . while production of @xmath12 pairs @xcite provides an excellent opportunity to probe the top s qcd properties , in order to carefully measure the top s electro - weak interactions it is also useful to consider single top production , in addition to studying the decay of the top quark in @xmath13 events .
single top production at the tevatron occurs within the sm in three different channels , the @xmath4-channel @xmath5 production , @xmath14 @xcite , the @xmath2-channel @xmath3-exchange mode , @xmath15 @xcite ( sometimes referred to as @xmath3-gluon fusion ) , and through @xmath16 production @xcite .
these three sub - processes have very different kinematics and experimental signatures @xcite , and as we shall discuss later , are sensitive to different types of new physics in the top quark sector .
thus they provide complimentary information about the properties of the top quark . in this paper
, we quantify the observation that the modes of single top production can be used to provide complimentary information about the top quark . in section [ singlet ]
, we discuss the properties of these separate modes of single top production in the context of the sm , summarizing the positive and negative aspects of each for probing the top quark sector at the tevatron run ii collider . at run
ii it is expected that a center - of - mass energy @xmath1 tev for the colliding protons and anti - protons will be realized . for the purposes of this study ,
we consider a data sample with an integrated luminosity of @xmath17 , and 30 . ] . in section [ resonance ]
we show the effect of an additional heavy resonance on each sub - process , and in section [ width ] we examine the effects from possible modifications to the top s couplings to the other particles of the sm .
finally , we examine what the possibility of new physics in the top quark sector implies about our ability to measure @xmath7 with confidence .
single top production at a hadron collider occurs dominantly through three sub - processes .
the @xmath5 mode of production shown in figure [ schannelfig ] occurs when a quark and an anti - quark fuse into a virtual @xmath3 boson , which then splits into a @xmath2 and @xmath18 quark .
the @xmath3-gluon fusion mode shown in figure [ tchannelfig ] , occurs when a @xmath19 quark fuses with a @xmath20 boson , producing a top quark .
the @xmath21 mode occurs when a @xmath19 quark radiates a @xmath22 , and is shown in figure [ twfig ] .
this mode may be important at the large hadron collider collider with @xmath23 14 tev . ]
( lhc ) , but is highly suppressed at the tevatron because of the massive @xmath3 and @xmath2 particles in the final state . because of its low cross section of about 0.1 pb ( for @xmath24 gev ) at the tevatron run ii @xcite , and relative insensitivity to the new physics effects we will be considering in this work , we do not present detailed studies of its cross section .
the three single top production processes contain the vertex of the sm , and thus are sensitive to any possible modification of this vertex from physics beyond the sm , and to the cabibbo - kobayashi - maskawa ( ckm ) parameter @xmath7 .
it has been shown @xcite that because of the difference in kinematics , it is possible to statistically disentangle the two sub - processes from each other , and from the sm backgrounds expected at the tevatron run ii . the total cross section for the @xmath5 production sub - process ( @xmath25 ) has been studied at next - to - leading - order ( nlo ) in qcd corrections ( order @xmath26 ) @xcite , by including nlo corrections to both the @xmath27 initial state ( c.f .
figure [ nlosi ] ) and the @xmath28 final state ( c.f .
figure [ nlosf ] ) .
detailed studies for the kinematics of this process at leading order ( lo ) @xcite exist in the literature , as well as more realistic studies in which the effects of soft gluons on the kinematics have been re - summed @xcite .
this sub - process is an attractive mode for studying top quark properties , because the initial state partons are quarks with relatively large momentum fraction , @xmath29 , and thus the parton densities are well understood . on the other hand , this mode suffers from a smaller cross section than the @xmath3-gluon fusion mode , and a larger percentage dependence on the mass of the top quark @xcite . in order to estimate the uncertainty in the nlo theoretical prediction for the @xmath5 production rate
, we examine the dependence of the result on the factorization and renormalization scales ( which in principle are separate quantities , however we will work with @xmath30 ) , and the dependence on the choice of the parton distribution functions ( pdf ) used . a physical observable
can not depend on the scale @xmath31 , and a theoretical calculation to all orders in perturbation theory must similarly be independent of the value of @xmath31 used in the calculation .
however , at fixed order in the perturbation series the theoretical prediction will still show some residual dependence on @xmath31 .
thus , by varying @xmath31 , one obtains an estimate of the dependence of the calculation on the uncalculated , higher order terms in the perturbation series . in @xcite
it was found that in order to cancel potentially large logarithms in the nlo perturbation coefficients , a scale choice of @xmath32 is appropriate , where @xmath33 is the center of mass energy of the incoming partons .
thus we vary the scale between @xmath34 and @xmath35 in order to estimate the scale dependence of the calculation . to examine the pdf dependence of the result
, we choose the cteq4 m @xcite and mrrs(r1 ) @xcite pdf .
for each of various values of @xmath36 , we combine all of these predictions ( due to varied pdf and scale ) to determine the upper and lower bounds on @xmath25 for each top mass under consideration , by taking the combination of pdf and scale choice which gives the highest and lowest result respectively . in this way we can study the correlated effect obtained by varying the pdf and scale ( and @xmath36 ) .
we also present the central value of @xmath25 , obtained as the point midway between the upper and lower bounds , and the mean , which is the average of the two pdf sets with the scale choice @xmath32 .
the mean value represents the best theoretical estimate of @xmath25 .
the fact that the mean and central values are the same for most of the top masses considered shows that the result does not depend more strongly on either raising or lowering the scale , and thus indicates that the choice of scale is appropriate .
our results are presented in table [ schantab ] and are shown graphically in figure [ topmassdep ] .
these results include the full ckm elements for the light quark fusion into @xmath5 , but do not include any ckm parameters in the @xmath3-@xmath2-@xmath19 vertex production and not @xmath37 or @xmath38 production , under the assumption that @xmath39 . ] .
the numbers include the rate of both @xmath40 and @xmath41 production through the @xmath5 mode . if we define the variable , m_t m_t - 175 , we can parameterize the curves in table [ schantab ] ( in units of pb ) as follows , @xmath42
-0.06 in .the central and mean values of @xmath25 ( in pb ) , along with the upper and lower bounds derived by varying the scale and by considering the pdf sets cteq4 m and mrrs(r1 ) . the central value is the value midway between the upper and lower bounds , while the mean is the averaged result of @xmath25 calculated using the cteq4 m and mrrs(r1 ) pdf with the canonical scale choice discussed in the text . [ cols="<,^,^,^,^",options="header " , ] 0.08 in
the authors would like to thank c. balazs for helpful discussion .
t. tait is grateful for conversations with e. l. berger , s. mrenna , and s. murgia .
yuan wishes to thank the cteq collaboration for useful discussions .
this work was supported in part by the nsf grant no .
part of t. tait s work was completed at argonne national laboratory , in the high energy physics division and was supported in part by the u.s .
department of energy , high energy physics division , under contract w-31 - 109-eng-38 .
f. abe _ et al .
lett . * 73 * , 225 ( 1994 ) ; + s. abachi _
et al._,phys .
* 72 * , 2138 ( 1994 ) ; + cdf collaboration , phys .
lett . * 74 * , 2626 ( 1995 ) ; + d0 collaboration , phys .
lett . * 74 * , 2632 ( 1995 ) .
p. nason , s. dawson , and r. k. ellis , nucl . phys . *
b303 * , 607 ( 1988 ) ; nucl . phys .
* b327 * , 49 ( 1989 ) ; nucl . phys . * b335 * , 260 ( 1990 ) .
+ g. altarelli , m. diemoz , g. martinelli , and p. nason , nucl . phys . *
b308 * , 724 ( 1988 ) .
+ w. beenakker , h. kuijf , w.l .
van neerven , and j. smith , phys . rev .
* d 40 * , 54 ( 1989 ) .
+ w. beenakker , w.l .
van neerven , r. meng , g.a .
schuler , and j. smith , nucl .
b351 * , 507 ( 1991 ) .
ellis , phys .
b259 * , 492 ( 1992 ) .
+ e. laenen , j. smith , and w.l .
van neervan , nucl . phys . *
b369 * , 543 ( 1992 ) ; phys . lett . *
b321 * , 254 ( 1994 ) .
+ n. kidonakis and j. smith , phys . rev .
* d 51 * , 6092 ( 1995 ) . s. catani , m. mangano , p. nason , and l. trentadue , phys . lett . * b378 * , 329 ( 1996 ) ; nucl . phys . * b478 * , 273 ( 1996 ) .
+ e. berger and h. contopanago , phys .
* b 361 * , 115 ( 1995 ) ; phys . rev . * d54 * , 3085 ( 1996 ) ; hep - ph/9706206 .
yuan , `` top quark physics at hadron colliders '' , published in ccast symposium 1993 , 259 ( 1993 ) ; hep - ph/9308240 .
+ d. o. carlson and c.p .
yuan , `` probing new physics from the single top production '' , particle phys . & phen .
1995 , 172 ( 1995 ) ; hep - ph/9509208 .
yuan , `` top quark physics '' , published in valencia elem .
part . phys .
1995 , 148 ( 1995 ) ; hep - ph/9509209 .
yuan , `` physics of top quark at the tevatron '' , talk given at 5th mexican workshop of particles and fields , puebla , mexico , oct .
30 - nov 3 , 1995 ; hep - ph/9604434 .
+ d. carlson , ph.d .
thesis , michigan state university , msuhep-050727 , august 1995 .
r.n . cahn and s. dawson , _ phys .
* b136 * , 196 ( 1984 ) , * b138 * , 464(e ) ( 1984 ) ; + m.s .
chanowitz and m.k .
gaillard , _ phys .
lett . _ * b142 * , 85 ( 1984 ) ; + g.l .
kane , w.w .
repko and w.r .
rolnick , _ phys .
_ * b148 * , 367 ( 1984 ) ; + s. dawson , _ nucl .
_ * b249 * , 427 ( 1985 ) ; + j. lindfors , _ z. phys .
_ * c28 * , 427 ( 1985 ) ; + w.b .
rolnick , _ nucl .
phys . _ * b274 * , 171 ( 1986 ) ; + p.w .
johnson , f.i .
olness and w .- k .
tung , _ phys .
* d36 * , 291 ( 1987 ) ; + z. kunszt and d.e .
soper , _ nucl .
phys . _ * b296 * , 253 ( 1988 ) ; + a. abbasabadi , w.w .
repko , d.a .
dicus and r. vega , _ phys .
* d38 * , 2770 ( 1988 ) ; s. dawson , _ phys . lett . _
* b217 * , 347 ( 1989 ) ; + s. cortese and r. petronzio , _ phys .
lett . _ * b276 * , 203 ( 1992 ) ; + i. kuss and h. spiesberger , _ phys .
rev . _ * d53 * , 6078 ( 1996 ) .
d. carlson and c.p .
yuan , phys .
b306 * , 386 ( 1993 ) .
+ d. carlson , e. malkawi and c.p .
yuan , phys .
* b337 * 145 ( 1994 ) .
+ f. larios and c.p .
yuan , phys . rev . * d 55 * , 7218 ( 1997 ) .
+ f. larios , t. tait , and c.p .
yuan , hep - ph/9709316 ( 1997 ) . | single top quark production at the fermilab tevatron run ii ( a @xmath0 collider with @xmath1 tev ) proceeds dominantly via two sub - processes , a @xmath2-channel @xmath3-gluon fusion process , and an @xmath4-channel @xmath5 process .
we show that these two sub - processes have different sensitivities to new physics effects in the top quark s electro - weak interactions .
the @xmath5 process is sensitive to new heavy charged resonances , such as a @xmath6 boson , while the @xmath3-gluon fusion process is more sensitive to modifications to the top s interaction , including flavor - changing neutral currents involving the top quark .
we examine the implications of these results on our ability to measure @xmath7 with confidence , and propose a quantity @xmath8 , which may be studied in order to characterize the confidence one may place upon a given measurement of @xmath7 from single top production .
psfig.tex 6.0 in 9.0 in -0.30truein 0.30truein 0.30truein = 1.5pc 0d hep - ph/9710372 + october 1997 + * the phenomenology of single top quark production at the fermilab tevatron * tim tait , and c .- p .
yuan _ department of physics and astronomy , michigan state university _ _ east lansing , michigan 48824 , usa _
pacs numbers : 14.65.ha , 12.39.fe , 12.60.-i |
to date , all direct determinations of the ckm matrix element magnitudes @xmath10 and @xmath11 were performed using measurements of @xmath16 meson semileptonic or leptonic decays at @xmath17 colliders . for both @xmath10 and @xmath11
, there are tensions between the most precise extractions from exclusive and inclusive semileptonic @xmath16 decays . the 2014 review of particle physics lists @xcite @xmath18 |v_{ub}|_{\rm incl . } = ( 4.41 \pm 0.15 ^{+0.15}_{-0.17 } ) \times 10^{-3 } , \hspace{2ex } & |v_{cb}|_{\rm incl . }
= ( 42.2 \pm 0.7 ) \times 10^{-3}. \label{eq : vubvcbrpp2014 } \end{array}\ ] ] the exclusive results in eq .
( [ eq : vubvcbrpp2014 ] ) are from the decays @xmath19 and @xmath20 ( where @xmath21 ) and use hadronic form factors from lattice qcd @xcite .
the discrepancy between the exclusive and inclusive results is a long - standing puzzle in flavor physics @xcite , and right - handed currents beyond the standard model have been considered as a possible explanation @xcite .
new lattice qcd calculations of the @xmath22 form factors published recently yield somewhat higher values of @xmath23 @xcite and @xmath24 @xcite , but the latest analysis of @xmath20 using lattice qcd gives @xmath25 @xcite and slightly increases the exclusive - inclusive tension .
moreover , the current experimental results for the ratios of the @xmath26 and @xmath27 ( @xmath21 ) branching fractions differ from the standard - model expectation with a combined significance of 3.4@xmath28 @xcite . on the experimental front
, new results are expected from the future belle ii detector at the superkekb @xmath29 collider , and in the near future also from lhcb at the large hadron collider .
the lhcb collaboration is currently analyzing the ratio of branching fractions of the baryonic @xmath30 and @xmath7 decays @xmath31 and @xmath32 , with the aim of determining @xmath12 for the first time at a hadron collider .
these decays were chosen over the more conventional @xmath33 and @xmath34 decays because , with the lhcb detector , final states containing protons are easier to identify than final states with pions @xcite .
note that the production rate of @xmath35 baryons at the lhc is remarkably high , equal to approximately @xmath36 times the production rate of @xmath37 mesons @xcite .
the extraction of @xmath10 and @xmath11 ( or their ratio ) from the measured @xmath31 and @xmath32 branching fractions requires knowledge of the form factors describing the @xmath38 and @xmath39 matrix elements of the relevant @xmath30 and @xmath7 currents in the weak effective hamiltonian .
these form factors have been studied using sum rules and quark models @xcite .
nonperturbative qcd calculations of the @xmath38 and @xmath39 form factors can be performed using lattice gauge theory .
the first lattice qcd calculation of @xmath38 form factors , published in ref .
@xcite , employed static @xmath4 quarks ( i.e. , leading - order heavy - quark effective theory ) to simplify the analysis .
the static limit reduces the number of independent @xmath38 form factors to two @xcite , but introduces systematic uncertainties of order @xmath40 and @xmath41 in the @xmath31 differential decay rate ( where @xmath42 is the momentum of the proton in the @xmath35 rest frame ) . here
we present a new lattice calculation which improves upon ref .
@xcite by replacing the static @xmath4 quarks by relativistic @xmath4 quarks , eliminating this systematic uncertainty .
in addition to the six form factors describing the hadronic part of the decay @xmath31 in fully relativistic qcd , we also compute the six analogous form factors for @xmath32 ( note that early lattice studies of @xmath39 form factors in the quenched approximation can be found in refs .
preliminary results from the present work were shown in ref .
@xcite . in sec .
[ sec : ffdefinitions ] we provide the definitions of the form factors employed here .
the lattice actions and parameters , as well as the matching of the @xmath30 and @xmath7 currents from the lattice renormalization scheme to the continuum @xmath43 scheme are discussed in sec .
[ sec : actions ] .
this calculation is based on the same lattice gauge - field ensembles as ref .
@xcite ; the ensembles include 2 + 1 flavor of dynamical domain - wall fermions and were generated by the rbc and ukqcd collaborations @xcite .
section [ sec : ratios ] explains our method for extracting the form factors from ratios of three - point and two - point correlation functions and removing excited - state contamination by extrapolating to infinite source - sink separation .
our fits of the quark - mass , lattice - spacing , and momentum - dependence of the form factors are discussed in sec .
[ sec : ccextrap ] .
the form factors in the physical limit are presented in terms of @xmath9-expansion @xcite parameters and their correlation matrices .
two different sets of parameters , referred to as the `` nominal parameters '' and the `` higher - order parameters '' are given .
the nominal parameters are used to obtain the central values and statistical uncertainties of the form factors ( and of derived quantities ) , while the higher - order parameters are used to calculate systematic uncertainties . in sec .
[ sec : dgamma ] we then present predictions for the @xmath0 and @xmath1 differential and integrated decay rates using our form factors . combined with experimental data , our results for
the @xmath44 and @xmath45 decay rates in the high-@xmath8 region will allow determinations of @xmath10 and @xmath11 with theory uncertainties of @xmath13 and @xmath14 , respectively .
allowing for possible right - handed currents beyond the standard model , the effective weak hamiltonian for @xmath46 transitions ( where @xmath47 ) can be written as @xmath48\ : \bar{\ell } \gamma_\mu ( 1-\gamma_5 ) \nu \label{eq : heff}\ ] ] ( in the standard model , @xmath49 and @xmath50 ) . to calculate the differential decay rate and other observables , we therefore need the hadronic matrix elements of the vector and axial vector currents , @xmath51 and @xmath52 . in the following ,
we denote the final - state baryon by @xmath53 ( @xmath54 ) .
lorentz and discrete symmetries imply that the matrix elements @xmath55 and @xmath56 can each be decomposed into three form factors . in this work we primarily use a helicity - based definition of the @xmath57 form factors , which was introduced in ref .
@xcite and is given by @xmath58 u_{\lambda_b}(p , s ) , \\
\nonumber \langle x(p^\prime , s^\prime ) | \overline{q } \,\gamma^\mu\gamma_5\ , b | \lambda_b(p , s ) \rangle & = & -\overline{u}_x(p^\prime , s^\prime ) \:\gamma_5 \bigg [ g_0(q^2)\ : ( m_{\lambda_b}+m_x)\frac{q^\mu}{q^2 } \\
\nonumber & & \phantom{\overline{u}_x \:\gamma_5\bigg[}+ g_+(q^2)\frac{m_{\lambda_b}-m_x}{s_-}\left ( p^\mu + p^{\prime \mu } - ( m_{\lambda_b}^2-m_x^2)\frac{q^\mu}{q^2 } \right ) \\ & & \phantom{\overline{u}_x\:\gamma_5 \bigg[}+ g_\perp(q^2 ) \left(\gamma^\mu + \frac{2m_x}{s_- } p^\mu - \frac{2 m_{\lambda_b}}{s_- } p^{\prime \mu } \right ) \bigg ] u_{\lambda_b}(p , s).\end{aligned}\ ] ] in these expressions , @xmath59 is the four - momentum transfer ( whereas @xmath60 is the @xmath61 or @xmath62 quark field ) , and @xmath63 is defined as @xmath64 the form factors with subscripts @xmath65 , @xmath66 , @xmath67 describe the contractions of the above matrix elements with virtual polarization vectors @xmath68 that are , respectively , time - like , longitudinal , and transverse to @xmath69 .
consequently , this choice of form factors leads to particularly simple expressions for observables such as the differential decay rate .
moreover , this choice simplifies the extraction of the form factors from correlation functions and clarifies the spin - parity quantum numbers of poles outside the physical kinematic region @xmath70 . an alternate definition of the form factors that can be found in the literature ( see , e.g. , ref .
@xcite ) is the following : @xmath71 u_{\lambda_b}(p , s ) , \label{eq : weinbergff1 } \\
\langle x(p^\prime , s^\prime ) | \overline{q } \,\gamma^\mu\gamma_5\ , b |
\lambda_b(p ) \rangle & = & \overline{u}_x(p^\prime , s^\prime ) \left [ f_1^a(q^2)\ : \gamma^\mu - \frac{f_2^a(q^2)}{m_{\lambda_b } } i\sigma^{\mu\nu}q_\nu + \frac{f_3^a(q^2)}{m_{\lambda_b } } q^\mu \right]\gamma_5\ : u_{\lambda_b}(p , s ) , \label{eq : weinbergff2}\end{aligned}\ ] ] where @xmath72 and , as before , @xmath59 .
this choice decomposes the matrix elements into form factors of the first and second class according to weinberg s classification @xcite .
the second - class form factors @xmath73 and @xmath74 would vanish in the limit @xmath75 ( for @xmath39 ) or @xmath76 ( for @xmath38 ) @xcite . in the following , we will refer to the form factors defined in eqs .
( [ eq : weinbergff1 ] ) , ( [ eq : weinbergff2 ] ) as `` weinberg form factors '' .
the helicity form factors are related to the weinberg form factors as follows : @xmath77 these relations also demonstrate the following endpoint constraints for the helicity form factors : @xmath78 where @xmath79 . at intermediate stages of our analysis of the lattice qcd data ,
it is beneficial to work with both definitions of the form factors .
however , we perform the chiral / continuum / kinematic extrapolations only in the helicity basis .
this calculation is based on lattice gauge field ensembles generated by the rbc and ukqcd collaborations @xcite with the iwasaki gauge action @xcite and 2 + 1 flavors of dynamical domain - wall fermions @xcite .
we implement the light ( @xmath80 or @xmath81 ) valence quarks with the same domain - wall action that was used in generating the ensembles .
our analysis uses six different combinations of light - quark masses and lattice spacings as shown in table [ tab : params ] .
these parameters are identical to those used in the earlier calculation of @xmath82 form factors in ref .
however , instead of the static eichten - hill action @xcite employed in ref .
@xcite , we now use anisotropic clover actions for the heavy ( @xmath5 and @xmath4 ) quarks @xcite .
these actions have the form @xmath83 q \ , , \label{eq : rhq}\ ] ] where @xmath84 is the lattice charm or bottom quark field , @xmath85 and @xmath86 are first- and second - order covariant lattice derivatives , and @xmath87 is a lattice expression for the field - strength tensor ( all of which are defined as in ref .
@xcite ) . by suitably tuning the parameters @xmath88 , @xmath89 , @xmath90 as functions of @xmath91 , heavy - quark discretization errors proportional to powers of @xmath91
can be removed to all orders .
the remaining discretization errors are of order @xmath92 , where @xmath93 is the typical magnitude of the spatial momentum of the heavy quark inside the hadron .
as the continuum limit @xmath94 is approached , the values @xmath95 and @xmath96 corresponding to the standard clover - improved wilson action are recovered .
for the bottom quark , we use the parameters that were tuned nonperturbatively by the rbc and ukqcd collaborations @xcite using the condition that the action reproduces the correct spin - averaged @xmath97 meson mass and relativistic dispersion relation , as well as the correct @xmath98 hyperfine splitting . for the charm quarks ,
we use the parameters from ref .
@xcite , where @xmath91 and @xmath88 were tuned nonperturbatively to obtain the correct spin - averaged charmonium mass and relativistic dispersion relation , while @xmath89 and @xmath90 were set to mean - field improved tree - level predictions .
note that after the parameters were tuned in this way , the calculated charmonium hyperfine splittings were also in agreement with experiment @xcite .
the values of all heavy - quark action parameters used here are given in table [ tab : hqparams ] .
.[tab : params ] parameters of the lattice gauge field ensembles @xcite and light - quark propagators @xcite .
the three groups of data sets @xmath99 , @xmath100 , and @xmath101 correspond to three different ensembles of lattice gauge fields : one with a `` coarse '' lattice spacing @xmath102 fm , and two with `` fine '' lattice spacings @xmath103 fm ( we use the lattice spacing values determined in ref .
@xcite ) . within each group ,
the valence - quark masses @xmath104 used for the propagators differ , resulting in different `` valence pion masses '' @xmath105 ; the number of light - quark propagators used in each data set is denoted as @xmath106 . [ cols="^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^",options="header " , ]
collaboration , k. olive _ et al .
_ , `` review of particle physics , '' http://dx.doi.org/10.1088/1674-1137/38/9/090001[chin.phys .
* c38 * ( 2014 ) 090001 ] .
j. a. bailey , c. bernard , c. e. detar , m. di pierro , a. el - khadra , _
et al . _ ,
`` the @xmath107 semileptonic form factor from three - flavor lattice qcd : a model - independent determination of @xmath10 , '' http://dx.doi.org/10.1103/physrevd.79.054507[phys.rev . * d79 * ( 2009 ) 054507 ] , http://arxiv.org/abs/0811.3640[arxiv:0811.3640 [ hep - lat ] ] . c. bernard , c. e. detar , m. di pierro , a. el - khadra , r. evans , _ et al .
_ , `` the @xmath108 form factor at zero recoil from three - flavor lattice qcd : a model independent determination of @xmath11 , '' http://dx.doi.org/10.1103/physrevd.79.014506[phys.rev .
* d79 * ( 2009 ) 014506 ] , http://arxiv.org/abs/0808.2519[arxiv:0808.2519 [ hep - lat ] ] .
collaboration , r. kowalewski , `` status of @xmath10 and @xmath11 determinations , '' pos * fpcp2010 * ( 2010 ) 028 . t. mannel , `` determination of @xmath10 and @xmath11 : a theory perspective , '' pos * fpcp2010 * ( 2010 ) 029 . g. ricciardi , `` status of @xmath11 and @xmath10 ckm matrix elements , '' http://arxiv.org/abs/1412.4288[arxiv:1412.4288 [ hep - ph ] ] .
chen and s .- h .
nam , `` left - right mixing on leptonic and semileptonic @xmath109 decays , '' http://dx.doi.org/10.1016/j.physletb.2008.07.095[phys.lett . *
b666 * ( 2008 ) 462466 ] , http://arxiv.org/abs/0807.0896[arxiv:0807.0896 [ hep - ph ] ] .
a. crivellin , `` effects of right - handed charged currents on the determinations of @xmath10 and @xmath11 , '' http://dx.doi.org/10.1103/physrevd.81.031301[phys.rev .
* d81 * ( 2010 ) 031301 ] , http://arxiv.org/abs/0907.2461[arxiv:0907.2461 [ hep - ph ] ] .
a. j. buras , k. gemmler , and g. isidori , `` quark flavour mixing with right - handed currents : an effective theory approach , '' http://dx.doi.org/10.1016/j.nuclphysb.2010.09.021[nucl.phys . *
b843 * ( 2011 ) 107142 ] , http://arxiv.org/abs/1007.1993[arxiv:1007.1993 [ hep - ph ] ] .
a. crivellin and s. pokorski , `` can the differences in the determinations of @xmath2 and @xmath3 be explained by new physics ?
, '' http://dx.doi.org/10.1103/physrevlett.114.011802[phys.rev.lett .
* 114 * ( 2015 ) 011802 ] , http://arxiv.org/abs/1407.1320[arxiv:1407.1320 [ hep - ph ] ] .
collaboration , j. a. bailey _ et al .
_ , `` @xmath10 from @xmath110 decays and ( 2 + 1)-flavor lattice qcd , '' http://arxiv.org/abs/1503.07839[arxiv:1503.07839 [ hep - lat ] ]
. j. flynn , t. izubuchi , t. kawanai , c. lehner , a. soni , _
et al . _ , `` the @xmath111 and @xmath112 form factors and @xmath10 from 2 + 1-flavor lattice qcd with domain - wall light quarks and relativistic heavy quarks , '' http://arxiv.org/abs/1501.05373[arxiv:1501.05373 [ hep - lat ] ] .
j. a. bailey , a. bazavov , c. bernard , c. bouchard , c. detar , _ et al .
_ , `` update of @xmath11 from the @xmath113 form factor at zero recoil with three - flavor lattice qcd , '' http://dx.doi.org/10.1103/physrevd.89.114504[phys.rev .
* d89 * no . 11 , ( 2014 ) 114504 ] , http://arxiv.org/abs/1403.0635[arxiv:1403.0635 [ hep - lat ] ] .
collaboration , j. lees _ et al .
_ , `` evidence for an excess of @xmath114 decays , '' http://dx.doi.org/10.1103/physrevlett.109.101802[phys.rev.lett . *
109 * ( 2012 ) 101802 ] , http://arxiv.org/abs/1205.5442[arxiv:1205.5442 [ hep - ex ] ] .
collaboration , m. adinolfi _ et al .
_ , `` performance of the lhcb rich detector at the lhc , '' http://dx.doi.org/10.1140/epjc/s10052-013-2431-9[eur.phys.j .
* c73 * ( 2013 ) 2431 ] , http://arxiv.org/abs/1211.6759[arxiv:1211.6759 [ physics.ins-det ] ] .
collaboration , r. aaij _ et al .
_ , `` study of the kinematic dependences of @xmath115 production in pp collisions and a measurement of the @xmath116 @xmath117 branching fraction , '' http://dx.doi.org/10.1007/jhep08(2014)143[jhep * 1408 * ( 2014 ) 143 ] , http://arxiv.org/abs/1405.6842[arxiv:1405.6842 [ hep - ex ] ] . f. cardarelli and s. simula , `` isgur - wise form - factors of heavy baryons within a light front constituent quark model , '' http://dx.doi.org/10.1016/s0370-2693(97)01581-5[phys.lett . * b421 * ( 1998 ) 295302 ] , http://arxiv.org/abs/hep-ph/9711207[arxiv:hep-ph/9711207 [ hep - ph ] ] . h. g. dosch , e. ferreira , m. nielsen , and r. rosenfeld , `` evidence from qcd sum rules for large violation of heavy quark symmetry in @xmath35 semileptonic decay , '' http://dx.doi.org/10.1016/s0370-2693(98)00566-8[phys.lett .
* b431 * ( 1998 ) 173178 ] , http://arxiv.org/abs/hep-ph/9712350[arxiv:hep-ph/9712350 [ hep - ph ] ] .
huang , c .- f .
qiao , and h .-
yan , `` decay @xmath118 in qcd sum rules , '' http://dx.doi.org/10.1016/s0370-2693(98)00909-5[phys.lett .
* b437 * ( 1998 ) 403407 ] , http://arxiv.org/abs/hep-ph/9805452[arxiv:hep-ph/9805452 [ hep - ph ] ] .
r. marques de carvalho , f. navarra , m. nielsen , e. ferreira , and h. g. dosch , `` form - factors and decay rates for heavy @xmath119 semileptonic decays from qcd sum rules , '' http://dx.doi.org/10.1103/physrevd.60.034009[phys.rev . *
d60 * ( 1999 ) 034009 ] , http://arxiv.org/abs/hep-ph/9903326[arxiv:hep-ph/9903326 [ hep - ph ] ] .
huang and d .- w .
wang , `` light cone qcd sum rules for the semileptonic decay @xmath120 , '' http://dx.doi.org/10.1103/physrevd.69.094003[phys.rev .
* d69 * ( 2004 ) 094003 ] , http://arxiv.org/abs/hep-ph/0401094[arxiv:hep-ph/0401094 [ hep - ph ] ] .
m. pervin , w. roberts , and s. capstick , `` semileptonic decays of heavy @xmath119 baryons in a quark model , '' http://dx.doi.org/10.1103/physrevc.72.035201[phys.rev .
* c72 * ( 2005 ) 035201 ] , http://arxiv.org/abs/nucl-th/0503030[arxiv:nucl-th/0503030 [ nucl - th ] ] .
ke , x .- q .
li , and z .- t . wei , `` diquarks and @xmath39 weak decays , '' http://dx.doi.org/10.1103/physrevd.77.014020[phys.rev . *
d77 * ( 2008 ) 014020 ] , http://arxiv.org/abs/0710.1927[arxiv:0710.1927 [ hep - ph ] ] . y .-
wang , y .-
shen , and c .- d .
lu , `` @xmath121 transition form factors from qcd light - cone sum rules , '' http://dx.doi.org/10.1103/physrevd.80.074012[phys.rev .
* d80 * ( 2009 ) 074012 ] , http://arxiv.org/abs/0907.4008[arxiv:0907.4008 [ hep - ph ] ] .
k. azizi , m. bayar , y. sarac , and h. sundu , `` semileptonic @xmath122 to nucleon transitions in full qcd at light cone , '' http://dx.doi.org/10.1103/physrevd.80.096007[phys.rev .
* d80 * ( 2009 ) 096007 ] , http://arxiv.org/abs/0908.1758[arxiv:0908.1758 [ hep - ph ] ] .
a. khodjamirian , c. klein , t. mannel , and y .- m .
wang , `` form factors and strong couplings of heavy baryons from qcd light - cone sum rules , '' http://dx.doi.org/10.1007/jhep09(2011)106[jhep * 1109 * ( 2011 ) 106 ] , http://arxiv.org/abs/1108.2971[arxiv:1108.2971 [ hep - ph ] ] .
t. gutsche , m. a. ivanov , j. g. k@xmath123rner , v. e. lyubovitskij , and p. santorelli , `` heavy - to - light semileptonic decays of @xmath35 and @xmath124 baryons in the covariant confined quark model , '' http://arxiv.org/abs/1410.6043[arxiv:1410.6043 [ hep - ph ] ] .
t. gutsche , m. a. ivanov , j. g. k@xmath123rner , v. e. lyubovitskij , p. santorelli , _
et al . _ , `` the semileptonic decay @xmath125 in the covariant confined quark model , '' http://arxiv.org/abs/1502.04864[arxiv:1502.04864 [ hep - ph ] ] . w. detmold , c .- j .
d. lin , s. meinel , and m. wingate , `` @xmath126 form factors from lattice qcd with static b quarks , '' http://dx.doi.org/10.1103/physrevd.88.014512[phys.rev .
* d88 * ( 2013 ) 014512 ] , http://arxiv.org/abs/1306.0446[arxiv:1306.0446 [ hep - lat ] ] .
t. mannel , w. roberts , and z. ryzak , `` baryons in the heavy quark effective theory , '' http://dx.doi.org/10.1016/0550-3213(91)90301-d[nucl.phys .
* b355 * ( 1991 ) 3853 ] .
f. hussain , j. k@xmath123rner , m. kramer , and g. thompson , `` on heavy baryon decay form - factors , '' http://dx.doi.org/10.1007/bf01475799[z.phys .
* c51 * ( 1991 ) 321328 ] .
f. hussain , d .- s .
liu , m. kramer , j. k@xmath123rner , and s. tawfiq , `` general analysis of weak decay form - factors in heavy to heavy and heavy to light baryon transitions , '' http://dx.doi.org/10.1016/0550-3213(92)90286-k[nucl.phys . * b370 * ( 1992 ) 259277 ] .
collaboration , k. bowler _
et al . _ , `` first lattice study of semileptonic decays of @xmath35 and @xmath127 baryons , '' http://dx.doi.org/10.1103/physrevd.57.6948[phys.rev .
* d57 * ( 1998 ) 69486974 ] , http://arxiv.org/abs/hep-lat/9709028[arxiv:hep-lat/9709028 [ hep - lat ] ] .
s. a. gottlieb and s. tamhankar , `` a lattice study of @xmath35 semileptonic decay , '' http://dx.doi.org/10.1016/s0920-5632(03)01612-8[nucl.phys.proc.suppl .
* 119 * ( 2003 ) 644646 ] , http://arxiv.org/abs/hep-lat/0301022[arxiv:hep-lat/0301022 [ hep - lat ] ] .
s. meinel , `` flavor physics with @xmath35 baryons , '' pos * lattice2013 * ( 2014 ) 024 , http://arxiv.org/abs/1401.2685[arxiv:1401.2685 [ hep - lat ] ] .
collaboration , y. aoki _ et al .
_ , `` continuum limit physics from 2 + 1 flavor domain wall qcd , '' http://dx.doi.org/10.1103/physrevd.83.074508[phys.rev
. * d83 * ( 2011 ) 074508 ] , http://arxiv.org/abs/1011.0892[arxiv:1011.0892 [ hep - lat ] ] .
c. bourrely , i. caprini , and l. lellouch , `` model - independent description of @xmath107 decays and a determination of @xmath10 , '' http://dx.doi.org/10.1103/physrevd.82.099902 , 10.1103/physrevd.79.013008[phys.rev . *
d79 * ( 2009 ) 013008 ] , http://arxiv.org/abs/0807.2722[arxiv:0807.2722 [ hep - ph ] ] .
t. feldmann and m. w. yip , `` form factors for @xmath128 transitions in scet , '' http://dx.doi.org/10.1103/physrevd.85.014035 , 10.1103/physrevd.86.079901[phys.rev .
* d85 * ( 2012 ) 014035 ] , http://arxiv.org/abs/1111.1844[arxiv:1111.1844 [ hep - ph ] ] .
s. weinberg , `` charge symmetry of weak interactions , '' http://dx.doi.org/10.1103/physrev.112.1375[phys.rev .
* 112 * ( 1958 ) 13751379 ] .
s. sasaki and t. yamazaki , `` lattice study of flavor su(3 ) breaking in hyperon beta decay , '' http://dx.doi.org/10.1103/physrevd.79.074508[phys.rev .
* d79 * ( 2009 ) 074508 ] , http://arxiv.org/abs/0811.1406[arxiv:0811.1406 [ hep - ph ] ] .
y. iwasaki , `` renormalization group analysis of lattice theories and improved lattice action .
2 . four - dimensional nonabelian @xmath129 gauge model , '' .
y. iwasaki and t. yoshie , `` renormalization group improved action for @xmath130 lattice gauge theory and the string tension , '' http://dx.doi.org/10.1016/0370-2693(84)91500-4[phys.lett . * b143 * ( 1984 ) 449 ] . d. b. kaplan , `` a method for simulating chiral fermions on the lattice , '' http://dx.doi.org/10.1016/0370-2693(92)91112-m[phys.lett .
* b288 * ( 1992 ) 342347 ] , http://arxiv.org/abs/hep-lat/9206013[arxiv:hep-lat/9206013 ] .
v. furman and y. shamir , `` axial symmetries in lattice qcd with kaplan fermions , '' http://dx.doi.org/10.1016/0550-3213(95)00031-m[nucl.phys .
* b439 * ( 1995 ) 5478 ] , http://arxiv.org/abs/hep-lat/9405004[arxiv:hep-lat/9405004 ] .
y. shamir , `` chiral fermions from lattice boundaries , '' http://dx.doi.org/10.1016/0550-3213(93)90162-i[nucl.phys .
* b406 * ( 1993 ) 90106 ] , http://arxiv.org/abs/hep-lat/9303005[arxiv:hep-lat/9303005 ] .
e. eichten and b. r. hill , `` renormalization of heavy - light bilinears and @xmath131 for wilson fermions , '' http://dx.doi.org/10.1016/0370-2693(90)90432-6[phys.lett .
* b240 * ( 1990 ) 193 ] . a. x. el - khadra , a. s. kronfeld , and p. b. mackenzie , `` massive fermions in lattice gauge theory , '' http://dx.doi.org/10.1103/physrevd.55.3933[phys.rev .
* d55 * ( 1997 ) 39333957 ] , http://arxiv.org/abs/hep-lat/9604004[arxiv:hep-lat/9604004 [ hep - lat ] ] .
s. aoki , y. kuramashi , and s .-
tominaga , `` relativistic heavy quarks on the lattice , '' http://dx.doi.org/10.1143/ptp.109.383[prog.theor.phys .
* 109 * ( 2003 ) 383413 ] , http://arxiv.org/abs/hep-lat/0107009[arxiv:hep-lat/0107009 [ hep - lat ] ] .
s. aoki , y. kayaba , and y. kuramashi , `` a perturbative determination of mass dependent @xmath132 improvement coefficients in a relativistic heavy quark action , '' http://dx.doi.org/10.1016/j.nuclphysb.2004.07.017[nucl.phys .
* b697 * ( 2004 ) 271301 ] , http://arxiv.org/abs/hep-lat/0309161[arxiv:hep-lat/0309161 [ hep - lat ] ] .
lin and n. christ , `` non - perturbatively determined relativistic heavy quark action , '' http://dx.doi.org/10.1103/physrevd.76.074506[phys.rev . * d76 * ( 2007 ) 074506 ] , http://arxiv.org/abs/hep-lat/0608005[arxiv:hep-lat/0608005 [ hep - lat ] ] .
collaboration , y. aoki _
et al . _ ,
`` nonperturbative tuning of an improved relativistic heavy - quark action with application to bottom spectroscopy , '' http://dx.doi.org/10.1103/physrevd.86.116003[phys.rev .
* d86 * ( 2012 ) 116003 ] , http://arxiv.org/abs/1206.2554[arxiv:1206.2554 [ hep - lat ] ] .
z. s. brown , w. detmold , s. meinel , and k. orginos , `` charmed bottom baryon spectroscopy from lattice qcd , '' http://dx.doi.org/10.1103/physrevd.90.094507[phys.rev . * d90 * ( 2014 ) 094507 ] , http://arxiv.org/abs/1409.0497[arxiv:1409.0497 [ hep - lat ] ] .
w. detmold , c .- j .
d. lin , s. meinel , and m. wingate , `` @xmath133 form factors and differential branching fraction from lattice qcd , '' http://dx.doi.org/10.1103/physrevd.87.074502[phys . rev .
d 87 , * 074502 * ( 2013 ) ] , http://arxiv.org/abs/1212.4827[arxiv:1212.4827 [ hep - lat ] ]
. s. meinel , `` bottomonium spectrum at order @xmath134 from domain - wall lattice qcd : precise results for hyperfine splittings , '' http://dx.doi.org/10.1103/physrevd.82.114502[phys.rev . *
d82 * ( 2010 ) 114502 ] , http://arxiv.org/abs/1007.3966[arxiv:1007.3966 [ hep - lat ] ] .
s. hashimoto , a. x. el - khadra , a. s. kronfeld , p. b. mackenzie , s. m. ryan , _
et al . _ ,
`` lattice qcd calculation of @xmath135 decay form - factors at zero recoil , '' http://dx.doi.org/10.1103/physrevd.61.014502[phys.rev .
* d61 * ( 1999 ) 014502 ] , http://arxiv.org/abs/hep-ph/9906376[arxiv:hep-ph/9906376 [ hep - ph ] ] .
a. x. el - khadra , a. s. kronfeld , p. b. mackenzie , s. m. ryan , and j. n. simone , `` the semileptonic decays @xmath107 and @xmath136 from lattice qcd , '' http://dx.doi.org/10.1103/physrevd.64.014502[phys.rev . *
d64 * ( 2001 ) 014502 ] , http://arxiv.org/abs/hep-ph/0101023[arxiv:hep-ph/0101023 [ hep - ph ] ] . c. lehner , `` automated lattice perturbation theory and relativistic heavy quarks in the columbia formulation , '' pos * lattice2012 * ( 2012 ) 126 , http://arxiv.org/abs/1211.4013[arxiv:1211.4013 [ hep - lat ] ] . c. lehner ,
`` physyhcal . ''
/ physyhcal/. collaboration , j. a. bailey _
et al . _ ,
`` the @xmath137 form factors at nonzero recoil and @xmath11 from @xmath138-flavor lattice qcd , '' http://arxiv.org/abs/1503.07237[arxiv:1503.07237 [ hep - lat ] ] .
n. h. christ , j. m. flynn , t. izubuchi , t. kawanai , c. lehner , _ et al .
_ , `` b - meson decay constants from 2 + 1-flavor lattice qcd with domain - wall light quarks and relativistic heavy quarks , '' http://arxiv.org/abs/1404.4670[arxiv:1404.4670 [ hep - lat ] ]
. w. detmold , c. d. lin , and s. meinel , `` calculation of the heavy - hadron axial couplings @xmath139 , @xmath140 , and @xmath141 using lattice qcd , '' http://dx.doi.org/10.1103/physrevd.85.114508[phys.rev . *
d85 * ( 2012 ) 114508 ] , http://arxiv.org/abs/1203.3378[arxiv:1203.3378 [ hep - lat ] ] . c. morningstar and m. j. peardon , `` analytic smearing of @xmath130 link variables in lattice qcd , '' http://dx.doi.org/10.1103/physrevd.69.054501[phys.rev . *
d69 * ( 2004 ) 054501 ] , http://arxiv.org/abs/hep-lat/0311018[arxiv:hep-lat/0311018 [ hep - lat ] ] .
g. colangelo , s. durr , a. juttner , l. lellouch , h. leutwyler , _
et al . _ , `` review of lattice results concerning low energy particle physics , '' http://dx.doi.org/10.1140/epjc/s10052-011-1695-1[eur.phys.j .
* c71 * ( 2011 ) 1695 ] , http://arxiv.org/abs/1011.4408[arxiv:1011.4408 [ hep - lat ] ] .
s. r. beane , `` nucleon masses and magnetic moments in a finite volume , '' http://dx.doi.org/10.1103/physrevd.70.034507[phys.rev . *
d70 * ( 2004 ) 034507 ] , http://arxiv.org/abs/hep-lat/0403015[arxiv:hep-lat/0403015 [ hep - lat ] ] .
s. r. beane and m. j. savage , `` baryon axial charge in a finite volume , '' http://dx.doi.org/10.1103/physrevd.70.074029[phys.rev .
* d70 * ( 2004 ) 074029 ] , http://arxiv.org/abs/hep-ph/0404131[arxiv:hep-ph/0404131 [ hep - ph ] ] .
w. detmold , c .- j .
d. lin , and s. meinel , `` axial couplings in heavy hadron chiral perturbation theory at the next - to - leading order , '' http://dx.doi.org/10.1103/physrevd.84.094502[phys.rev .
* d84 * ( 2011 ) 094502 ] , http://arxiv.org/abs/1108.5594[arxiv:1108.5594 [ hep - lat ] ] .
a. s. kronfeld , `` application of heavy quark effective theory to lattice qcd .
power corrections , '' http://dx.doi.org/10.1103/physrevd.62.014505[phys.rev . * d62 * ( 2000 ) 014505 ] , http://arxiv.org/abs/hep-lat/0002008[arxiv:hep-lat/0002008 [ hep - lat ] ] . j. harada , s. hashimoto , k .-
ishikawa , a. s. kronfeld , t. onogi , _
et al . _ ,
`` application of heavy quark effective theory to lattice qcd .
radiative corrections to heavy light currents , '' http://dx.doi.org/10.1103/physrevd.71.019903 , 10.1103/physrevd.65.094513[phys.rev .
* d65 * ( 2002 ) 094513 ] , http://arxiv.org/abs/hep-lat/0112044[arxiv:hep-lat/0112044 [ hep - lat ] ] .
j. harada , s. hashimoto , a. s. kronfeld , and t. onogi , `` application of heavy quark effective theory to lattice qcd .
radiative corrections to heavy - heavy currents , '' http://dx.doi.org/10.1103/physrevd.65.094514[phys.rev .
* d65 * ( 2002 ) 094514 ] , http://arxiv.org/abs/hep-lat/0112045[arxiv:hep-lat/0112045 [ hep - lat ] ] .
s. shivashankara , w. wu , and a. datta , `` @xmath142 decay in the standard model and with new physics , '' http://arxiv.org/abs/1502.07230[arxiv:1502.07230 [ hep - ph ] ] .
collaboration , r. g. edwards and b. joo , `` the chroma software system for lattice
qcd , '' http://dx.doi.org/10.1016/j.nuclphysbps.2004.11.254[nucl.phys.proc.suppl .
* 140 * ( 2005 ) 832 ] , http://arxiv.org/abs/hep-lat/0409003[arxiv:hep-lat/0409003 [ hep - lat ] ] .
collaboration , r. aaij _ et al .
_ , `` determination of the quark coupling strength @xmath10 using baryonic decays , '' http://arxiv.org/abs/1504.01568[arxiv:1504.01568 [ hep - ex ] ] . | measurements of the @xmath0 and @xmath1 decay rates can be used to determine the magnitudes of the ckm matrix elements @xmath2 and @xmath3 , provided that the relevant hadronic form factors are known . here
we present a precise calculation of these form factors using lattice qcd with 2 + 1 flavors of dynamical domain - wall fermions .
the @xmath4 and @xmath5 quarks are implemented with relativistic heavy - quark actions , allowing us to work directly at the physical heavy - quark masses .
the lattice computation is performed for six different pion masses and two different lattice spacings , using gauge - field configurations generated by the rbc and ukqcd collaborations .
the @xmath6 and @xmath7 currents are renormalized with a mostly nonperturbative method .
we extrapolate the form factor results to the physical pion mass and the continuum limit , parametrizing the @xmath8-dependence using @xmath9-expansions .
the form factors are presented in such a way as to enable the correlated propagation of both statistical and systematic uncertainties into derived quantities such as differential decay rates and asymmetries . using these form factors , we present predictions for the @xmath0 and @xmath1 differential and integrated decay rates .
combined with experimental data , our results enable determinations of @xmath10 , @xmath11 , and @xmath12 with theory uncertainties of @xmath13 , @xmath14 , and @xmath15 , respectively .
= 1 |
the three - dimensional distribution of galaxies in the redshift surveys differ from the true one since the distance to each galaxy can not be determined by its redshift @xmath2 only ; for @xmath3 the peculiar velocity of galaxies , typically @xmath4 , contaminates the true recession velocity of the hubble flow , while the true distance for objects at @xmath5 sensitively depends on the ( unknown and thus assumed ) cosmological parameters .
this hampers the effort to understand the true distribution of large - scale structure of the universe .
nevertheless such redshift - space distortion effects are quite useful since through the detailed theoretical modelling , one can derive the peculiar velocity dispersions of galaxies as a function of separation , and also can infer the cosmological density parameter @xmath0 and the dimensionless cosmological constant @xmath1 , for instance . in
what follows , i will present two specific topics concerning the redshift distortion ; the small - scale pair - wise peculiar velocity dispersions of galaxies@xcite , and anisotropies in the two - point correlation functions@xcite at high redshifts .
assuming that particle pairs in expanding universes are in `` statistical equilibrium '' on small scales , their relative peculiar velocity dispersion can be computed as a function of their separation .
the result is called the cosmic virial theorem @xcite ( cvt , hereafter ) which predicts the one - dimensional pair - wise relative peculiar velocity dispersion @xmath6 as a function of @xmath0 . the observed two- and three - point correlation functions of _ galaxies _ , @xmath7 and @xmath8 , are well approximated by the simple form@xcite : @xmath9\end{aligned}\ ] ] where @xmath10 , @xmath11 , @xmath12 .
thus it is reasonable to assume that the two- and three - point correlation functions of _ mass _ , @xmath13 and @xmath14 , also obey the same scaling except for the overall amplitude : @xmath15 then the cvt prediction of the small - scale peculiar velocity dispersion is given by@xcite @xmath16 \right\ } , \end{aligned}\ ] ] and numerically @xmath17 , @xmath18 , and @xmath19 .
to what extent is the cvt prediction reliable ? in order to examine this , i compared @xmath20 with the one - dimensional peculiar velocity dispersions of particle pairs with separation @xmath21 ^ 2 \rangle^{1/2 } $ ] .
figure [ fig : ptpcvt ] summarizes the comparison which implies that @xmath20 reproduces the simulation result excellently for @xmath22 ; cvt is quite reliable in predicting @xmath23 for @xmath24 .
it should be stressed that the crucial assumption in deriving the prediction ( [ eq : cvt1d ] ) is equation ( [ eq : hier ] ) , and the result is independent of the theoretical model for dark matter . in this sense
the prediction ( [ eq : cvt1d ] ) is general , and the good agreement in cdm models should be ascribed to the fact that the cdm models actually satisfy the relation ( [ eq : cvt1d ] ) . using the anisotropy of the two - point correlation function of the cfa1 galaxy redshift survey , davis and peebles@xcite estimated the line of sight peculiar velocity dispersions @xmath25 of galaxy pairs
seen projected at separation @xmath26 .
this is related to the dispersions @xmath27 for galaxy pairs with separation @xmath28 in three dimension as @xmath29 note that davis and peebles@xcite found that @xmath30 which justifies our assumption of the _ mass _ correlation function @xmath31 as @xmath32 from a dynamical point of view . if @xmath33 and @xmath34 , the corresponding cvt estimator corrected for the projection becomes @xmath35 where @xmath36 is the gamma function .
thus @xmath37 is the correction factor for the projection effect@xcite ; @xmath38 , @xmath39 , @xmath40 .
independent estimates of @xmath41 are available from several galaxy redshift catalogues@xcite .
as pointed out earlier by mo , jing , and brner@xcite , the currently available redshift catalogues are far from the fair sample , and the observational estimate of @xmath25 may differ from its true cosmic average due to the limited survey volume and the selection effect . therefore it is meaningful to re - examine the problem in more details .
if the hierarchical relation ( [ eq : hier ] ) holds as is indicated from the galaxy distribution , the amplitudes of the two- and three - point correlation functions , or equivalently @xmath42 and @xmath43 , are the two uncertain parameters in the cvt prediction ( [ eq : cvt1d ] ) .
recent numerical and analytical studies in nonlinear gravitational clustering seem to indicate that @xmath43 can be approximated as a constant in the range of @xmath44 and @xmath45 which depends very weakly on the underlying cosmological model . thus if the value of @xmath42 is fixed from the cobe data assuming cdm models , for example , the cvt prediction ( [ eq : cvt1d ] ) is completely specified .
figure [ fig : cobesigma8 ] plots the empirical fit for the cobe normalized @xmath46 and @xmath47 by nakamura@xcite to the numerical computation by sugiyama@xcite .
there i adopt the baryon density parameter @xmath48 .
incidentally it is amusing to note that @xmath49 , @xmath50 , and @xmath51 cdm model just corresponds to @xmath52 , i.e. , _ galaxies faithfully trace mass _ in this model according to the cobe normalization . once the @xmath42 is specified , it is easy to compute the cvt prediction ( [ eq : cvt1d ] ) .
figure [ fig : sigma12rp ] plots @xmath53 ( eq.[[eq : cvtproj ] ] ) as a function of @xmath0 .
the symbols denote several recent observational estimates which are shifted along the x - axis just for illustration .
the left panel shows the dependence on @xmath43 and the right panel shows how the cvt prediction is sensitive to their _ local _ value , or sample - to - sample variation of the data .
figure [ fig : sigma12rp ] exhibits that both observational estimates and cvt predictions of @xmath54 vary by a factor of @xmath55 depending on the local clustering degree of each sample . even allowing for these
, however , it looks unlikely that high - density cdm models ( @xmath56 ) are compatible with the currently observed values of @xmath54 . in a similar argument , one may obtain strong upper limits on @xmath0 in general cosmological scenarios as long as @xmath42 is not much greater than unity .
as shown in the previous section , small - scale velocity dispersions of galaxies place strong upper limits on @xmath0 , fairly independently of @xmath1 .
in turn , detailed analysis of clustering of objects at high redshifts will place a potentially important constraint on @xmath1 via an effect which we call the _ cosmological redshift distortion_@xcite . a very similar idea
was put forward independently by ballinger , peacock and heavens@xcite although they work entirely in k - space .
the result that i describe below considers the analysis of two - point correlation functions , which would be more straightforward to derived from the quasar catalogues .
let us consider a pair of objects located at redshifts @xmath57 and @xmath58 whose redshift difference @xmath59 is much less than the mean redshift @xmath60 .
then the observable separations of the pair parallel and perpendicular to the line - of - sight direction , @xmath61 and @xmath62 , are given as @xmath63 and @xmath64 , respectively , where @xmath65 is the hubble constant and @xmath66 denotes the angular separation of the pair on the sky .
the cosmological redshift - space distortion originates from the anisotropic mapping between the redshift - space coordinates , @xmath67 , and the real comoving ones@xcite , @xmath68 ; @xmath69 is written in terms of the angular diameter distance @xmath70 as @xmath71 , and @xmath72 the relation between the two - point correlation functions of quasars in redshift space , @xmath73 , and that of _ mass _ in real space @xmath74 can be derived in linear theory@xcite : @xmath75 ^
2\right ) \xi_0(x ) p_0(\mu ) -\left({4\over 3}\beta(z ) + { 4\over7}[\beta(z)]^2\right)\xi_2(x ) p_2(\mu ) \nonumber\\ & & \qquad\qquad + \frac{8}{35}[\beta(z)]^2 \xi_4(x ) p_4(\mu ) , \end{aligned}\ ] ] where @xmath76 , @xmath77 , @xmath78 s are the legendre polynomials , @xmath79 and @xmath80 is the linear growth rate . for specific examples , we compute @xmath81 in linear theory applying equations ( [ eq : xis ] ) and ( [ eq : xi2l ] ) in cdm models with @xmath82 km@xmath83s@xmath84mpc@xmath85@xcite .
the resulting contours are plotted in figure [ fig : xicont ] .
the four sets of values of @xmath0 and @xmath1 are indicated at the top of each panel .
we adopt the cobe normalization@xcite . in order to quantify the cosmological redshift distortion in figure [ fig : xicont ] ,
let us introduce the anisotropy parameter @xmath86 , where @xmath87 and @xmath88 .
the left four panels in figure [ fig : aniso ] show the anisotropy parameter against @xmath2 in cdm models .
this clearly exhibits the extent to which one can discriminate the different @xmath1 models on the basis of the anisotropies in @xmath89 at high redshifts .
one can roughly divide the source of cosmological information into four regimes ; ( i ) linear regime ( @xmath90 ) at @xmath91 , ( ii ) nonlinear regime ( @xmath92 ) at @xmath91 , ( iii ) linear regime at @xmath93 , and ( iv ) nonlinear regime at @xmath93 .
the clustering feature at @xmath91 is best probed by galaxy redshift surveys , and the current samples , albeit statistically limited by the number of galaxies , are heavily used for a variety of cosmological tests .
the conventional redshift - space distortion analysis@xcite in the regime ( i ) yields generally @xmath94 ; @xmath95 for instance from the durham / ukst galaxy redshift survey of @xmath96 galaxies .
i have shown in the first part of the talk that the small - scale peculiar velocity dispersions of galaxies in comparison withe the cvt prediction constrains the value of @xmath97 . allowing for the sample - to - sample variation of the available redshift surveys and the uncertainties for @xmath43 and @xmath98 ,
i still conclude that the observation in the regime ( ii ) favors low - density universes @xmath99 at most .
in contrast to the regimes ( i ) and ( ii ) , the proper analysis of the regimes ( iii ) and ( iv ) requires a large number of quasars homogeneously samples which are not currently available .
therefore the cosmological tests in theses regimes have not been fully explored even theoretically . in the second part of the talk , i presented an example in this line of investigations , cosmological redshift - space distortion@xcite .
i derived the formula which describes the degree of the anisotropies of two - point correlation functions of quasars in linear regime ( iii ) , and argued that it is a potentially powerful discriminator of @xmath1 once @xmath0 is determined from the nearby observation . from the observational point of view , however , it would be much easier to detect the the anisotropies of two - point correlation functions in the regime ( iv ) .
we are currently approaching this important area of research using the nonlinear theory and simulations@xcite .
the second part of the present talk is based on my collaborative work with takahiko matsubara .
i thank him for the fruitful and enjoyable collaboration , and yi - peng jing , takahiro t. nakamura and david weinberg for discussions .
i am grateful to john peacock for calling my attention to their work@xcite prior to publication after we submitted the paper@xcite of the second part of the present talk .
this research was supported in part by the grant - in - aid by the ministry of education , science , sports and culture of japan ( 07ce2002 ) to resceu ( research center for the early universe ) , the university of tokyo .
i thank the session organizers , chul hoon lee and katsuhiko sato , for inviting me to give the present talk at gravitation and cosmology session of the apctp inauguration conference . | i discuss two cosmological tests to determine the cosmological density parameter @xmath0 the cosmological constant @xmath1 , which make use of the anisotropy of the two - point correlation functions due to the peculiar velocity field and the cosmological redshift - space distortion . *
@xmath0 and @xmath1 from galaxy and quasar clustering : + cosmic virial theorem and cosmological redshift - space distortion * yasushi suto |
low energy supersymmetry ( susy)@xcite has been one of the most promising candidates for a theory of fundamental particles and interactions going beyond the standard model ( sm ) ; the so - called bsm physics .
the minimal extension of the sm including susy , namely , the minimal supersymmetric standard model ( mssm ) , extends the particle spectrum of the sm by one additional higgs doublet and the supersymmetric partners of all the sm particles - the * s*particles .
the supersymmetric extension of the sm provides a particularly elegant solution to the problem of stabilizing the electroweak ( ew ) symmetry breaking scale against large radiative correction and keeps the higgs naturally light .
in fact , a very robust upper limit on the mass of the lightest higgs boson is perhaps one of the important predictions of this theory .
further , this upper limit is linked in an essential way to the values of some of the susy breaking parameters in the theory . in addition , in @xmath10parity conserving susy , the lightest supersymmetric particle ( lsp ) emerges as the natural candidate for the dark matter ( dm),the existence of which has been proved beyond any doubt in astrophysical experiments . the search for evidence of the realization of this symmetry in nature(in the context of high energy collider experiments , precision measurements at the high intensity b factories and in the dm detection experiments ) has therefore received enormous attention of particle physicists , perhaps only next to the higgs boson .
the recent observation of a boson with mass around @xmath11 gev at the large hadron collider ( lhc)@xcite and the rather strong lower limits on the masses of sparticles that possess strong interactions that the lhc searches have yielded@xcite , necessitates careful studies of the mssm in the context of all the recent low energy data . in these studies ,
it is also very important to seek suitable guiding principles which could possibly reduce the associated large number of susy breaking parameters of mssm .
thus , looking for modes of specific susy breaking mechanisms that involve only a few input quantities given at a relevant scale can be useful . here , the soft susy breaking parameters at the electroweak scale are found via renormalization group ( rg ) analyses . apart from the simplicity of having a few parameters as input ,
such schemes create challenging balancing acts . on the one hand , various soft breaking masses and couplings become correlated with one another in such schemes . on the other hand ,
overall one has to accommodate a large number of very stringent low energy constraints in a comprehensive model consisting of only a few parameters .
a simple and well - motivated example of a susy model is the minimal supergravity ( msugra)@xcite . here , susy is broken spontaneously in a hidden sector and the breaking is communicated to the observable sector where mssm resides via planck mass suppressed supergravity interactions .
the model involves soft - susy breaking parameters like ( i ) universal gaugino mass parameter @xmath12 , ( ii ) the universal scalar mass parameter @xmath13 , ( iii ) the universal trilinear coupling @xmath14 , ( iv ) the universal bilinear coupling @xmath15 , all given at the gauge coupling unification scale .
in addition to it , one has the superpotential related higgsino mixing parameter @xmath16 with its associated sign parameter .
the two radiative electroweak symmetry breaking ( rewsb ) conditions may then be used so as to replace @xmath15 and @xmath16 with the z - boson mass @xmath17 and @xmath18 , the ratio of higgs vacuum expectation values . similar to msugra , one has other susy breaking scenarios like the gauge mediated susy breaking and models with anomaly mediated susy breaking ( amsb ) etc@xcite . apart from direct collider physics data , one has to satisfy constraints from flavor changing neutral current ( fcnc ) as well as flavor conserving phenomena like the anomalous magnetic moment of muon , constraints like electric dipole moments associated with cp violations or to check whether there is a proper amount of dark matter content in an r - parity conserving scenarios or proper neutrino masses in r - parity violating scenarios@xcite .
a single model is yet to be found that can adequately explain various stringent experimental results and at the same time possesses a sufficient degree of predictiveness .
howevr , it is always important to continue the quest of a simple model and check the degree of agreement with low energy constraints . in this work ,
we pursue a predictive theory of susy breaking by considering a field theory on a superspace where the grassmannian coordinates are essentially fluctuating / stochastic@xcite .
we note that in a given susy breaking scenario , our limitation of knowing the actual mechanism of breaking susy is manifested in the soft parameters . here , in stochastic superspace framework we assume that a manifestation of an unknown but a fundamental mechanism of susy breaking may effectively lead to stochasticity in the grassmannian parameters of the superspace . with a suitably chosen probability distribution
, this causes a given khler potential and a superpotential to lead to soft breaking terms that carry signatures of the stochasticity .
as we will see , the susy breaking is parametrized by @xmath1 which is nothing but @xmath19 , where the symbol @xmath20 refers to averaging over the grassmannian coordinates .
the other scale that is involved is @xmath3 .
values of various soft parameters at this scale are the input parameters of the scheme .
the values of the same at the electroweak scale are then obtained from these input values by using the renormalization group evolution . considering the superpotential of mssm ,
the soft terms obtained are readily recognized as the ones supplied by the externally added soft susy breaking terms of constrained mssm ( cmssm)@xcite , except that the model is unable to produce a scalar mass soft term@xcite .
reference @xcite used @xmath3 and @xmath1 as free parameters while analyzing the low energy signatures within mssm .
@xmath3 was chosen between @xmath4 to @xmath21 , the scale of gauge coupling unification and the planck mass scale respectively .
the model as given in ref.@xcite is called as stochastic supersymmetric model ( ssm ) and it is characterized by universal gaugino mass parameter @xmath12 , universal trilinear soft susy breaking parameter @xmath14 , and universal bilinear soft susy breaking parameter @xmath15 , all being related to @xmath1 , the parameter related to susy breaking .
we note that with the bilinear soft susy parameter being given , @xmath22 , becomes a derived quantity .
however , as already mentioned , in spite of the fact that the ssm generates soft susy breaking terms , it produces no scalar mass soft term .
scalar masses start from zero at the scale @xmath3 and renormalization group evolution is used to generate scalar masses at the electroweak scale @xmath23 .
scalar masses at @xmath23 severely constrain the model because typically scalars are very light . in particular , quite often sleptons turn to be the lightest supersymmetric particles or even become tachyonic over a large part of the parameter space .
this is partially ameliorated when one takes the high scale @xmath3 to be larger than the gauge coupling unification scale @xmath24 gev .
however , in spite of obtaining valid parameter space that would provide us with a lightest neutralino as a possible dark matter candidate in r - parity preserving framework , we must note that the low values that one obtains for the masses of the first two generation of squarks are hardly something of an advantage in view of the constraints coming from fcnc as well as those from the lhc data@xcite . on the other hand
, ssm has a natural advantage of being associated with a nonvanishing trilinear coupling that is favorable to produce a relatively light spectra for a given value of higgs boson mass @xmath2 .
the recent announcement from the cms and atlas collaborations of the lhc experiment about the discovery of a higgs - like boson at @xmath11 gev@xcite thus makes this model potentially attractive .
however , as explored in ref.@xcite , ssm as such is unable to accommodate such a large @xmath2 in spite of having a built - in feature of having a nonvanishing @xmath14 .
it could at most reach 116 gev for @xmath2@xcite and the constraint due to @xmath25 as used in ref.@xcite was much less stringent in comparison to the same of present day@xcite .
furthermore , it is also important to investigate the effect of the direct detection rate of dark matter as constrained by the recent xenon100 data@xcite . in this analysis
, we would like to give all the input parameters at the grand unification scale @xmath4 , the scale at which the standard model gauge group , namely , @xmath26 , comes into existence .
any evolution above @xmath4 would obviously demand choosing a suitable gauge group ; a question that is not going to be addressed in this work . in this way we would like to avoid unknown issues arising out of a post - grand unified theory(gut)@xcite physics .
however , we would rather try to meet the phenomenological demand of confronting the issue of sleptons becoming tachyonic or avoiding scalar masses to become light in general in a _ minimal modification _ by considering an externally given scalar mass soft parameter @xmath27 as a manifestation of an additional origin of susy breaking
. it would be useful to have the first two generations of scalar masses adequately heavy so as to overcome the fcnc related constraints and lhc data@xcite on squark masses .
additionally this will also be consistent with having the lighter higgs boson mass ( @xmath2 ) to be in the vicinity of 125 gev .
we will henceforth denote the model as * mod - ssm*. we may note that traditionally minimal versions of models of susy breaking have been extended for phenomenological reasons .
it is also true that extending a minimal model often lowers predictiveness and may even cause partial dilution of the main motivations associated with the building of the model .
for example , considering nonuniversal gaugino or scalar mass scenarios may be more suitable than cmssm or msugra so as to obtain a relatively lighter spectra in the context feasibility of exploring via lhc .
another example may be given in the context of the minimal amsb model .
as we know , a pure amsb scenario@xcite is associated with form invariance of the renormalization group equations ( rge ) of scalar masses and absence of flavor violation .
however , it produces tachyonic sleptons . in the minimal amsb model@xcite
one introduces an additional common mass parameter @xmath27 for all the scalars of the theory .
this ameliorates the tachyonic slepton problem but it is true that we sacrifice the much cherished feature of form invariance and accept some degree of flavor violations at the end .
we would like to explore a non - minimal scenario of stochastic supersymmetry model , namely , mod - ssm in this spirit .
we particularly keep in mind that the stochastic supersymmetry formalism may be used not only within the mssm framework but it may be extended to superpotentials beyond that of the mssm@xcite .
the fact that the model with a minimal modification can easily accommodate the recent higgs boson mass range by its generic feature of having a nonvanishing trilinear coupling parameter makes it further attractive .
we believe that our approach of considering an additional susy breaking scalar mass term is justified for phenomenological reasons .
thus , in mod - ssm we consider the input of soft term parameters @xmath12 , @xmath15 and @xmath14 ( all are either proportional to @xmath28 or @xmath29 ) and a universal scalar mass soft parameter @xmath27 , all being given at a suitable scale which we simply choose as the gauge coupling unification scale @xmath4 considering the lep data on gauge couplings could be a hint of the existence of grand unification . as in ref .
@xcite we would also restrict @xmath1 to be real , either positive or negative .
we clearly like to emphasize that the ssm framework produces no scalar mass soft term at a scale @xmath3 . with @xmath3
set to the gauge coupling unification scale @xmath4 , ssm is plagued with tachyonic sleptons . with @xmath30 one
can avoid tachyonic scalar but there is no scope of obtaining the currently accepted higgs boson mass . in mod - ssm
we consider @xmath5 and add the extra scalar mass term for an additional susy breaking effect . with the trilinear coupling parameter @xmath14 and the bilinear coupling parameter @xmath15 becoming correlated with @xmath12 , mod - ssm parameter space is essentially a subset of the one for the constrained mssm ( cmssm ) . in mod - ssm ,
the ssm inspired nonvanishing @xmath14 parameter is suitable for producing an appropriately large loop correction to the lighter higgs boson mass while keeping the overall sparticle spectra relatively at a lower range .
cmssm , on the other hand , does not pinpoint with any fundamental physical distinction / motivation for such a zone of parameter space that is capable of producing the correct higgs mass with similarly smaller sparticle mass scale .
even if we consider the stochastic superspace model in its original form as only a toy idea , we believe that it may be worthwhile to explore it with minimal modifications in regard to the current higgs boson mass , as well as other phenomenological constraints .
as seen in ref.@xcite we consider an @xmath0 superspace where the grassmannian coordinates @xmath31 and @xmath32 are taken to be stochastic in nature .
one starts with identifying the terms involving superfields in the superpotential and the kinetic energy terms that could be used to construct the susy invariant lagrangian density for a given model .
each term is then multiplied with a probability distribution function @xmath33 and integrated over the grassmannian coordinates appropriately .
@xmath34 can be expanded into terms involving @xmath35 and @xmath32 which obviously has a finite number of terms because of the grassmannian nature of @xmath35 and @xmath32 .
one then imposes the normalization condition @xmath36 and vanishing of lorentz nonscalar moments like @xmath37 and @xmath38 .
the stochasticity parameter @xmath1 is defined as @xmath39 . here
, @xmath1 is a complex parameter with mass dimension unity . as computed in ref.@xcite , and as worked out in this analysis explicitly in the appendix , the above leads to the following hermitian probability distribution . @xmath40 for the simple case of a wess - zumino type of scenario@xcite where the kinetic term is obtained from @xmath41 and the superpotential is given as @xmath42 , where @xmath43 is a chiral superfield , one finds that the effect of stochasticity as described above leads to the following susy breaking term : @xmath44 applying the stochasticity idea to the superpotential of mssm , along with considering the effect on the gauge kinetic energy function , the above formalism leads to the following tree level soft susy breaking parameters to be given at the high scale @xmath3 : + a. universal gaugino mass parameter @xmath45 , + b. universal trilinear soft parameter @xmath46 , + c. universal bilinear higgs soft parameter @xmath47 .
+ recall that there is no scalar mass soft susy breaking term in ssm .
+ for convenience we take @xmath1 to be a real positive number with an additional input sign@xmath48 . if we count the universal gaugino mass parameter @xmath12 as the independent parameter , we have , @xmath49 as has already been discussed before , we introduce a nonvanishing scalar mass parameter @xmath27 and fix @xmath3 at @xmath4.@xcite . ]
thus with the above extension , the input quantities for the stochastic susy model are : @xmath12 , @xmath27 , sign(@xmath50 ) and sign(@xmath1 ) .
we note that the model quite naturally is associated with nonvanishing trilinear soft breaking terms .
as we will see , this is quite interesting in view of the recent lhc announcement for the higgs mass range centering around 125 gev@xcite . in this analysis
, we will discuss only the case of @xmath51 because the other sign of @xmath1 does not produce a spectra compatible with the dark matter relic density constraint .
the requirement of the rewsb then results in the following relations at the electroweak scale : @xmath52 and , @xmath53 where @xmath54 denote the one - loop corrections@xcite .
here , @xmath55 refers to the value of bilinear higgs coupling at the electroweak scale which has to be consistent with its given value @xmath15 at @xmath4 .
@xmath15 is determined via @xmath12 apart from a sign of the stochasticity parameter as mentioned before .
consequently , @xmath22 is a derived quantity in the model .
@xmath15 at the scale @xmath4 and @xmath55 at the electroweak scale are connected via the following rge written here at the one - loop level : @xmath56 where @xmath57 with @xmath58 being the renormalization scale . @xmath59 for @xmath60 refer to scaled gauge coupling constants ( with @xmath61 ) and @xmath62 for @xmath60 are the running gaugino masses .
@xmath63 are the squared yukawa couplings , e.g , @xmath64 where @xmath65 is the top yukawa coupling . in this analysis ,
the value of @xmath22 is determined via eqs.[mueqn]-[beqn ] along with @xmath66 at the scale @xmath4 .
we use suspect@xcite for solving the rges and obtaining the spectra .
the code takes @xmath22 as an input quantity .
hence , we implement a self - consistent method of solution that starts from a guess value of @xmath22 resulting into a @xmath67 that in general would not agree with the input of @xmath15 .
use of a newton - raphson root finding scheme ensures a fast convergence toward the correct value of @xmath22 when @xmath67 matches with the input of @xmath15 .
here we stress that we do not encounter any parameter point with multiple values of @xmath22 in our analysis .
is given as an input . ]
the fact that the model has @xmath22 as a derived quantity necessitates studying the behavior of the evolution of the bilinear higgs parameter @xmath55 .
figure[rgeneg ] shows the evolution of a few relevant couplings for a specimen input of @xmath68 gev , @xmath69 tev and @xmath70 in mod - ssm . for @xmath51 , both @xmath15 as well as @xmath14
are negative , namely , @xmath71 and @xmath72 . for a valid parameter point within the model with @xmath70 , we require @xmath73 , a necessity in order to have a positive @xmath74 from eq.([maeqn ] ) . , which is the ratio of two vacuum expectation values , is positive .
hence , @xmath75 is also positive . ]
we note that the denominator in the right - hand side of eq.([maeqn ] ) is the square of pseudoscalar higgs mass which needs to be positive .
the fact that @xmath55 is originally negative at @xmath4 and has to change to a positive value at @xmath23 puts a strong constraint on the parameter space of the model .
numerically , this results in @xmath22 assuming large values . in regard to the evolution of @xmath76 parameters we defer our discussion until fig.[mhalf - m0 ] .
+ gev , @xmath69 tev and @xmath70 . with @xmath51 ,
one has @xmath71 and @xmath72 . for a valid parameter point within the model with @xmath70
we require @xmath77 , a necessity in order to have a positive @xmath74 from eq.([maeqn ] ) . ]
figure [ mhalf - tanbeta ] shows a scatter plot of parameter points in the @xmath78 plane that satisfy the rewsb constraints of eqs.([mueqn ] ) and ( [ maeqn ] ) .
as @xmath27 is varied up to 7 tev and @xmath12 up to 2 tev , @xmath22 is seen to have a range of 32 to 48 .
the spread of @xmath22 for a given @xmath12 arises from variation of @xmath27 . for smaller values of @xmath12
, there is a larger dependence of @xmath27 on @xmath22 .
hence , there is a larger spread of @xmath22 when @xmath27 is varied . for larger @xmath12 ,
the valid solution of @xmath22 has a lesser dependence on @xmath27 .
hence , for larger @xmath12 values the spread in values of @xmath22 decreases .
the white region embedded within the blue - green area corresponds to parameter points with no valid solution satisfying rewsb .
typically , the two rewsb conditions are nonlinear in nature with respect to @xmath12 and @xmath27 .
the code unsuccessfully tries with a large number of iterations to find consistent @xmath79 and @xmath80 solutions for parameter points within the white region ( for all @xmath81 .
thus a lack of a valid @xmath22 for any value of @xmath27 results in the above white region .
it is worth mentioning that the range of valid @xmath22 is much larger for the case of @xmath82 where @xmath55 stays positive throughout the range from @xmath4 to @xmath23 .
this is unlike the case of @xmath51 under discussion , where @xmath55 is negative at @xmath4 and necessarily has to become positive at the electroweak scale , thus adding stringency to @xmath22 in its range .
however , as already mentioned , we will not discuss the case of @xmath82 further because of the resulting overabundance of dark matter for the entire parameter space for this sign of @xmath1 .
+ plane when @xmath27 and @xmath12 are scanned up to 7 tev and 2 tev , respectively .
here we only consider the validity of rewsb constraints of eqs.([mueqn ] ) and ( [ maeqn ] ) .
the values of @xmath22 that satisfy the rewsb constraint vary from 32 to 48 .
the white region inside the shaded ( blue - green ) region corresponds to invalid parameter points where no consistent solution satisfying rewsb could be found even after trying with a large number of iterations .
the spread of @xmath22 for a given @xmath12 arises from variation of @xmath27 .
this spread decreases as @xmath12 becomes larger because of decreased sensitivity on @xmath27 while satisfying rewsb . ]
+ we will study now the effect of low energy constraints particularly in the context of the recent discovery of the higgs - like boson@xcite .
figure .[mhalf - m0 ] shows the result in the @xmath83 plane for @xmath51 .
a sufficiently nonvanishing @xmath84 ( see , for example , figure [ rgeneg ] ) helps in producing a large loop correction to the lighter cp - even higgs boson @xmath85 .
this can be understood by looking at the expression for the dominant part of loop correction to the higgs boson mass coming from the top - stop sector@xcite @xmath86 . \label{higgscor}\ ] ] here , @xmath87 , @xmath88 , @xmath89 gev and @xmath90 is the running top - quark mass that also takes into account qcd and electroweak corrections . the loop correction is maximized if @xmath91 .
clearly , a nonvanishing @xmath14 can be useful to increase @xmath92 so that @xmath2 reaches the lhc specified zone without a need to push up the average sparticle mass scale by a large amount . as particularly mentioned in ref@xcite ,
the ssm is additionally attractive in this context since it naturally possesses a nonvanishing and large @xmath93 .
we further note that in this model , the lighter higgs boson @xmath85 has couplings similar to those in the standard model because the cp - odd higgs boson mass ( @xmath94 ) is in the decoupling zone@xcite .
the atlas and cms results for the possible higgs boson masses are @xmath95 gev and @xmath96 gev , respectively@xcite . in regard to mssm light higgs boson mass
, we note that there is about a 3 gev uncertainty arising out of uncertainties in the top - quark mass , renormalization scheme , as well as scale dependence and uncertainties in higher order loop corrections up to three loop@xcite .
hence , in this analysis we consider the following limits for @xmath2 : @xmath97 plane for @xmath70 and @xmath51 for mod - ssm . @xmath14 and
@xmath15 in the model satisfy @xmath98 and @xmath99 .
@xmath22 becomes a derived quantity that varies between 32 and 48 .
the higgs boson limits are shown as two solid blue lines .
@xmath100 limit is shown as a maroon dot - dashed line .
the lower part corresponds to discarded region via eq.([bsgammalimits ] ) where the branching ratio goes below the lower limit of the constraint .
@xmath25 limit is shown as a brown solid line of which the lower region exceeds the upper limit of eq.([bsmumulimit ] ) .
the top green region ( i ) corresponds to discarded zone via rewsb .
the bottom blue - green region ( iii ) refers to the zone where stau becomes lsp or tachyonic .
the gray region ( ii ) has discontinuous patches of valid parameter zones , the details of which are mentioned in the text .
red points / areas falling in region - ii satisfy the wmap-7 data only for the upper limit of eq.[wmap7data ] .
typically , the red points bordering region - iii and some part of the extreme left red points ( for very small @xmath12 ) satisfy both the upper and the lower limits of eq.([wmap7data ] ) . ]
we will outline the other relevant limits used in this analysis .
a susy model parameter space finds a strong constraint from @xmath100 .
in sm , the principal contribution that almost saturates the experimental value comes from the loop comprising of top - quark and w - boson@xcite . in mssm ,
principal contributions arise from loops containing top quark and charged higgs bosons , and the same containing top squarks and charginos@xcite .
the chargino loop contributions are proportional to @xmath101 and this may cause cancellations or enhancements between the principal terms of the mssm contribution depending on the sign of @xmath101 .
similar to msugra , both the ssm and mod - ssm with @xmath51 also typically have @xmath102 . with @xmath70 ,
this primarily means cancellation between the chargino and the charged higgs contributions .
this leads to a valid region for @xmath100 for a larger sparticle mass scale compared to the case of @xmath103 .
@xmath100 constraint thus favors the positive sign of @xmath50 by allowing larger areas of parameter space .
we consider the experimental value @xmath104@xcite .
this results in the following @xmath105 level zone as used in this analysis .
@xmath106 the above constraint is displayed as a maroon dot - dashed line .
the left region of this line would be a discarded zone .
next , the fact that the stochastic model with @xmath51 selects appreciably large values for @xmath22 necessitates checking the @xmath107 limit .
this is required because @xmath107 increases with @xmath22 as @xmath108 and decreases with increase in @xmath109 , the mass of pseudoscalar higgs boson as @xmath110@xcite .
we use the recent experimental limit from lhcb@xcite : @xmath111 .
this is contrasted with the sm evaluation @xmath112 @xcite . as in ref.@xcite , combining the errors of the lhcb data along with that of the sm result one finds the following : @xmath113 the upper limit of @xmath25 is shown as a brown solid line going across fig.[mhalf - m0 ] .
parameter values in the region below this curve lead to values of @xmath114 higher than the above limit .
we also compute @xmath115 in this analysis . the susy contribution to @xmath115
is typically effective for large @xmath22 and small charged higgs boson mass scenarios@xcite . the experimental data from
_ babar_@xcite reads @xmath116 . the recent result from belle@xcite for @xmath117 that used the hadronic tagging method is given by @xmath118 .
the same branching ratio from belle extracted by using a semileptonic tagging method is @xmath119@xcite .
we use ref.@xcite for the result of averaging of all the recent belle and _ babar _ data which is @xmath120 .
the sm result strongly depends on the ckm element @xmath121 and the b - meson decay constant .
we use @xmath122@xcite . using the above theoretical and experimental errors appropriately , we obtain the following : @xmath123 this translates into @xmath124 at @xmath125 . here
@xmath126 denotes the branching ratio in a susy framework , of course including the sm contribution .
in general , we find that the model parameter space of mod - ssm is not constrained by @xmath127 since charged higgs bosons are sufficiently heavy .
we have not , however , included the constraint from muon @xmath128 in this analysis considering the tension arising out of large deviation from the sm value , uncertainty in hadronic contribution evaluations and accommodating susy models in view of the lhc sparticle mass lower limits@xcite .
we now explore the cosmological constraint for neutralino dark matter relic density@xcite . at 3@xmath129 ,
the wmap-7 data@xcite are considered as shown below : @xmath130 the conclusions in regard to the relic density constraint is additionally found to be sensitive on the top - quark mass in this model .
we divide the dark matter analysis into two parts depending on ( a ) the top - quark pole mass set at 173.3 gev and ( b ) using a spread of top - quark pole mass within its range @xmath131 gev following the result of the recent analysis performed in ref.@xcite . in this context
, we note that the experimental value as measured by the cdf and d0 collaborations of tevatron is : @xmath132 gev@xcite .
gev . ] the lightest neutralino , the lsp of the model is typically highly bino dominated except in a few regions where the higgsino mixing parameter @xmath50 turns out to be small .
the parameter points within all of the white region in figure [ mhalf - m0 ] have bino - dominated lsp .
at this point we note that the implementation of the rewsb conditions as manifest in eqs.([mueqn ] ) and ( [ maeqn ] ) has to be done by keeping in mind ( i ) positivity of @xmath74 and ( ii ) positivity of @xmath79 and and ( iii ) the requirement of @xmath134 being related to @xmath12 as given by eq .
( [ neweqn ] ) , as well as requiring that the lighter chargino mass lower limit is respected .
all these requirements lead the top green shaded region ( labeled as i ) to be a discarded zone .
on the other hand , the gray shaded region ( ii ) has discontinuous zones of valid parameter points , shown in red .
the red points have considerably small values of @xmath50 , thus giving the lsp a large degree of higgsino mixing .
apart from the red points , there are no solutions in the gray areas of region ii .
the need to satisfy the point ( i ) as explained above along with the requirement to satisfy the condition ( iii ) ( implemented via a newton - raphson method of finding the correct @xmath22 ) stringently negates the existence of solution zones within region ii . as a result , either there are solutions with appreciably small @xmath50 or no solution at all within this region .
this on the other hand leads to a large amount of @xmath135 coannihilation .
such degrees of coannihilations indeed cause the lsp to be only a subdominant component of dm .
thus , in this part of the analysis we consider the possibility of an underabundant dark matter candidate and ignore the lower limit of the wmap-7 data . typically we see that the relic density falls below the lower limit of eq.([wmap7data ] ) in region ii by an order of magnitude .
such underabundant lsp scenarios have been discussed in several works@xcite .
the lower shaded region iii is disallowed as the mass of the stau ( @xmath136 ) turns negative or it is the lsp .
typically the red strip near region iii refers to the lsp - stau coannihilation zone where the relic density can be consistent with both the upper and lower limits of eq.([wmap7data ] ) . quite naturally
, the coannihilation may be stronger and this would additionally produce some underabundant dm points for this zone .
finally , the leftmost red region ( with very small @xmath12 ) satisfying wmap-7 data is discarded by all other constraints .
we note that only a small region satisfying the higgs mass bound is discarded via @xmath137 . on the other hand
, the higgs mass bound line of 122 gev supersedes the constraint imposed by the recent data on inclusive search for susy by the atlas experiment@xcite . the most potent constraint to eliminate a large region of parameter space with @xmath27 up to 1.5 tev or so is due to @xmath25 data ( eq.([bsmumulimit ] ) ) .
the constraint is effective simply because of the large values of @xmath22 involved in the model .
the discarded part of parameter space via the above constraint includes a large zone that satisfies the dark matter limit via lsp - stau coannihilation .
we will now describe the spin - independent direct detection scattering cross section for scattering of the lsp with proton .
the scalar cross section depends on t - channel higgs exchange diagrams and s - channel squark diagrams .
unless the squark masses are close to that of the lsp , the higgs exchange diagrams dominate@xcite .
we note that for the cases of parameter points with @xmath138 , where @xmath139 refers to the lower limit of eq.([wmap7data ] ) , one must appropriately include the fraction of local dm density contributed by the specific candidate of dm under discussion while evaluating the event rate .
+ scattering cross section vs lsp mass .
the scaling factor is given as @xmath140 , where @xmath139 refers to the lower limit of eq.([wmap7data ] ) . ]
this translates into multiplying @xmath141 for such underabundant scenarios by @xmath142 . here
, @xmath143 is the actual dm density contributed by the specific dm candidate contributing to @xmath144 where the latter is the local dark matter density .
we thus use @xmath145 where @xmath140 . on the other hand , @xmath146 is simply @xmath147 for abundant or overabundant dark matter cases .
thus , we conveniently define @xmath148@xcite .
figure[sigmasifigscaled ] shows the rescaled cross section as computed via micromegas version 2.4@xcite .
we wish to emphasize that while some region of parameter space where the lsp typically has a large degree of higgsino component is eliminated via xenon100 data as announced in the summer of 2012@xcite , a large section of parameter space remains to be explored via future direct detection of dm experiments .
this consists of both types of coannihilation zones , namely , the chargino as well as the stau coannihilation zones .
we must also keep in mind the issue of theoretical uncertainties , particularly the hadronic uncertainties in evaluating @xmath141 .
the strangeness content of nucleon finds a large reduction in the evaluation of relevant couplings via lattice calculations@xcite .
this is not incorporated in our computation while using micromegas to calculate the cross section .
thus , the above itself will cause a reduction of @xmath141 by almost an order of magnitude .
there is also an appreciable amount of uncertainty of the local dark matter density@xcite .
all these points need to be kept in mind while evaluating the implications of fig . 4 for susy models .
+ it is to be noted that in the gray area ( shown as region ii ) of fig.[mhalf - m0 ] the existence of a valid solution depends very critically on the parameters of the model .
it was found that either ( i ) we obtain a very small @xmath50 ( barely satisfying the lighter chargino mass lower limit ) in the gray area that would only provide us with extreme coannihilation between lighter chargino and lsp leading to underabundance of dm , or ( ii ) we find no valid solution at all .
the sensitivity arises from the stringency of satisfying rewsb on the parameter space for this region .
in other words , for a given @xmath12 , a small change in assumed @xmath27 for a valid parameter point would produce a small change in @xmath22 in the _ white _ region producing most probably another valid parameter point . however , the change may not be allowed via rewsb , particularly eq.([maeqn ] ) if the parameter point is considered in the _ gray _ region .
eq.([maeqn ] ) means @xmath74 needs to be a positive quantity less than unity .
this typically becomes a severe constraint even if condition of satisfying the lighter chargino mass lower limit is met . with the top - quark mass having a strong influence on rewsb
, it may be useful to investigate whether varying @xmath149 may extract newer valid points within the gray region that would have a suitable @xmath50 so as to satisfy a well - tempered@xcite lsp situation .
this will then have the right abundance of dark matter satisfying both the lower and the upper limit of dm of eq.([wmap7data ] ) .
the present experimental data on top - quark mass read : @xmath132 gev@xcite .
recently ref.@xcite predicted the pole mass of top - quark to be @xmath150 gev .
the analysis used the next - to - next - to - leading order ( nnlo ) in the qcd prediction of the inclusive @xmath151 cross section and the tevatron and lhc data of the the same cross section .
the comparison between the experimental and theoretical results helped extracting the top - quark mass in the modified minimal subtraction ( @xmath152 ) scheme .
this was then used to compute the pole mass @xmath153 .
we now extend our analysis by investigating the effect of varying top - quark pole mass within the above range ( @xmath150 gev ) on the solution space of the model .
indeed , we will see that even with a variation of @xmath154 gev , the range of experimental error would be enough to have a substantial effect on the conclusions .
+ is varied by 2.8 gev on either side of 173.3 gev . here
@xmath12 and @xmath27 are scanned up to 2 tev and 7 tev , respectively , for @xmath70 and @xmath51 .
the two blue lines are the wmap-7 limits of eq.([wmap7data ] ) .
the value of @xmath133 gev as used in figure [ mhalf - m0 ] visibly falls in a disfavored zone in the context of obtaining the correct relic density , particularly toward the lower limit of of eq.([wmap7data ] ) . ]
figure [ top - dependence ] shows the scattered points that satisfy the wmap-7 given range of relic density for the aforesaid variation of @xmath155 . here ,
@xmath12 and @xmath27 are varied up to 2 tev and 7 tev , respectively .
we see from this figure that the central value of @xmath149 ( = 173.3 ) gev as considered in figure [ mhalf - m0 ] is indeed the one which has the _ least amount of possibility to satisfy the wmap-7 data .
the occurrence of points which satisfies wmap-7 data is particularly rare around this value of @xmath149 , near the lower part of the limit of relic density .
_ we already found that the gray region ( region ii ) of figure [ mhalf - m0 ] is a sensitive zone because of rewsb , where @xmath50 can be quite small . for such small values of @xmath50 , one can only expect a large degree of @xmath135 coannihilation which results into very small relic density .
the latter goes below the lower limit of eq.([wmap7data ] ) .
hence no red point exists near the bottom blue line of figure [ top - dependence ] for this value of @xmath149 . on the contrary ,
a value of @xmath149 less than 1 gev from the central value , which is only within the experimental error of @xmath156 , would cause to have an lsp with correct abundance for dm .
the most favored zone for @xmath149 , however , would be from 171 to 172 gev for satisfying the relic density limits .
in fact , a reduced degree of sensitivity for satisfying rewsb is the reason for obtaining a well - tempered lsp while considering a top - quark mass little away from the central value .
+ plane for @xmath70 and @xmath51 when top - quark mass is varied by 2.8 gev on either side of 173.3 gev .
region i is the same as that in fig.[mhalf - m0 ] except that it is now a discarded zone for all values of @xmath149 within its limit .
the region ii is discarded via stau turning the lsp or turning itself tachyonic for all @xmath149 .
the constraints from @xmath100 and @xmath25 are as shown .
the lines show the boundary of purely discarded zones irrespective of variation of @xmath149 within its range .
there is no region toward the left of the blue line labeled as @xmath157 gev for which @xmath2 may become larger than 124 gev irrespective of values of @xmath149 .
three benchmark points a , b and c are shown corresponding to table [ spectratable ] . ]
figure [ m0mhalfwithtopscanned ] shows the effect of scanning the top - quark mass on the @xmath83 plane .
region i shown in green is a disallowed area of parameter space via rewsb similar to fig.[mhalf - m0 ] , except that it is now an invalid region for all values of @xmath149 within its limit .
thus , this region is smaller in extension than the corresponding region of figure [ mhalf - m0 ] .
the region ii is disallowed because of stau turning tachyonic or the least massive .
the constraints from @xmath100 and @xmath25 are as shown .
the lines denote the boundary of purely discarded zones irrespective of variation of @xmath149 within its range . the blue line for @xmath157
gev means that @xmath158 gev for all the region left of the line irrespective of the value of @xmath149 within the range .
the atlas specified limit@xcite for squarks also falls well within this left zone .
the neutralino relic density satisfied areas are below region i and above region ii .
the _ red _ points satisfy both the limits of the wmap-7 data , and thus correspond to having the right degree of abundance of dm . on the other hand ,
we have also shown _ blue - green _ points that only satisfy the upper limit of the wmap-7 data . in this part of the analysis
, we consider the lsp to have the correct abundance so as to be a unique candidate for dm .
finally we show the effect of varying @xmath149 on the spin - independent lsp - proton scattering cross section in figure [ sigmasifigtopscanned ] .
only the wmap-7 satisfied points are shown ( in red ) . considering an order of magnitude of uncertainty ( reduction ) , primarily because of the issue of strangeness content of nucleon as well as astrophysical uncertainties as mentioned before , we believe that the recent xenon100 data still can accommodate mod - ssm even while considering the lsp as a unique candidate of dm
. + is scanned along with @xmath12 and @xmath27 for @xmath70 and @xmath51 . only wmap-7 satisfied points ( for both the lower and the upper limit ) are shown along with the xenon100 exclusion limit . ]
table [ spectratable ] shows three benchmark points of the model .
the top - quark mass @xmath149 is as shown for each of the cases .
points a and b correspond to values of @xmath149 which are entirely within the experimental error .
the points a and c correspond to the upper region of figure [ m0mhalfwithtopscanned ] , and these two points correspond to the hyperbolic branch ( hb)/focus point ( fp ) zone@xcite .
these points are associated with a large degree of @xmath135 coannihilation .
the degree of agreement between the desired value of @xmath15 and the one obtained via the newton - raphson iteration may be seen from the fifth and the sixth rows .
we allowed a maximum deviation of 1@xmath159 in the iterative procedure while scanning the parameter space .
a majority of the points are consistent within 0.1@xmath159 in this regard .
the charginos and the neutralinos are relatively lighter for points a and c in comparison to those in point b. the scalar particles , on the other hand , are relatively lighter for point - b . undoubtedly , like many susy models analyzed or reanalyzed after the higgs boson discovery , the spectra is on the heavier side .
the fact that @xmath14 is nonvanishing and adequately large helps in reducing the average sparticle mass to a great extent compared to vanishing @xmath14 scenarios satisfying the current lighter higgs boson limit .
the @xmath2 for point - c , however , goes below the assumed limit of eq.([higgslimits ] ) .
however , we , still believe that it is within an acceptable zone considering the various uncertainties to compute @xmath2 as mentioned before .
points a and c have larger spin - independent scattering cross section @xmath160 than the xenon100 limit .
but , we believe this is within acceptable limit considering the existing uncertainties arising out of strangeness content of nucleon as well as those from astrophysical origins , particularly from local dm density . finally , it is also possible to satisfy all the limits in full subject to a @xmath161 heavier spectra and/or considering a multicomponent dm scenario .
+ .spectra of three specimen parameter points a , b and c as shown in fig.[m0mhalfwithtopscanned ] .
results with marginal deviation from the assumed limits are shown in italics ( red ) .
muon @xmath128 is not imposed as a constraint in this analysis .
@xmath162 has two entries .
the first one is the desired value while the second one is the value obtained with a suitable @xmath22 as found from the newton - raphson root finding scheme .
see text for further details . [
cols="<,^,^,^",options="header " , ]
in the ssm@xcite , one assumes that a manifestation of an unknown but a fundamental mechanism of susy breaking may effectively lead to stochasticity in the grassmannian parameters of the superspace . with a suitable probability distribution decided out of physical requirements , stochasticity in grassmannian coordinates for a given khler potential and a superpotential may lead to well - known soft breaking terms .
when applied to the superpotential of the mssm , the model leads to soft breaking terms like the bilinear higgs coupling term , a trilinear soft term as well as a gaugino mass term , all related to a parameter @xmath1 , the scale of susy breaking .
the other scale considered in the model of ref.@xcite at which the input quantities are given is @xmath3 , where the latter can assume a value between the gauge coupling unification scale @xmath4 and the planck mass scale @xmath21 .
there is an absence of a scalar mass soft term at @xmath3 in the original model that leads to stau turning to be the lsp or even turning itself tachyonic if @xmath3 is chosen as @xmath4 .
this is only partially ameliorated when @xmath3 is above @xmath4 . however , the model because of its nonvanishing trilinear parameter is potentially accommodative to have a larger lighter higgs boson mass via large stop scalar mixing . as recently shown in ref.@xcite the model in spite of its nice feature of bringing out the desired soft breaking terms of mssm
is not able to produce @xmath2 above 116 gev , and it has an lsp which is only a subdominant component of dm . in this work , a minimal modification ( referred as mod - ssm ) is made by allowing a nonvanishing single scalar mass parameter @xmath27 as an explicit soft breaking term .
phenomenologically the above addition is similar to what was considered in minimal amsb while confronting the issue of sleptons turning tachyonic in a pure amsb framework .
the modified model successfully accommodates the lighter higgs boson mass near 125 gev .
additionally , it can accommodate the stringent constraints from the dark matter relic density , @xmath25 , @xmath163 and xenon100 data on direct detection of dark matter .
a variation of top - quark mass within its allowed range is included in the analysis and this shows the lsp to be a suitable candidate for dark matter satisfying both the limits of wmap-7 data .
finally , we remind that the idea of stochastic superspace can easily be generalized to various scenarios beyond the mssm .
* acknowledgments * + m.c . would like to thank the council of scientific and industrial research , government of india for support .
u.c . and r.m.g are thankful to the cern thph division ( where the work was initiated ) for its hospitality .
r.m.g . wishes to acknowledge the department of science and technology of india , for financial support under grant no .
sr / s2/jcb-64/2007 .
we would like to thank b. mukhopadhyaya , s. kraml , p. majumdar , k. ray , s. roy and s. sengupta for valuable discussions .
using ref.@xcite , we consider the following hermitian probability distribution : @xmath164 here , a , b , c , d and @xmath165 are complex numbers .
@xmath166 , @xmath167 , @xmath168 , and @xmath169 are grassmann numbers . * vanishing fermionic moments : * + next , we require vanishing of moments of fermionic type because of the requirement of lorentz invariance . the fact that @xmath181 means @xmath182 .
similarly , @xmath183 means @xmath184 , and @xmath185 leads to @xmath186 .
finally , @xmath187 gives @xmath188 and @xmath189 gives @xmath190 .
we consider the wess - zumino model with a single chiral superfield @xmath43 .
@xmath43 has the following expansion@xcite : @xmath199 correspondingly , for @xmath200 we have : @xmath201 the kinetic term of the lagrangian @xmath202 is @xmath203}_d$ ] .
one finds , @xmath204}\phi^ * + \sqrt 2 \theta^2 \bar \theta \bar \psi
f - 2i\theta \sigma^\mu \bar \theta \phi^*{[\partial_\mu]}\phi \nonumber \\ & & + i\sqrt 2 { \bar \theta}^2 \theta \sigma^\mu\bar \psi { [ \partial_\mu]}\phi + \sqrt 2 { \bar \theta}^2 \theta \psi f^ * \nonumber \\ & & + \theta^2 { \bar \theta}^2 \left(f^ * f + \frac12 \partial_\mu \phi^ * { [ \partial^\mu]}\phi -\frac12 \phi^ * { [ \partial_\mu ] } \partial^\mu \phi + i\psi\sigma^\mu{[\partial_\mu]}\bar \psi\right ) .
\label{kineticinfull}\end{aligned}\ ] ] here , @xmath205}y=\frac12\left(x\partial_\mu y -y \partial_\mu x \right)$ ] . upon vanishing the appropriate surface terms ,
the d term in particular reads : @xmath206}_d = f^*f + \partial_\mu \phi^ * \partial^\mu \phi+\frac{i}{2}\left(\psi \sigma^\mu \partial_\mu \bar \psi- \partial_\mu \psi \sigma^\mu \bar \psi\right ) . \label{actualkinetic}\ ] ] next , we consider the superpotential @xmath207 given by @xmath208 .
one has : @xmath209 thus , @xmath210 the potential energy density term will be as follows : @xmath211}_f= \left(m \phi f -\frac12 m \psi \psi + h f \phi^2 -h \psi \psi \phi\right)+h.c .
\label{potentialdensity}\ ] ] * kinetic and potential terms averaged over @xmath35 and @xmath212 and emergence of soft susy breaking terms : * + averaging over the grassmannian coordinates , we compute @xmath213 . here
@xmath214 is the usual super - lagrangian density : @xmath215 .
then , using eq.([kineticinfull ] ) and eq.([actualprobability ] ) @xmath216 , namely , the kintetic part of @xmath217 is found as , @xmath218}_d + \overbrace{{|\xi|}^2{|\phi|}^2 + \xi^ * \phi f^ * + \xi \phi^ * f}^{\cancel{susy}}. \label{derivedke}\ ] ] similarly , the potential energy density averaged over @xmath35 and @xmath212 is given by , @xmath219 using eq.([superpotentialexpr ] ) and eq.([actualprobability ] ) we find , @xmath220+h.c .
\label{derivedpe}\ ] ] total lagrangian is then : @xmath221 using eqs.([potentialdensity ] ) , ( [ derivedke ] ) and ( [ derivedpe ] ) we break @xmath217 into susy invariant and susy breaking parts as follows : @xmath222 where @xmath223}_d + { [ w+ h.c.]}_f,\ ] ] and @xmath224.\ ] ] the equations of motion of auxiliary fields are then @xmath225 substituting @xmath226 and @xmath227 in @xmath217 , one finds , @xmath228 where @xmath229 is the usual on shell susy invariant lagrangian for interacting wess - zumino model and is given by , @xmath230,\end{aligned}\ ] ] and @xmath231 is given by , @xmath232 .
\label{lsofteqn}\ ] ] we remind that a negative sign in the left - hand side of eq.[lsofteqn ] appears simply because of considering a positive sign before @xmath233 in eq.([totlagrangiansum ] ) while writing the total lagrangian .
we note that it is only the superpotential term in the original theory that leads to soft breaking terms in the resulting lagrangian ( @xmath234 and @xmath235 going from @xmath207 to @xmath231 ) .
thep presence of a vector field will not lead to any soft susy breaking term .
this can easily be seen by considering a vector superfield in wess - zumino gauge . +
* mssm : * + in the mssm , as mentioned in ref.@xcite , the superpotential term @xmath236 will lead to @xmath237 and terms like @xmath238 will lead to @xmath239 as soft susy breaking terms . here
, fields with tildes denote the scalar component of the corresponding chiral superfields .
one further obtains a gaugino mass term @xmath240 .
thus , one finds a universal gaugino mass @xmath241 , a bilinear higgs soft parameter @xmath242 and a universal trilinear soft parameter @xmath46 . with no resulting scalar mass term
, one has the universal scalar mass parameter @xmath243 .
these are the input quantities to be given at a scale @xmath3 .
low energy spectra are then found via rg evolutions .
reference @xcite considered @xmath1 and @xmath3 as the input quantities and considered @xmath244 .
m. drees , p. roy and r. m. godbole , _ theory and phenomenology of sparticles _ , ( world scientific , singapore , 2005 ) . h. baer and x. tata , _ weak scale supersymmetry : from superfields to scattering events _ , cambridge , uk : univ .
( 2006 ) 537 p. g. aad _ et al . _ [ atlas collaboration ] , phys .
b * 716 * , 1 ( 2012 ) [ arxiv:1207.7214 [ hep - ex ] ] ; + s. chatrchyan _ et al . _
[ cms collaboration ] , phys .
b * 716 * , 30 ( 2012 ) [ arxiv:1207.7235 [ hep - ex ] ] .
a. h. chamseddine , r. arnowitt and p. nath , phys .
* 49 * , 970 ( 1982 ) ; r. barbieri , s. ferrara and c. a. savoy , phys .
b * 119 * , 343 ( 1982 ) ; l. j. hall , j. lykken and s. weinberg , phys .
d * 27 * , 2359 ( 1983 ) ; p.
nath , r. arnowitt and a. h. chamseddine , nucl .
b * 227 * , 121 ( 1983 ) ; n. ohta , prog .
. phys . * 70 * , 542 ( 1983 ) ; p. nath , r. arnowitt and a.h .
chamseddine , _ applied n = 1 supergravity _
( world scientific , singapore , 1984 ) .
a. kobakhidze , n. pesor and r. r. volkas , phys .
d * 79 * , 075022 ( 2009 ) [ arxiv:0809.2426 [ hep - ph ] ] .
a. kobakhidze , n. pesor and r. r. volkas , phys .
d * 81 * , 095019 ( 2010 ) [ arxiv:1003.4782 [ hep - ph ] ] .
a. kobakhidze , n. pesor , r. r. volkas and m. j. white , phys .
d * 85 * , 075023 ( 2012 ) [ arxiv:1201.1624 [ hep - ph ] ] .
r. aaij _ et al . _
[ lhcb collaboration ] , phys .
lett . * 110 * , 021801 ( 2013 ) arxiv:1211.2674 .
e. aprile _ et al .
_ [ xenon100 collaboration ] , arxiv:1207.5988 [ astro-ph.co ] .
r. l. arnowitt and p. nath , phys .
d * 56 * , 2833 ( 1997 ) [ hep - ph/9701325 ]
. g. f. giudice , m. a. luty , h. murayama , and r. rattazzi , j. high energy phys . *
12 * , 027 ( 1998 ) ; l. randall and r. sundrum , nucl . phys .
* b557 * , 79 ( 1999 ) ; j. bagger , t. moroi , and e. poppitz , j. high energy phys . * 04 * , 009 ( 2000 ) .
t. gherghetta , g. f. giudice , and j. d. wells , nucl .
b559 * , 27 ( 1999 ) ; j. l. feng , t. moroi , l. randall , m. strassler , and s. su , phys .
lett . * 83 * , 1731 ( 1999 ) ; j. l. feng and t. moroi , phys . rev .
* d61 * , 095004 ( 2000 ) ; u. chattopadhyay , d. k. ghosh and s. roy , phys .
d * 62 * , 115001 ( 2000 ) .
a. djouadi , j. -l .
kneur and g. moultaka , comput .
commun . * 176 * , 426 ( 2007 ) [ hep - ph/0211331 ] .
m. drees and m. m. nojiri , nucl .
b * 369 * , 54 ( 1992 ) .
a. djouadi , phys .
rept . * 459 * , 1 ( 2008 ) [ hep - ph/0503173 ] .
haber and r. hempfling , phys .
lett . * 66 * , 1815 ( 1991 ) ; y. okada , m. yamaguchi and t. yanagida , prog .
phys . * 85 * , 1 ( 1991 ) , phys .
b * 262 * , 54 ( 1991 ) ; j. ellis , g. ridolfi and f. zwirner , phys .
b * 257 * , 83 ( 1991 ) , phys .
b * 262 * , 477 ( 1991 ) .
haber and y. nir , phys .
b306 * 327 ( 1993 ) ; h.e .
haber , hep - ph/9505240 ; a. dobado , m.j .
herrero and s. penaranda , eur .
j. * c17 * 487 ( 2000 ) ; j.f .
gunion and h.e .
haber , phys .
* d67 * 075019 ( 2003 ) ; a. djouadi and r. m. godbole , arxiv:0901.2030 [ hep - ph ] .
a. arbey , m. battaglia , a. djouadi and f. mahmoudi , jhep * 1209 * , 107 ( 2012 ) [ arxiv:1207.1348 [ hep - ph ] ] .
s. heinemeyer , o. stal and g. weiglein , phys .
b * 710 * , 201 ( 2012 ) [ arxiv:1112.3026 [ hep - ph ] ] .
b. c. allanach , a. djouadi , j. l. kneur , w. porod and p. slavich , jhep * 0409 * , 044 ( 2004 ) [ hep - ph/0406166 ] .
g. degrassi , s. heinemeyer , w. hollik , p. slavich and g. weiglein , eur .
j. c * 28 * , 133 ( 2003 ) [ hep - ph/0212020 ] .
r. v. harlander , p. kant , l. mihaila and m. steinhauser , phys .
* 100 * , 191602 ( 2008 ) [ phys .
rev . lett .
* 101 * , 039901 ( 2008 ) ] [ arxiv:0803.0672 [ hep - ph ] ] .
s. bertolini , f. borzumati and a. masiero , phys .
lett . *
59 * , 180 ( 1987 ) ; n. g. deshpande , p. lo , j. trampetic , g. eilam and p. singer , phys .
lett . * 59 * , 183 ( 1987 ) ; b. grinstein and m. b. wise , phys .
b * 201 * , 274 ( 1988 ) ; b. grinstein , r. p. springer and m. b. wise , phys .
b * 202 * , 138 ( 1988 ) ; w. -s .
hou and r. s. willey , phys .
b * 202 * , 591 ( 1988 ) ; b. grinstein , r. p. springer and m. b. wise , nucl .
b * 339 * , 269 ( 1990 ) .
s. bertolini , f. borzumati , a. masiero and g. ridolfi , nucl .
b * 353 * , 591 ( 1991 ) ; r. barbieri and g. f. giudice , phys .
b * 309 * , 86 ( 1993 ) [ hep - ph/9303270 ] ; r. garisto and j. n. ng , phys . lett .
b * 315 * , 372 ( 1993 ) [ hep - ph/9307301 ] ; p.
nath and r. l. arnowitt , phys .
b * 336 * , 395 ( 1994 ) [ hep - ph/9406389 ] ; m. ciuchini , g. degrassi , p. gambino and g. f. giudice , nucl .
b * 534 * , 3 ( 1998 ) [ hep - ph/9806308 ] .
d. asner _
[ heavy flavor averaging group collaboration ] , arxiv:1010.1589 [ hep - ex ] .
s. r. choudhury and n. gaur , phys .
b * 451 * , 86 ( 1999 ) ; k. s. babu and c. kolda , phys .
lett . * 84 * , 228 ( 2000 ) ; a. dedes , h. k. dreiner , and u. nierste phys .
lett . * 87 * , 251804 ( 2001 ) ; p. h. chankowski and l. slawianowska , phys .
d * 63 * , 054012 ( 2001 ) [ hep - ph/0008046 ] ; r. arnowitt , b. dutta , t. kamon and m. tanaka , phys .
b * 538 * ( 2002 ) 121 ; j. k. mizukoshi , x. tata and y. wang , phys .
d * 66 * , 115003 ( 2002 ) ; s. baek , p. ko , and w. y. song , jhep * 0303 * , 054 ( 2003 ) ; g. l. kane , c. kolda and j. e. lennon , hep - ph/0310042 ; t. ibrahim and p. nath , phys .
d * 67 * , 016005 ( 2003 ) ; j.r .
ellis , k.a .
olive and v.c .
spanos , phys .
b * 624 * , 47 ( 2005 ) ; s. akula , d. feldman , p. nath and g. peim , phys .
d * 84 * , 115011 ( 2011 ) [ arxiv:1107.3535 [ hep - ph ] ] ; a. j. buras , j. girrbach , d. guadagnoli and g. isidori , arxiv:1208.0934 [ hep - ph ] .
a. j. buras , j. girrbach , d. guadagnoli and g. isidori , eur .
j. c * 72 * , 2172 ( 2012 ) [ arxiv:1208.0934 [ hep - ph ] ] .
l. roszkowski , s. trojanowski , k. turzynski and k. jedamzik , arxiv:1212.5587 [ hep - ph ] . g. isidori and p. paradisi , phys .
b * 639 * , 499 ( 2006 ) [ hep - ph/0605012 ] ; b. bhattacherjee , a. dighe , d. ghosh and s. raychaudhuri , phys .
d * 83 * , 094026 ( 2011 ) [ arxiv:1012.1052 [ hep - ph ] ] .
i. adachi _ et al . _
[ belle collaboration ] , arxiv:1208.4678 [ hep - ex ] .
k. hara _ et al . _ [ belle collaboration ] , phys .
d * 82 * , 071101 ( 2010 ) [ arxiv:1006.4201 [ hep - ex ] ] .
w. altmannshofer , m. carena , n. shah and f. yu , arxiv:1211.1976 [ hep - ph ] . s. bodenstein , c. a. dominguez and k. schilcher , phys .
d * 85 * , 014029 ( 2012 ) [ arxiv:1106.0427 [ hep - ph ] ] ; g. -c .
cho , k. hagiwara , y. matsumoto and d. nomura , jhep * 1111 * , 068 ( 2011 ) [ arxiv:1104.1769 [ hep - ph ] ] ; d. ghosh , m. guchait , s. raychaudhuri and d. sengupta , phys .
d * 86 * , 055007 ( 2012 ) [ arxiv:1205.2283 [ hep - ph ] ] ; s. akula , p. nath and g. peim , phys .
b * 717 * , 188 ( 2012 ) arxiv:1207.1839 [ hep - ph ] .
g. bertone , d. hooper and j. silk , phys .
rept . * 405 * , 279 ( 2005 ) [ hep - ph/0404175 ] ; g. jungman , m. kamionkowski and k. griest , phys .
rept . * 267 * , 195 ( 1996 ) [ hep - ph/9506380 ] .
e. komatsu _ et al .
_ [ wmap collaboration ] , astrophys . j. suppl . * 192 * , 18 ( 2011 ) [ arxiv:1001.4538 [ astro-ph.co ] ] .
s. alekhin , a. djouadi and s. moch , phys .
b * 716 * , 214 ( 2012 ) [ arxiv:1207.0980 [ hep - ph ] ] .
[ tevatron electroweak working group and cdf and d0 collaborations ] , arxiv:1107.5255 [ hep - ex ] .
f. boudjema and g. d. la rochelle , phys .
d * 86 * , 115007 ( 2012 ) arxiv:1208.1952 [ hep - ph ] .
s. caron , j. laamanen , i. niessen and a. strubig , jhep * 1206 * , 008 ( 2012 ) [ arxiv:1202.5288 [ hep - ph ] ] ; u. chattopadhyay and d. p. roy , phys . rev .
d * 68 * , 033010 ( 2003 ) [ hep - ph/0304108 ] .
s. mohanty , s. rao and d. p. roy , j. high energy phys.11 ( 2012 ) 175 arxiv:1208.0894 [ hep - ph ] ; m. a. ajaib , t. li and q. shafi , phys .
d * 85 * , 055021 ( 2012 ) [ arxiv:1111.4467 [ hep - ph ] ] ; d. feldman , z. liu , p. nath and b. d. nelson , phys .
d * 80 * , 075001 ( 2009 ) [ arxiv:0907.5392 [ hep - ph ] ] ; u. chattopadhyay , d. das , d. k. ghosh and m. maity , phys .
d * 82 * , 075013 ( 2010 ) [ arxiv:1006.3045 [ hep - ph ] ] ; d. feldman , z. liu and p. nath , phys .
d * 80 * , 015007 ( 2009 ) [ arxiv:0905.1148 [ hep - ph ] ] ; u. chattopadhyay , d. das and d. p. roy , phys .
d * 79 * , 095013 ( 2009 ) [ arxiv:0902.4568 [ hep - ph ] ] ; u. chattopadhyay and d. das , phys .
d * 79 * , 035007 ( 2009 ) [ arxiv:0809.4065 [ hep - ph ] ] ; r. m. godbole , m. guchait and d. p. roy , phys .
d * 79 * , 095015 ( 2009 ) [ arxiv:0807.2390 [ hep - ph ] ] ; u. chattopadhyay , d. das , p. konar and d. p. roy , phys .
d * 75 * , 073014 ( 2007 ) [ hep - ph/0610077 ] ; u. chattopadhyay , d. das , a. datta and s. poddar , phys . rev .
d * 76 * , 055008 ( 2007 ) [ arxiv:0705.0921 [ hep - ph ] ] ; g. belanger , f. boudjema , a. cottrant , r. m. godbole and a. semenov , phys .
b * 519 * , 93 ( 2001 ) [ hep - ph/0106275 ] .
n. fornengo , s. scopel and a. bottino , phys .
d * 83 * , 015001 ( 2011 ) [ arxiv:1011.4743 [ hep - ph ] ] ; a. bottino , f. donato , n. fornengo and s. scopel , phys .
d * 81 * , 107302 ( 2010 ) [ arxiv:0912.4025 [ hep - ph ] ] ; a. bottino , v. de alfaro , n. fornengo , s. mignola and s. scopel , astropart .
* 2 * , 77 ( 1994 ) [ hep - ph/9309219 ] ; t. k. gaisser , g. steigman and s. tilav , phys . rev .
d * 34 * , 2206 ( 1986 ) .
g. belanger , f. boudjema , a. pukhov and a. semenov , comput .
commun . * 180 * , 747 ( 2009 ) [ arxiv:0803.2360 [ hep - ph ] ] .
m. perelstein and b. shakya , arxiv:1208.0833 [ hep - ph ] ; c. beskidt , w. de boer , d. i. kazakov and f. ratnikov , jhep * 1205 * , 094 ( 2012 ) [ arxiv:1202.3366 [ hep - ph ] ] ; j. giedt , a. w. thomas and r. d. young , phys .
lett . * 103 * , 201802 ( 2009 ) [ arxiv:0907.4177 [ hep - ph ] ] ; h. ohki , h. fukaya , s. hashimoto , t. kaneko , h. matsufuru , j. noaki , t. onogi and e. shintani _ et al .
_ , phys .
d * 78 * , 054502 ( 2008 ) [ arxiv:0806.4744 [ hep - lat ] ] .
j. r. ellis , k. a. olive and c. savage , phys .
d * 77 * , 065026 ( 2008 ) [ arxiv:0801.3656 [ hep - ph ] ] .
see for example the second reference of ref.@xcite k. l. chan , u. chattopadhyay and p. nath , phys .
d * 58 * , 096004 ( 1998 ) ; [ arxiv : hep - ph/9710473 ] ; u. chattopadhyay , a. corsetti and p. nath , phys .
d * 68 * , 035005 ( 2003 ) [ arxiv : hep - ph/0303201 ] ; s. akula , m. liu , p. nath and g. peim , phys .
b * 709 * , 192 ( 2012 ) [ arxiv:1111.4589 [ hep - ph ] ] .
j. l. feng , k. t. matchev and t. moroi , phys .
d * 61 * , 075005 ( 2000 ) ; phys .
lett . * 84 * , 2322 ( 2000 ) ; j. l. feng , k. t. matchev and f. wilczek , phys . lett .
b * 482 * , 388 ( 2000 ) ; j. l. feng and f. wilczek , phys .
b * 631 * , 170 ( 2005 ) ; u. chattopadhyay , t. ibrahim and d. p. roy , phys .
d * 64 * , 013004 ( 2001 ) ; u. chattopadhyay , a. datta , a. datta , a. datta and d. p. roy , phys .
b * 493 * , 127 ( 2000 ) ; s. p. das , a. datta , m. guchait , m. maity and s. mukherjee , eur .
j. c * 54 * , 645 ( 2008 ) , [ arxiv:0708.2048 [ hep - ph ] ] . | we revisit the issue of considering stochasticity of grassmannian coordinates in @xmath0 superspace , which was analyzed previously by kobakhidze _
et al_. in this stochastic supersymmetry(susy ) framework , the soft susy breaking terms of the minimal supersymmetric standard model(mssm ) such as the bilinear higgs mixing , trilinear coupling as well as the gaugino mass parameters are all proportional to a single mass parameter @xmath1 , a measure of supersymmetry breaking arising out of stochasticity . while a nonvanishing trilinear coupling at the high scale is a natural outcome of the framework , a favorable signature for obtaining the lighter higgs boson mass @xmath2 at 125 gev , the model produces tachyonic sleptons or staus turning to be too light .
the previous analyses took @xmath3 , the scale at which input parameters are given , to be larger than the gauge coupling unification scale @xmath4 in order to generate acceptable scalar masses radiatively at the electroweak scale .
still this was inadequate for obtaining @xmath2 at 125 gev .
we find that higgs at 125 gev is highly achievable provided we are ready to accommodate a nonvanishing scalar mass soft susy breaking term similar to what is done in minimal anomaly mediated susy breaking ( amsb ) in contrast to a pure amsb setup .
thus , the model can easily accommodate higgs data , lhc limits of squark masses , wmap data for dark matter relic density , flavor physics constraints and xenon100 data .
in contrast to the previous analyses we consider @xmath5 , thus avoiding any ambiguities of a post - grand unified theory physics . the idea of stochastic superspace
can easily be generalized to various scenarios beyond the mssm .
+ pacs nos : 12.60.jv , 04.65.+e , 95.30.cq , 95.35.+d -1 cm 16.5 cm 21.5 cm -0.5 cm # 1 # 1 ( # 1,#2,#3)*#1 * , # 2 ( # 3)#1#2#3#4#1 * # 2 * , # 3 ( # 4 ) # 1#2#3jhep * # 1 * , # 3 ( # 2 ) # 1#2#3jhep * # 1 * , # 3 ( # 2 ) # 1#2#3phys .
rev .
d * # 1 * , # 3 ( # 2 ) # 1#2#3phys .
rev .
d * # 1 * , # 3 ( # 2 ) # 1#2#3phys .
rev .
lett . *
# 1 * , # 3 ( # 2 ) # 1#2#3phys .
rev .
lett . *
# 1 * , # 3 ( # 2 ) # 1#2#3nucl . phys .
b * # 1 * , # 3 ( # 2 ) # 1#2#3nucl .
phys .
b * # 1 * , # 3 ( # 2 ) # 1arxiv : hep - ph/#1 ( # 1,#2,#3)phys
. rep .
( # 1,#2,#3 ) ( # 1,#2,#3)eur .
phys .
j. c ( # 1,#2,#3 ) 22 2 ^ 2_w 2m_z^2 2b2 v#1v_#1 # 1 ^ 0_#1 # 1m_#1 # 1m_#1 # 1m_^0_#1 # 1m_^+_#1 0m_0 2wsec^2_w 2(g-2 ) _ = 11 # 1#2 # 1#1| # 1| # 1 # 1#1 # 1#1 versim#1#2 2_2^+ 2_2 ^ - 2 _ 2^ 2 m _ 2 ^ - # 1#2 * implication of higgs at 125 gev within stochastic superspace framework + * 0.5 cm manimala chakraborti@xmath6 , utpal chattopadhyay@xmath6 and rohini m. godbole@xmath7 + @xmath8 department of theoretical physics , indian association for the cultivation of science , + 2a & b raja s.c .
mullick road , jadavpur , kolkata 700 032 ,
india + @xmath9 centre for high energy physics , indian institute of science , bangalore 560 012 , india |
this paper derives the averaged mhd - equations for oscillating flows .
the resulting equations are similar to the original mhd - equations , but surprisingly ( instead of commonly expected reynolds stresses ) the _ drift velocity _ ( or just _ the drift _ ) plays a part of an additional advection velocity .
it is known that the drift can appear from either lagrangian or eulerian considerations .
the _ lagrangian drift _ appears as the average motion of lagrangian particles and its theory is often based on the averaging of odes , see @xcite . in this paper
we focus on the _ eulerian drift _ , which appears as the results of the eulerian averaging of related pdes without addressing the motion of particles , see @xcite .
the detailed materials about the eulerian drift can be found in @xcite . to derive the averaged equations we employ the two - timing method ,
e.g. _ @xcite .
we expose it as an elementary , systematic , and justifiable procedure that follows the form developed by @xcite .
this mathematical procedure is complemented by a novel material on the distinguished limit , which allows to find the proper slow time - scale .
we introduce functions of variables @xmath0 , @xmath1 , and @xmath2 , which in the text below serve as dimensionless cartesian coordinates , slow time , and fast time .
the class @xmath3 of _ hat - functions _ is defined as @xmath4 where the @xmath2-dependence is always @xmath5-periodic ; the dependencies on @xmath6 and @xmath7 are not specified . for an arbitrary @xmath8 the _ averaging operation _ is @xmath9 where during the @xmath2-integration @xmath10 and @xmath11 does not depend on @xmath12 .
the class @xmath13 of _ tilde - functions _ is such that @xmath14 the tilde - functions are also called purely oscillating functions ( _ @xmath13-function _ represents a special case of @xmath3-function with zero average ) . the class @xmath15 of _ bar - functions _ is defined as @xmath16 ( any @xmath3-function can be uniquely separated into its @xmath15- and @xmath13- parts with the use of ( [ oper-1 ] ) ) _ @xmath13-integration ( or tilde - integration ) _ : for a given @xmath17 we introduce a new function @xmath18 called the @xmath13-integral of @xmath17 : @xmath19 which represents the unique solution of a pde @xmath20 ( with an unknown function @xmath18 and a known function @xmath17 ) supplemented by the condition @xmath21 ( [ oper-2 ] ) .
the @xmath2-derivative of @xmath13-function always represents @xmath13-function .
however the @xmath2-integration of @xmath13-function can produce an @xmath3-function .
an example : @xmath22 where @xmath23 is an arbitrary @xmath15 : one can see that @xmath24 , however @xmath25 , unless @xmath26 . formula ( [ oper-7 ] ) keeps the result of integration inside the @xmath13-class .
@xmath13-integration is inverse to @xmath2-differentiation @xmath27 ; the proof is omitted .
a dimensionless function @xmath28 belongs to the class @xmath29 @xmath30 if @xmath31 and all required partial @xmath6- , @xmath1- , and @xmath2-derivatives of @xmath32 are also @xmath33 . here
we emphasize that through all the text below all large or small parameters are represented by various degrees of @xmath34 only ; these parameters appear as explicit multipliers in all formulae containing tilde- and bar - functions ; these functions always belong to @xmath29-class .
we also will use some properties of @xmath2-derivatives such as @xmath35 the product of two @xmath13-functions @xmath17 and @xmath36 represents a @xmath3-function : @xmath37 , say . separating @xmath13-part @xmath38 from @xmath39
we write @xmath40 where the notation @xmath41 for the tilde - part is introduced to avoid two levels of tildes .
we will use that the unique solution of a pde inside the tilde - class is @xmath42 which follows from ( [ oper-7 ] ) . since the average operation ( [ oper-1 ] ) is proportional to the integration over @xmath2 , then from the integration by parts we have @xmath43\rangle=-\langle[\widetilde{{\boldsymbol{a}}}_\tau,\widetilde{{\boldsymbol{b}}}]\rangle=- \langle[\widetilde{{\boldsymbol{a}}}_\tau , \widehat{{\boldsymbol{b}}}]\rangle,\ \langle[\widetilde{{\boldsymbol{a}}},\widetilde{{\boldsymbol{b}}}^\tau]\rangle=-\langle[\widetilde{{\boldsymbol{a}}}^\tau,\widetilde{{\boldsymbol{b}}}]\rangle=- \langle[\widetilde{{\boldsymbol{a}}}^\tau , \widehat{{\boldsymbol{b}}}]\rangle \label{oper-15}\end{aligned}\ ] ] where @xmath44 $ ] stands for the commutator of two vector fields @xmath45 and @xmath46 which is antisymmetric and satisfies jacobi s identity for vector fields @xmath45 , @xmath46 , and @xmath47 : @xmath48\equiv({\boldsymbol{b}}\cdot\nabla){\boldsymbol{a}}-({\boldsymbol{a}}\cdot\nabla){\boldsymbol{b}},\label{commutr}\\ & & [ { \boldsymbol{a}},{\boldsymbol{b}}]=-[{\boldsymbol{b}},{\boldsymbol{a}}],\quad [ { \boldsymbol{a}},[{\boldsymbol{b}},{\boldsymbol{c}}]]+ [ { \boldsymbol{c}},[{\boldsymbol{a}},{\boldsymbol{b}}]]+[{\boldsymbol{b}},[{\boldsymbol{c}},{\boldsymbol{a}}]]=0 \label{oper-13}\end{aligned}\ ] ] a useful property of the commutator is @xmath49=0\label{oper-11a } \ ] ] for any tilde - function @xmath50 and bar - function @xmath51 ( [ oper-15]),([oper-13 ] ) give @xmath52\rangle=[\overline{{\boldsymbol{b}}},\overline{{\boldsymbol{v}}}_a]\quad \text{where}\quad\overline{{\boldsymbol{v}}}_a\equiv\frac{1}{2}\langle[\widetilde{{\boldsymbol{a}}},\widetilde{{\boldsymbol{a}}}^\tau]\rangle\label{lem1}\end{aligned}\ ] ]
the governing equation for mhd - dynamics of a homogeneous inviscid incompressible fluid with velocity field @xmath53 , magnetic fields @xmath54 , vorticity @xmath55 and current @xmath56 is taken in the vorticity form @xmath57^*-[{\boldsymbol{j}}^*,{\boldsymbol{h}}^*]^ * = 0,\quad \text{in}\quad \mathcal{d}^*\\ & & \frac{\partial{\boldsymbol{h}}^*}{\partial { t}^*}+[{\boldsymbol{h}}^*,{\boldsymbol{u}}^*]^*=0,\quad\nonumber\\ & & \nabla^*\cdot{\boldsymbol{u}}^*=0,\quad \nabla^*\cdot{\boldsymbol{h}}^*=0\nonumber\end{aligned}\ ] ] where asterisks mark dimensional variables , @xmath58-time , @xmath59-cartesian coordinates , @xmath60 , and @xmath61^*$ ] stands for the dimensional commutator ( [ commutr ] ) . in this paper
we deal with the transformations of equations , so the form of flow domain @xmath62 and particular boundary conditions can be specified at the later stages .
we accept that the considered class of ( unknown ) oscillatory solutions @xmath63 possesses characteristic scales of velocity @xmath64 , magnetic field @xmath65 , length @xmath66 , and high frequency @xmath67 @xmath68 where @xmath69 is a dependent time - scale . in the chosen system of units the dimensions of @xmath64 and
@xmath65 coincide ; we choose them being the same order @xmath70 . the dimensionless variables and frequency are @xmath71 we assume that the flow has its own intrinsic slow - time scale @xmath72 ( which can be different from @xmath69 ) and consider solutions of ( [ exact-1 ] ) in the form of hat - functions ( [ tilde - func - def ] ) @xmath73 then the use of the chain rule and transformation to dimensionless variables give @xmath74-\frac{1}{\sigma}[\widehat{{\boldsymbol{j}}},\widehat{{\boldsymbol{h } } } ] = 0\label{exact-6-h}\\ & & \left(\frac{\partial}{\partial\tau}+ \frac{\omega}{\sigma}\frac{\partial}{\partial s}\right)\widehat{{\boldsymbol{h}}}+ \frac{1}{\sigma}[\widehat{{\boldsymbol{h}}},\widehat{{\boldsymbol{u}}}]= 0\nonumber \\ & & { \mathrm{div}\,}\widehat{{\boldsymbol{h}}}=0,\quad { \mathrm{div}\,}\widehat{{\boldsymbol{u}}}=0\nonumber\end{aligned}\ ] ] in order to keep variable @xmath1 ` slow ' in comparison with @xmath2 we have to accept that @xmath75 .
then eqn.([exact-6-h ] ) contains two independent small parameters : @xmath76 here we must make an auxiliary ( technically essential ) assumption : after the use of the chain rule ( [ exact-6-h ] ) variables @xmath1 and @xmath2 are ( temporarily ) considered to be _ mutually independent _ : @xmath77 from the mathematical viewpoint the increasing of the number of independent variables in a pde represents a very radical step , which leads to an entirely new pde .
this step should be justified _ a posteriori _ by the estimations of the error of the obtained solution ( rewritten back to the original variable @xmath7 ) substituted to the original equation ( [ exact-1 ] ) . in a rigorous asymptotic procedure with
@xmath78 one has to consider asymptotic paths on the @xmath79-plane such that @xmath80 each such path can be prescribed by a particular function @xmath81 .
one may expect that there are infinitely many different ( although some of them can coincide ) solutions to ( [ exact-6-h ] ) corresponding to different @xmath81 .
however for these equations ( as well as for many others ) a unique path can be found , which is called the _ distinguish limit _ ( or the _ distinguished path _ ) .
the notion of the _ distinguish limit _ is practical and heuristic , see @xcite ; its definition can vary for different equations and in different books and papers . for our problem
we write that _ the distinguished limit is given by such a function @xmath82 that allows to build a self - consistent asymptotic solution_. here the term _ self - consistent asymptotic solution _ means that the required successive approximations can be calculated .
these calculations include the elimination of the reducible secular in @xmath1 terms ; the _ reducible secular terms _ are such terms which can be excluded by increasing the slow time - scale .
for instance , a non - secular term proportional to @xmath83 gives secular terms in its taylor s decomposition with respect to @xmath84 ) .
below we show that the choice @xmath85 allows to build the distinguished limit solution .
the uniqueness of such a path for the considered class of solutions can be proven but we avoid such details in this paper . hence the governing equations are @xmath86-\varepsilon[\widehat{{\boldsymbol{j}}},\widehat{{\boldsymbol{h } } } ] + \varepsilon^2\widehat{{\boldsymbol{\omega}}}_s= 0 , \quad \varepsilon\equiv 1/\sigma\to 0\label{exact-6eps}\\ & & \widehat{{\boldsymbol{h}}}_\tau+ \varepsilon[\widehat{{\boldsymbol{h}}},\widehat{{\boldsymbol{u}}}]+\varepsilon^2\widehat{{\boldsymbol{h}}}_s= 0\nonumber\label{exact-6eps - h}\\ & & { \mathrm{div}\,}\widehat{{\boldsymbol{u}}}=0 , \quad { \mathrm{div}\,}\widehat{{\boldsymbol{h}}}=0\nonumber\end{aligned}\ ] ] where the subscripts @xmath2 and @xmath1 denote the related partial derivatives .
let us look for solutions of ( [ exact-6eps ] ) in the form of regular series @xmath87 in this paper we enforce the restriction @xmath88 which is natural physically if one considers , say , how the secondary vorticity develops on the background of a wave motion .
the substitution of ( [ basic-4aa-0-h]),([mean0 ] ) into ( [ exact-6eps ] ) produces the equations of successive approximations .
the equations of zero approximation are @xmath89 their unique solution ( [ f - tilde=0 ] ) is @xmath90 and @xmath91 . taking into account ( [ mean0 ] ) we can write @xmath92 which means that in zero approximation the flow is potential , purely oscillating , and the magnetic field vanishes .
this leads to the similar equations for the first approximation of ( [ exact-6eps]),([basic-4aa-0-h])-([sol-0 ] ) @xmath93 which have the unique solution @xmath94 where mean functions remain undetermined .
the equations of second approximation that take into account ( [ mean0]),([sol-0]),([sol-1 ] ) are @xmath95=0,\quad \widetilde{{\boldsymbol{\omega}}}_{2\tau}+[\overline{{\boldsymbol{\omega}}}_1,\widetilde{{\boldsymbol{u}}}_0]=0 \label{eqn-2}\end{aligned}\ ] ] which after @xmath13-integration ( [ oper-7 ] ) yield @xmath96,\quad \widetilde{{\boldsymbol{h}}}_2=[\widetilde{{\boldsymbol{u}}}_0^\tau,\overline{{\boldsymbol{h}}}_1],\quad \overline{{\boldsymbol{\omega}}}_2=\boxed{?},\quad\overline{{\boldsymbol{h}}}_2=\boxed{?}\label{sol-2}\end{aligned}\ ] ] the equations of third approximation that take into account ( [ mean0]),([sol-0]),([sol-1 ] ) are @xmath97 + [ \overline{{\boldsymbol{\omega}}}_1,\widehat{{\boldsymbol{u}}}_1]-[\overline{{\boldsymbol{j}}}_1,\overline{{\boldsymbol{h}}}_1]=0 \label{eqn-3}\\ & & \widetilde{{\boldsymbol{h}}}_{3\tau}+\overline{{\boldsymbol{h}}}_{1s}+[\widehat{{\boldsymbol{h}}}_2,\widetilde{{\boldsymbol{u}}}_0]+ [ \overline{{\boldsymbol{h}}}_1,\widehat{{\boldsymbol{u}}}_1]=0\nonumber \label{eqn-3}\end{aligned}\ ] ] the bar - part ( [ oper-1 ] ) of this system is @xmath98-[\overline{{\boldsymbol{j}}}_1,\overline{{\boldsymbol{h}}}_1]+ \langle[\widetilde{{\boldsymbol{\omega}}}_2,\widetilde{{\boldsymbol{u}}}_0]\rangle=0\label{eqn-3-bar}\\ & & \overline{{\boldsymbol{h}}}_{1s}+ [ \overline{{\boldsymbol{h}}}_1,\overline{{\boldsymbol{u}}}_1]+ \langle[\widetilde{{\boldsymbol{h}}}_2,\widetilde{{\boldsymbol{u}}}_0]\rangle=0\nonumber \label{eqn-3-tilde}\end{aligned}\ ] ] which can be transformed with the use of ( [ sol-2 ] ) and ( [ lem1 ] ) to the final form @xmath99-[\overline{{\boldsymbol{j}}}_1,\overline{{\boldsymbol{h}}}_1]=0 \label{eqn-3-bar1 - 1}\\ & & \overline{{\boldsymbol{h}}}_{1s}+ [ \overline{{\boldsymbol{h}}}_1,\overline{{\boldsymbol{u}}}_1+\overline{{\boldsymbol{v}}}_0]=0 \nonumber\\ & & \overline{{\boldsymbol{v}}}_0\equiv\frac{1}{2}\langle[\widetilde{{\boldsymbol{u}}}_0,\widetilde{{\boldsymbol{u}}}_0^\tau]\rangle\label{drift - vel}\end{aligned}\ ] ] if one uses these equations as a closed mathematical model , then all the subscripts and bars can be deleted : @xmath100-[{\boldsymbol{j}},{\boldsymbol{h}}]=0 \label{eqn-3-bar1 - 1-model}\\ & & { \boldsymbol{h}}_{s}+ [ { \boldsymbol{h}},{\boldsymbol{u}}+{\boldsymbol{v}}]=0 \nonumber\\ & & { \boldsymbol{v}}\equiv\frac{1}{2}\langle[\widetilde{{\boldsymbol{u}}},\widetilde{{\boldsymbol{u}}}^\tau]\rangle\label{drift - vel - model}\end{aligned}\ ] ] one can see that : \1 .
the equation for the oscillating velocity @xmath101 in our consideration is absent , there are only two restrictions : @xmath101 is incompressible and potential .
hence the drift velocity @xmath102 ( [ drift - vel ] ) represents a function that is ` external ' to the equations .
. the derived system of equations ( [ eqn-3-bar1 - 1 ] ) looks similar to the original one ( [ exact-1 ] ) .
one may think that ( [ eqn-3-bar1 - 1 ] ) describes ` just ' an additional advection of vorticity and magnetic field by the drift .
however , the fact that the averaged vorticity is transported with such an additional velocity is highly non - trivial ; in particular , it contains the possibility of the _ vortex dynamo _ or langmuir circulations , which we consider below .
our description of the drift differs from a classical one , therefore we first demonstrate the match of ( [ drift - vel ] ) with the classical stokes drift .
let velocity field @xmath101 and @xmath103 be @xmath104 with arbitrary @xmath15-functions @xmath105 and @xmath106 .
straightforward calculations yield @xmath107=[\overline{{\boldsymbol{p}}},\overline{{\boldsymbol{q}}}]\label{example-1 - 3}\end{aligned}\ ] ] hence the commutator is surprisingly not oscillating .
the drift velocity ( [ drift - vel ] ) is @xmath108\rangle= \frac{1}{2}[\overline{{\boldsymbol{p}}},\overline{{\boldsymbol{q}}}]\label{example-1 - 5}\end{aligned}\ ] ] the _ dimensional _ solution for a plane potential harmonic travelling wave is @xmath109 where @xmath110 are cartesian coordinates and @xmath111 is a wavenumber . in @xcite
one can see that @xmath112 where @xmath113 and @xmath114 are dimensional spatial wave amplitude and gravity ; however these physical details are excessive for our analysis . the dimensionless velocity field ( [ stokes ] ) and @xmath115 are @xmath116 where both fields ( [ example-3 - 1 ] ) are unbounded as @xmath117 , but it is not essential for our purposes .
fields @xmath118 , @xmath119 ( [ example-1 - 1 ] ) are @xmath120 the calculations ( with the use of ( [ example-1 - 3 ] ) ) yield @xmath121 the dimensional version of ( [ example-3 - 3 ] ) is @xmath122 which coincides with the classical expression for the drift velocity given by @xcite . to obtain ( [ example-3 - 3d ] )
one should take into account the difference between the time @xmath7 and @xmath123 ( [ tau - s - h ] ) .
a special case of ( [ eqn-3-bar1 - 1]),([eqn-3-bar1 - 1-model ] ) without a magnetic field is @xmath100=0.\quad{\mathrm{div}\,}{\boldsymbol{u}}=0 \label{eqn-3-bar1 - 1-e}\end{aligned}\ ] ] different @xmath1-independent versions of eqn.([eqn-3-bar1 - 1 ] ) were derived in the studies of langmuir circulation by @xcite and for the steady streaming problems by @xcite ; the methods employed by these authors are different and more cumbersome than our method . in order to demonstrate the possibility of vortex dynamo we first notice that eqn.([eqn-3-bar1 - 1-e ] ) can be integrated ( in space ) as @xmath124 where @xmath125 is a function of integration and the second equation follows from the continuity equation in ( [ exact-1 ] ) .
let the zero approximation ( [ sol-0 ] ) represent the plane potential travelling gravity wave ( [ example-3 - 1 ] ) with the drift velocity ( [ example-3 - 3 ] ) .
let cartesian coordinates @xmath126 be such that @xmath127 , @xmath128 , @xmath129 where all components are @xmath130-independent ( translationally - invariant ) , and @xmath131-variables coincide with ones in ( [ example-3 - 1 ] ) .
then the component form of ( [ eqn-3-bar1-vel ] ) is @xmath132 which can be rewritten as ( see @xcite ) @xmath133 where @xmath134 , @xmath135 , and @xmath136 is a modified pressure .
one can see that ( [ strat - eqns ] ) is mathematically equivalent to the system of equations for an incompressible stratified fluid , written in boussinesq s approximation .
the effective ` gravity field ' @xmath137 is non - homogeneous that makes the analogy with a ` standard ' stratified fluid non - complete .
nevertheless one can see that any increasing function @xmath138 ( taken from the shear flow @xmath139 ) produces ` taylor instability ' of an inversely stratified equilibrium .
it leads to the growth of longitudinal vortices and can be connected to langmuir circulations , see @xcite .
the consideration of this paper is based on the assumption that the enforced frequency @xmath67 ( [ scales ] ) of oscillations is higher than all intrinsic frequencies .
this frequency appears in our theory via the prescribed potential velocity @xmath101 ( [ eqn-0 - 0-h ] ) .
the prescribed oscillatory velocity @xmath101 can be caused by different factors .
for example , it can be produced by oscillations of boundaries or appear in full viscous theory after the matching of external flow with boundary - layer solution .
the latter option is often considered , see @xcite . \3 . to justify the distinguished limit ( [ tau - s - h ] ) mathematically
, one should prove that any different path @xmath81 produces an asymptotic solution which contains terms secular in @xmath1 or does not produce any asymptotic solution at all .
the following statement can be proven : for @xmath140 ( [ drift - vel ] ) and the function @xmath141 ( with a constant @xmath142 ) all solutions with @xmath143 contain secular terms , while all equations with @xmath144 produce a controversial ( unsolvable ) equations of successive approximations . if @xmath145 then the statement is different but we do not describe it here .
the proof is omitted .
the higher approximations of the averaged equation ( [ eqn-3-bar1 - 1 ] ) can be derived .
they are especially useful for the study of motions with @xmath145 .
in particular , one can show that in this case langmuir circulations cab be still generated by a similar mechanism .
the viscosity and diffusivity can be routinely incorporated in ( [ eqn-3-bar1 - 1 ] ) as the rhs - terms @xmath147 and @xmath148 .
accordingly , viscous and diffusion terms will appear in the equations ( [ eqn-3-bar1-vel ] ) and ( [ strat - eqns ] ) .
at the same time after the incorporation of viscosity one more small parameter appears in the list ( [ small - pars - h ] ) , and the distinguished limit should be reconsidered .
the abolishing of the requirement of a vanishing mean flow in zeroth approximation ( [ mean0 ] ) is also straightforward .
however in this case the distinguished limit ( [ tau - s - h ] ) is different and the resulted averaged equations are more complicated than ( [ eqn-3-bar1 - 1 ] ) . \11 . for the finite and time - dependent flow domain
@xmath149 the definition of average ( [ oper-1 ] ) directly works only if @xmath150 at any instant .
if at some instant @xmath151 then the theory should include a ` projection ' of the boundary condition on the ` undisturbed ' boundary .
such a consideration requires the smallness of the amplitude @xmath152 of spatial oscillations of fluid particles .
however one can see that @xmath153 and hence @xmath154 ( [ small - pars - h ] ) .
therefore the considering of a time - dependent domain does not introduce any new small parameter , and the distinguished limit ( [ tau - s - h ] ) will stay the same . in particular , it means that our small parameter @xmath155 is the same as the dimensional slope of free surface in the theory of langmuir circulations by @xcite
the determining of the function @xmath138 in ( [ strat - eqns ] ) ( for real langmuir circulations ) requires an additional theory which includes viscosity or turbulent tangential stresses at the free surface .
it is interesting , that from this viewpoint an ` unstable stratification ' can be continuously generated and amplified by tangential stresses applied at free surface .
we do not compete here with the theories by @xcite , which describe this phenomenon well .
one can suggest that since the equations ( [ eqn-3-bar1 - 1 ] ) for @xmath156 do describe a mechanism of vortex dynamo , and the mathematical structure of the full averaged equations ( [ eqn-3-bar1 - 1]),([eqn-3-bar1 - 1-model ] ) with @xmath157 is similar , then these full equations could also describe a possible mechanism of mhd - dynamo , such as the generation of the magnetic field of the earth .
the author thanks the department of mathematics of the university of york for the research - stimulating environment .
the author is grateful to profs .
a.d.d.craik , k.i.ilin , s.leibovich , h.k.moffatt , a.b.morgulis , n.riley , and v.a.zheligovsky for helpful discussions .
vladimirov , v. a. , moffatt , h. k. , and ilin , k.i .
1996 on general transformations and variational principles for the magnetohydrodynamics of ideal fluids .
stability criteria for two - dimensional flows
.. _ j. fluid mech .
_ , * 329 * , 187 - 207 . | we have derived the closed system of averaged mhd - equations for general oscillating flows , which are purely oscillating in the main approximation .
we have used the mathematical approach which combines the two - timing method and the notion of the distinguished limit .
properties of the commutators are used to simplify calculations . the direct connection with a vortex dynamo ( or the langmuir circulations ) has been demonstrated and a conjecture on the mhd - dynamo has been formulated . |
an important goal of modern cosmology is to understand the evolution of active galaxies and their relationship to normal galaxy evolution . a key piece of information needed to further this understanding is an accurate knowledge of the luminosity function ( lf ) for agn over a wide range of absolute magnitudes and redshifts .
the agn lf is populated by quasars at the bright end ( e.g. hartwick & schade 1990 ) and seyfert galaxy nuclei , considered to be their intrinsically fainter counterparts , at the low luminosity end ( cheng @xmath7 1985 ; huchra & burg 1992 ; maiolino & rieke 1995 ; koehler @xmath7 1997 ; ulvestad & ho 2001 ; londish @xmath7 2002 ) . understanding how the faint end of the agn lf evolves is of particular importance for determining the frequency and total space density of agns at earlier epochs .
this has obvious implications for determining their total contribution to the x - ray , ir and uv backgrounds .
large numbers of low - luminosity agn have been proposed to explain the ionization of the intergalactic medium at high redshift ( steidel & sargent 1989 ) , although recent observational and theoretical results differ on the amount of their true contribution ( barger 2003 ; schirber & bullock 2003 ) .
additionally , the faint end is an important constraint on evolutionary models for the agn lf such as pure luminosity and luminosity - dependent density evolution ( e.g. boyle @xmath7 2000 ) .
one method to identify agn in imaging surveys is via their variable nature .
qsos and seyfert galaxies have long been known as variable objects , with significant optical flux changes occurring on timescales of months to several years .
variability has been a successful method for identifying agns , primarily qsos ( hawkins 1986 ; koo @xmath7 1986 ; hook @xmath7 1994 ) .
a survey for variable sources in sa57 revealed many optically extended , seyfert - like galaxies ( bershady @xmath7 1998 ) which generally had higher variability amplitudes than the more luminous qsos .
this result suggests that variability is a good technique for selecting intrinsically faint qsos and seyfert nuclei . while variability and other selection techniques ( spectroscopy , color selection , etc . )
have been successful in finding qsos out to high redshifts , the intrinsically fainter agn ( i.e. seyfert galaxies ) have been difficult to detect beyond the local universe .
ground - based variability surveys , like spectroscopic surveys , are quickly limited to agn which dominate the host galaxy light at higher redshift .
recent x - ray surveys with chandra and xmm - newton have revealed hundreds of x - ray sources at z@xmath01 .
many of these are confirmed agn but still many more are ambiguous in nature ( e.g barger @xmath7 2002 ) . in summary ,
the population of moderate - to - high redshift , low - luminosity agn has not been well surveyed and the evolution of the faint end of the agn lf is still poorly known . in this study , we have searched for optical nuclear variability in hdf galaxies to identify the population of active galaxies out to z@xmath31 .
the primary data sets analyzed for galactic variability are the original hdf observations ( williams @xmath7 1996 ) obtained in december 1995 and a second v - band ( f606w ) image obtained almost exactly 5 years later in december of 2000 .
the main advantage of hst over ground - based surveys is the ability to do accurate photometry within smaller apertures than can be done from the ground thus allowing us to probe much lower agn / host galaxy luminosity ratios .
we discuss the details of the image processing and photometry in section 2 , followed by a discussion of the selected varying galactic nuclei in sections 3 and 4 . in section 5 , we cross - reference our sources with various multi - wavelength surveys .
a discussion of the optical spectra for the varying sources is presented in section 6 . in section 7
, we present the luminosity function for variable nuclei .
we close with a summary of our conclusions in section 8 .
the original epoch hdf data in f606w consisted of 108 individual exposures with a total integration time of 113,050 seconds .
these data were obtained with a total of 9 unique dithers covering both sub - pixel offsets and multiple pixels on a scale of about @xmath8 10 wf pixels .
the primary data processing technique was detailed in gilliland , nugent , and phillips ( 1999 ) for creation of over - sampled , by a factor of four , summed images with resulting noise and resolution properties competitive with , and sometimes slightly better than those reported in williams @xmath7 ( 1996 ) .
the mean epoch of the original f606w data was mjd 50076.9 .
exposure times used for the original epoch ranged from 350 to 1500 seconds with a mean sky rate of 0.0758 e-/s . the second epoch hdf data in f606w was taken with go-8389 ( pi rodrigo ibata ) with a science goal distinct from that here , but one that also required precise differential measurements across the two epochs . these data consisted of 70 individual exposures each of 1300 seconds for a total exposure time of 91,000 seconds .
the goal was to have the new data obtained at the same pointing as the original hdf , and at the same orientation , and with dithering over the same spatial scale , but with an improved level of sampling using 35 unique dithers .
this goal was quite well met , thus resulting in quite similar data in the two f606w epochs .
the mean epoch of the second f606w data was mjd 51882.9 implying a delta of 4.94 years .
the mean sky rate for the newer epoch was 0.0958 e-/s .
processing of the second epoch data was identical to that of the first , using the same software as described in gilliland , nugent and phillips ( 1999 ) . to facilitate simple , robust analysis of data from the two epochs the registrations of the second epoch
were evaluated using a grid of galaxies from the original epoch ( stars could not be used since these show unique proper motions over five years ) ; this results in second epoch over - sampled combined images that are accurately aligned to the first epoch with _ maximum _ errors of @xmath30.05 pixels .
therefore , comparative analysis for change across epochs can be made using precisely the same pixels , mapping back to common physical domains .
for both epochs , each individual frame was sky subtracted before combining , cosmic rays were eliminated ( effected pixels assigned zero weight in the combination ) and hot pixels subtracted . the extreme stability provided by _
hst _ is an essential factor in being able to probe to unique sensitivity levels with these extensive observations . limiting factors , to be discussed further below ,
arise from ( 1 ) a time - varying charge transfer efficiency ( cte ) that must be empirically corrected for , and ( 2 ) limited sub - pixel dithering in the first hdf epoch that results in minor mismatches noticeable for the objects with tight cores at high signal to noise .
we created a catalog of hdf objects using daofind in iraf to ensure a systematic selection of all sources in the v - band images .
all objects above the 8@xmath9 detection threshold were identified in the 1995 images ( the deeper of the two ) using a kernel size of 1.8@xmath10 the semi - major axis of the gaussian convolution kernel ( 1.3 pixels ) to allow more sensitivity to extended objects .
the catalog was then edited to remove a minority of objects that did not appear to be centered on galaxies ( i.e. multiple detections within single galaxies ) .
for very amorphous galaxies with no clear center , all bright knots " appearing within the source were retained in the catalog . for our photometry , we chose a small aperture consistent with the fwhm of stellar sources in the field
. stars have [email protected] pixels in the over - sampled images corresponding to @xmath30.14@xmath11 .
we chose an aperture size of 6 pixels in diameter to be most sensitive to photometric changes in an unresolved nucleus .
this approach differs from that taken in sarajedini , gilliland & phillips ( 2000 ; hereafter sgp ) , where the aperture size was scaled to the size of the galaxy .
our goal here is to provide better sensitivity to variable active nuclei with less dilution from the underlying , non - varying host galaxy light .
aperture photometry for all sources was performed on the direct images using the phot algorithm in iraf .
because the second epoch image was aligned with the first , exactly the same pixels were used in both epochs . during image processing ,
the background was subtracted and the images were scaled to an exposure time of 6000s .
this allowed the sky value to be held constant at zero for both epochs and all wf ccds .
the resulting nuclear magnitudes for each galaxy in the hdf were then adjusted by a zero - point offset to bring them into agreement with the integrated , isophotal galaxy v - band ab magnitudes of williams @xmath7 ( 1996 ) . after performing aperture photometry , nuclear variability
can then be determined by simply obtaining the difference in magnitude for each source over the two epochs .
figure 1 shows the magnitude difference between the 1995 and 2000 images as a function of magnitude within an aperture of r@xmath23 pixels , the equivalent of 0.15@xmath11 in diameter .
the solid line represents no difference in magnitude between the two epochs .
the main culprit producing the obvious offset from the solid line is charge transfer efficiency ( cte ) losses . as in sgp
, we find a clear correlation of the photometric differences with position on the ccd and magnitude .
this is the well known cte effect ( whitmore @xmath7 1999 ; biretta @xmath7 2001 ) which causes targets far from the ccd readout amplifier to appear fainter than similar targets near the amplifier .
for this reason , the effect is usually more significant in the y - direction , ( the direction of the readout down the ccd ) than the x - direction ( the direction of the readout of the shift register at the bottom of the ccd ) .
although the 19951997 i - band study did not show significant correlations in the x - direction , the longer time interval sampled here does show a significant correlation in the x - direction for all 3 wf ccds . additionally , we note that the slope of the relation is generally steeper in the y - direction than had been determined in sgp .
this is expected due to the longer 5 year time interval between epochs in this study .
we first attempted to fit the relationship between object position , magnitude difference and magnitude globally for all 3 wf ccds as done in sgp .
however , we found that this approach was not sufficient to equally remove the cte effect from all wf ccds .
differences in the slope of the relation between the ccds were significant and could easily be determined for each ccd individually .
we adopted a two - step approach to determine this relationship .
first , we fit a linear surface to the x and y object positions and the magnitude differences using surfit in iraf for each ccd . while the slope of the fit was found to be significant with both the x and y positions , the value of the slope was steeper with y position in all 3 wf ccds , consistent with what is expected for cte losses .
the slope of the relation between y position and magnitude difference was found to be -2.6@xmath1010@xmath12 , -1.3@xmath1010@xmath12 and -1.46@xmath1010@xmath12 mag / pixel for wf2 , wf3 and wf4 respectively .
the slope of the relation with x position , while significant , was found to be about an order of magnitude lower in all 3 wf ccds .
these fits were used to correct the magnitude difference values for position - dependent offsets .
secondly we fit the dependence of the magnitude difference with nuclear magnitude .
a 3rd order fit was required to properly model the dependence in all 3 wf ccds . after applying this correction to the data in figure 1 ,
the resulting photometric differences are shown in figure 2 as a function of nuclear magnitude . finally , we determine the expected noise level for non - variable sources . as in sgp , we divided the full set of individual exposures into two sets composed simply of the odd - numbered frames and the even - numbered frames .
each set represents an epoch " without any real time difference .
these images effectively carry through the effects of object and sky poisson noise , readout noise , and possible errors in the adopted sky zero points .
the magnitude of each galaxy was then measured in the odd and even data sets and difference in magnitude vs. the average magnitude within r@xmath23 pixel apertures is shown in figure 3a .
figure 3b is the rms of the intensity differences shown in a ) within unity magnitude intervals .
the solid line represents a quadratic fit to the points which is the adopted 1@xmath9 galaxy photometric error as a function of magnitude .
the precision of the photometry must be independently limited at the bright end of the distribution , since systematic effects not captured by the odd - even data set comparisons become limiting factors ( e.g. the sometimes poor distribution of sub - pixel phase - space dithers in the first epoch data ) . in sgp
, we adopted a floor of 1.2% . here
we adopt a more conservative value of 1.5% consistent with the use of much smaller apertures on average for bright sources in the current study than in sgp .
as should be apparent , this is a reasonable limit for photometric precision to adopt , but in detail remains somewhat arbitrary .
figure 4a is the absolute value , cte - corrected magnitude differences for sources in the hdf .
the solid line represents the 3@xmath9 limit for variability significance ( 3@xmath10 the rms indicated by the solid line in figure 3b ) .
objects above this limit are selected as significant variables and are indicated with open hexagons .
figure 4b is the result of normalizing the magnitude difference by this solid line . here
the y - axis indicates the level of significance of each source in units of @xmath9 .
the x - axis extends at the faint end to v@[email protected] , which is the estimated photometric completeness magnitude limit for galaxy nuclei in this survey . beyond this limit ,
the number counts for galaxies in the hdf begin to decrease .
we find sixteen galaxies whose nuclei have undergone at least a 3@xmath9 variation over the 5 year time interval .
these sources are listed in table 1 with columns as follows : ( 1 ) williams @xmath7 ( 1996 ) i d , ( 2 ) & ( 3 ) ra & dec ( j2000 ) , ( 4 ) redshift from the literature , ( 5 ) spectral type based on photometry from fernndez - soto , lanzetta and yahil ( 1999 , hereafter : fly ) , ( 6 ) v@xmath1 internal to r=3 pixel aperture , ( 7 ) magnitude difference between 1995 and 2000 ; positive implies brighter in 1995 , ( 8) significance of change obtained when normalized by the expected error as a function of magnitude , ( 8) bulge - to - total ( b / t ) 2-dimensional model fits for the i(f814w ) images from marleau & simard ( 1998 ) .
the last several columns relate to the source detection at other wavelengths discussed in section 5 .
there are two distinct completeness issues that effect our survey .
the first is related to the incomplete time - sampling of the variable sources we wish to detect .
most variability surveys employ a method where the survey field is imaged several times over many years ( e.g. trevese @xmath7 1994 ; hawkins 2002 ) . depending on the quality of the data
, these surveys have shown that virtually all known agn will be found to vary if observed periodically over several years .
our study is limited by the fact that we have only two epochs with which to determine variability and therefore sample just two points on the lightcurve of a varying source .
because of this , we will be incomplete in our census of agn since some varying sources could lie at magnitudes close to their original magnitude measured 5 years earlier and would thus go undetected in our survey .
we estimate our incompleteness due to undersampling of the lightcurve by using variability data for agn obtained over many years .
we randomly select points along the lightcurve separated by 5 years , the time interval sampled by our hdf images .
because long term variability surveys for low - luminosity agn have not yet been published , we have conducted this test with two sets of qso lightcurve data ; a sample of pg quasars from giveon @xmath7 ( 1999 ) and a sample of sa57 quasars from trevese @xmath7 ( 1994 ) .
giveon @xmath7 monitored 42 quasars over 7 years with a typical sampling interval of 40 days while trevese @xmath7 monitored 64 over 15 years about once a year . sampling these lightcurves every 5 years and assuming photometric errors typical of our hdf images , we can estimate the probability that these sources would be detected in our survey .
we find that @xmath37580% would be detected as varying by at least 3@xmath9 with only 2 observations separated by 5 years .
therefore , the results of our analysis are likely to underrepresent the true number of variable nuclei by @xmath32025% . the second incompleteness issue results from the use of a fixed aperture to detect nuclear variability .
we have chosen a small , fixed aperture to minimize dilution of the agn light from the underlying host galaxy to be more sensitive to agn varying within bright hosts .
we set this aperture to the size of an unresolved point source in the hst images ( 0.15@xmath11 in diameter ) .
the fixed aperture will include the same amount of light from an unresolved source regardless of its redshift .
however , the aperture will contain a larger fraction of the underlying , resolved galaxy for higher redshift objects than for low - z sources . because of this , the dilution of the nuclear light by the galaxy increases as a function of redshift and , consequently , as a function of nuclear magnitude .
if we assume a disk - dominated , r@xmath13=1 kpc galaxy containing an agn that is 10% of the host galaxy light and varying by @xmath14m=0.3 magnitudes , the measured @xmath14 m within our fixed aperture would be 0.15 magnitudes at z=0.2 but only 0.06 magnitudes at z=1 .
the observed total nuclear magnitude would drop by @xmath33 magnitudes over this interval .
the increasing dilution with redshift is compounded by the increasing photometric errors with magnitude . for an agn that has varied by @xmath14m=0.30.2 mags ,
consistent with typical structure function values for agn ( trevese @xmath7 1994 ) , and reasonable estimates for the galaxy magnitude , morphology , physical size and true agn fraction , we estimate that incompleteness due to significant agn dilution begins to effect our survey at magnitudes between v@[email protected] to v@[email protected] . as can be seen in figure 4 ,
all of the variables lie at magnitudes greater than v@xmath1=27.3 . while this may partially be due to a real astrophysical effect ( see discussion in section 4 )
, this limiting magnitude is within the range of our incompleteness estimate above .
therefore , although we survey galaxy nuclei to v@xmath1@xmath029 , our limit for detecting _ variable agn _ at a uniformly varying level within galaxy nuclei is probably closer to v@[email protected] , with decreasing completeness fainter than v@[email protected] .
the 16 variable nuclei detected in our survey represent 2.2% of the galaxies brighter than v@[email protected] or 7.4% of the galaxies brighter than v@[email protected] . to emphasize
the significance of the variable sources , figure 5 is a histogram of the sigma distribution for galaxies in the hdf .
the x - axis is the absolute value of the normalized magnitude difference ( @xmath9 ) and the y - axis is the natural logarithm of the number of sources in 0.25@xmath9 bins .
errorbars represent the poisson statistical errors .
filled circles show the histogram for all 719 nuclei brighter than v@[email protected] and open squares represent the 217 nuclei brighter than v@xmath1=27.5 .
the curved lines are gaussian fits to the data within 2.5@xmath9 ( solid is the fit to data brighter than v@[email protected] and dashed is the fit to data brighter than 27.5 ) . in both distributions ,
the data are well fit by gaussians out to @xmath33@xmath9 and show a `` tail '' of significant variables extending to higher @xmath9 values . in a normal distribution , we would expect a total of @xmath35 sources to be greater than 3@xmath9 brighter than v@[email protected] with @xmath31.5 expected to lie brighter than v@[email protected] . based on gaussian statistics , we estimate that @xmath312 of our variables ( all detected brighter than v@[email protected] ) are spurious , implying that the true number of variables is @xmath314 .
if we also correct by 2025% for incompleteness due to lightcurve sampling , the incompleteness corrected number of variables in the hdf is @xmath318 representing 8.3% of the nuclei brighter than v@[email protected] .
one of the 16 variables detected in this survey , _ 2 - 251.0 _ , was also identified as variable in the i - band variability survey of the hdf ( sgp ) .
eight variables were detected in that survey with significant ( @xmath153@xmath9 ) magnitude changes over the 2 year time interval ( 19951997 ) .
object _ 2 - 251.0 _ is by far the brightest of these sources , having a nuclear v magnitude @xmath313@xmath10 brighter than the next brightest source . of the 7 i - band variables not detected in the present survey , 5 are fainter than v@xmath1=28.0 , placing them below the expected sensitivity limits for this study .
the other two i - band variables , _ 3 - 266.0 _ and _ 3 - 404.0 _ , have v@[email protected] .
in the present study , object _ 3 - 266.0 _ changed by 0.025 mag ( 1.1@xmath9 significance ) and object _ 3 - 404.0 _ changed by 0.052 mag ( 2.0@xmath9 significance ) .
the latter source is a marginally significant detection and would lie above the 3@xmath9 limit if it had been brighter than @xmath326.5 in the v - band .
3 - 266.0 _ , at just over 1@xmath9 significance , may be variable but at lower significance due to the 2-epoch statistical incompleteness effect discussed above .
figure 6 contains the 1995 v - band images for galaxies with variable nuclei .
the postage stamp images are 3@xmath11 on a side and are scaled to the same maximum pixel value .
it can be seen from this figure and table 1 that the selected galaxies cover a range of morphologies , magnitudes and colors .
their redshifts range from 0.09 to 1.8 with a median @xmath4z@[email protected] .
many of the galaxies have significant bulge components based to the 2-dimensional model fits .
we note , however , that several sources with b / t@xmath20 also have high chi - square values ( @xmath161.25 ) which in some cases appears to be due to a poorly fit bulge component .
in addition to the morphological b / t parameter , we also list the spectral type for the source based upon the photometry of fly to determine photometric redshifts for galaxies in the hdf .
the selected galaxies appear to be evenly distributed over all spectral classes with no particular preference towards early or late - type galaxies .
figure 7 is a color magnitude diagram of the galaxy nuclei for sources in the hdf .
the variable nuclei are indicated with larger solid triangles .
the variables cover the full range of v
i colors and do not appear to be preferentially bluer or redder than the non - varying nuclei of galaxies in the hdf .
all of the variables lie at the bright end of the magnitude distribution .
as previously discussed , incompleteness due to increasing agn light dilution as a function of redshift is likely to contribute to this observation
. however , an astrophysical effect may be at least partly responsible .
agn are generally found in brighter host galaxies than those that do not harbor agn ( huchra & burg 1992 ; ho @xmath7 1997 ; hamilton @xmath7 2002 ) .
if these sources are indeed agn , we might expect to find them in the brighter galaxies .
an additional factor is the inclusion of the agn light in the nuclear magnitude .
we can minimize the above mentioned incompleteness effect by considering only those sources brighter than v@xmath1=26.5 , where incompleteness is not expected to be significant .
we find that 23% ( 5/22 ) of nuclei are variable at 21.8@xmath4v@[email protected] while only 15% ( 6/39 ) are variable at 25.5@xmath4v@[email protected] .
while the small numbers do not provide significance , this trend is consistent with the previous studies concerning agn host galaxies .
the hdf has been surveyed with a variety of other telescopes and instruments providing a wonderful resource of multi - wavelength information for this particular region of the sky .
we have cross - referenced our list of variable nuclei with published catalogs of x - ray , radio and far - infrared sources as well as other photometric and spectroscopic surveys used to search for nuclear activity . agn have long been recognized as x - ray sources , being associated with a wide range of activity levels from the brightest quasars to low - luminosity seyferts .
even highly obscured agn can often be detected at x - ray wavelengths .
some of the light obscured by dust at other wavelengths will be reprocessed and emitted in the mid - ir . for these reasons ,
the x - ray and mid - ir regimes are excellent wavelengths in which to detect and study agn .
recent deep surveys with the chandra x - ray observatory have resolved many x - ray sources in the hdf and provide information about the nature of the x - ray radiation . the 2ms chandra x - ray survey (
alexander @xmath7 2003 ) has detected 20 sources in the original hdf and an additional 2 sources of lower significance which are listed in their supplementary catalog .
of these 22 detections , 18 are included in our variability study .
two were not included because their optical fluxes were too faint to be included in our survey and an additional pair fell too close to the edge of the ccd to obtain accurate nuclear photometry in both epochs .
of these 18 sources , we find seven that match ( within 1.2@xmath11 ) the positions of our variable nuclei ( see columns 10 and 11 of table 1 ) .
the filled symbols in figure 4 represent the 18 x - ray sources included in our survey and show the level of variability significance for each source .
in addition , we have found several variables coincident with source positions from the isocam 15@xmath5 survey of the hdf ( aussel @xmath7 1999 ) and the 1.4ghz radio survey ( richards @xmath7 1999 ) .
these are listed in columns 12 through 14 of table 1 .
below , we discuss the variable galaxy nuclei that overlap with one or more of these multi - wavelength surveys .
_ 2 - 251.0 : _ this z=0.960 early - type spiral galaxy is one of only two broad - line agn in the hdf ( see section 6 ) and is a bright x - ray source detected with chandra in the soft , hard and ultrahard bands ( hornscheiemer @xmath7 2000 ; brandt @xmath7 2001a
hereafter h00 , b01a ) with a photon index of 0.67 .
we detect it as an optical variable at 6.5@xmath9 significance . as previously mentioned , it is the only object that was also detected in our i - band survey of the hdf ( sgp ) .
spectroscopic detection of broad mgii ( phillips @xmath7 1997 ) , as well as 1.4ghz radio ( richards @xmath7 1999a ) and isocam ( aussel @xmath7 1999 ) detections for this source all corroborate the agn nature of this galaxy .
spectral fitting of the x - ray data has revealed a large intrinsic column density which appears to be related to the agn ( b01a ) . _
3 - 355.0 : _ this elliptical galaxy at z=0.474 is a 3.2@xmath9 variable .
the original 166ks chandra exposure did not detect this source in the hard band , but b01a report hard x - ray counts from the deeper 479.7ks exposure .
the photon index determined from the 2ms exposure is 1.8 .
this galaxy is also a radio source with a steep radio spectrum ( @xmath17=1.0 ) .
its position is coincident with an iso source from the aussel @xmath7 ( 1999 ) supplementary table of lower significance detections .
h00 note that the x - ray luminosity for this source is consistent with that expected from hot gas in an elliptical but the nuclear variability detected here suggests this galaxy may also host an agn .
_ 3 - 659.1 : _ the variable source is a spiral galaxy at z=0.401 ( barger @xmath7 2002 ) with a variability significance of 4@xmath9 .
this optical galaxy is @xmath31.2@xmath11 from the hardest x - ray source ( @xmath18=0.56 ) in the hdf .
the x - ray position is only 0.22@xmath11 from the second reddest source in the hdf nicmos survey ( dickinson @xmath7 2000 ) .
even though the optical galaxy is positioned further from the center of the x - ray emission , the detected variability suggests that perhaps some of the x - ray emission originates from the galaxy nucleus .
there is also significant 1.4ghz radio emission ( richards @xmath7 1999a ) and 15@xmath5 emission detected with iso ( aussel @xmath7 1999 ) .
the radio position is again more closely aligned with the nicmos source than the optical galaxy ( see figure 1 in richards 1999b ) but both the nicmos and optical sources are within the positional uncertainties . _
3 - 965.111111 : _ this is the most significant variable in our survey ( 9.4@xmath9 ) .
it was not detected in either the radio or mid - ir bands and was first detected in the 1ms chandra survey ( brandt @xmath7 2001b ) .
it was detected in both the soft and hard x - ray bands in the 2ms survey and has a photon index of 0.91 .
it is a bright elliptical galaxy near the edge of the hdf at z=0.663 . _ 4 - 752.1 : _
this optically variable ( 3.1@xmath9 ) x - ray source is also an fr i radio galaxy ( richards 2000 ) with extensive radio structure .
the host galaxy is a red elliptical at z=1.05 .
this source has not been detected in the hard x - ray band and thus appears to fall into the low x - ray luminosity / soft x - ray spectrum grouping of very red objects ( hogg @xmath7 2000 ) . because of its faintness , the photon index can not be determined .
b01a suggests that the x - ray emission is associated with the central agn or the hot interstellar medium and x - ray binaries of the elliptical galaxy .
it is not detected in the mid - infrared .
_ 4 - 976.1 : _ this bright spiral galaxy at z=0.089 is variable at the 3.6@xmath9 significance level .
b01a notes that the position of the x - ray source is not coincident with the galaxy nucleus ( which is 1.1@xmath11 away ) but does lie near ( @xmath30.14@xmath11 ) an off - nuclear bright spot .
they suggest that this bright spot may be a background agn , starbursting region , or super eddington " x - ray binary ( b01a , h00 ) .
our variability measurements have been made on the central nuclear region of this galaxy which , if hosting an agn , may be responsible for at least part of the x - ray emission .
the photon index for this source is 1.7 .
this galaxy is within 3.3@xmath11 of a mid - ir source ( aussel @xmath7 1999 ) . _
3 - 386.111 : _ this is a spiral galaxy at z=0.474 which is listed in the 2ms x - ray survey supplementary catalog containing lower significance sources associated with optically bright objects .
the nucleus has a variability significance of 3.8@xmath9 .
this source was not detected in the hard x - ray band .
it is associated with an iso source listed in the supplementary iso source catalog . one additional variable nucleus , _ 4 - 254.0 _ , matched the position of a lower significance x - ray source indicated in the 1.38ms chandra survey ( brandt @xmath7 2002 ) .
this x - ray source , however , was not listed among the 2ms detections .
the object is an elliptical galaxy at z=0.901 which is also associated with an iso detection in the supplementary catalog . of the remaining 8 variable nuclei , one ( _ 2 - 860.0 _ )
is associated with a mid - ir detection and another ( _ 3 - 943.0 _ ) is within @xmath32@xmath11 of a 1.4ghz radio source . the other 6
are not associated with any of the multi - wavelength surveys discussed here .
overall , 44% ( 7/16 ) of our variable galaxy nuclei have x - ray counterparts from the 2ms chandra survey ( or 31% excluding the two matches with greatest positional offsets ) .
optical variables make up 39% of the x - ray sources . of the brightest x - ray sources ( full - band flux @xmath192.4@xmath1010@xmath20 ergs / s )
five out of nine ( 56% ) are associated with optical variables .
if we assume that the observed optical variability indicates the presence of an agn , this is an important check on the nature of the x - ray emitting source and could potentially help discriminate between low - luminosity x - ray emitting agn and other x - ray emitting phenomena such as supernova remnants or x - ray binaries .
there does not appear to be a relation between photon index and the detection of optical variability .
five of the eight x - ray sources in the hdf with enough signal to determine photon indices are optical variables .
the x - ray / optical variables make up two of the softer x - ray sources and three of the harder sources in this small sample .
seven of our 16 variables are associated with mid - ir detections ( 44% ) and five ( 31% ) have radio emission at 1.4ghz .
variable galaxies make up 22% of the detected sources at 15@xmath5 .
this is comparable to the portion of the mid - ir integrated light attributed to agn based on the correlation of mid - ir sources with chandra sources ( elbaz @xmath7 1999 ) .
finally , we have compared our results with two photometric studies to identify qsos in the hdf on the basis of multi - band colors .
jarvis & macalpine ( 1998 ) identified 12 high redshift ( [email protected] ) qso candidates .
eleven of these sources were included in our variability survey but none showed significant variability amplitudes , having a median @xmath9 of @xmath30.6 . based on structure functions for agn / qsos ( trevese @xmath7 1994 ; hook @xmath7 1994 ; hawkins @xmath7 2002 ) and taking into account time - dilation effects ( @xmath14t@xmath21=@xmath14t/(1+z ) ) , the average magnitude change expected in our survey for high - z qsos is @xmath30.10.2 magnitudes .
even sources dominated by an agn component , as expected for these color - selected candidates , would be difficult to detect above the 3@xmath9 significance threshold due to their faint apparent magnitudes ( v@[email protected] ) .
nonetheless , an average magnitude change of only @xmath30.05 is significantly less than the 0.10.2 average magnitude change predicted from qso structure functions .
conti @xmath7 ( 1999 ) also identify 20 compact sources having qso - like colors and morphologies with estimated redshifts between z@xmath31 and 5.5 .
the nuclear magnitudes for these objects lie at v@xmath1@xmath02728 .
none were identified as variable above the 3@xmath9 threshold in our survey .
the average nuclear magnitude of this sample is @xmath31 magnitude brighter than the jarvis & macalpine qsos .
if these are variable qsos / agn - dominated galaxies , we would again expect an average magnitude change of @xmath30.10.2 .
the average change in magnitude for the conti @xmath7 sample over the 5 year interval is @xmath30.03 magnitudes , with all sources lying well below the 2@xmath9 significance limit .
the extensive redshift surveys of the hdf from cohen @xmath7 ( 1996 ; 2000 ) , phillips @xmath7 ( 1997 ) and barger @xmath7 ( 2002 ) have revealed only two broad - line agns ( blagns ) in the hdf proper . these are _ 2 - 251.0 _ ( z=0.96 variable galaxy and x - ray source discussed above ) and _ 4 - 852.12 _ ( z=0.943 x - ray source )
. the measured variability significance of _ 4 - 852.12 _ is 0.35@xmath9 .
it is possible that this source is variable but has been observed at two points in its light curve that are close to the same magnitude .
further monitoring of the hdf would be necessary to rule out optical variability for this blagn .
we have studied the available spectra for all of the galaxies hosting variable nuclei to determine if any of the sources show specific emission lines or line flux ratios indicative of type 2 agn . of the 16 variables ,
optical spectra exist for 13 .
one of these is the blagn _ 2 - 251.0 _ already discussed , which shows broad mgii ( @xmath222800 ) emission and absorption in its spectrum .
almost all of the remaining 12 show emission lines in their spectra .
nine out of 11 show oii(@xmath223727 ) when in range , with several sources also displaying h@xmath23 and oiii(@xmath225007 ) .
the x - ray and strong radio frii galaxy _ 4 - 752.1 _ , displays only strong absorption lines in its spectrum .
only one source , _ 4 - 254.0 _ , shows weak nev(@xmath223426 ) emission , a line indicative of the presence of an agn ( hall @xmath7 2000 ) .
neiii(@xmath223869 ) , also stronger in agn than in starforming galaxies ( rola @xmath7 1997 ) , is seen weakly in 2 sources , _ 3 - 143.0 _ and _ 3 - 386.111111_. these galaxies also show oiii and h@xmath23 , but with flux ratios consistent with star formation rather than agn activity ( veilleux & osterbrock 1997 ) . only the spiral galaxy _ 4 - 976.1 _ is low enough redshift to reveal strong h@xmath17 , sii(@xmath226713 + 6731 ) and oi(@xmath226300 ) emission in the optical spectrum
. the line flux ratios of oiii / h@xmath23 , sii / h@xmath17 and oi / h@xmath17 all indicate that this source is near the border that divides starforming galaxies and agn ( veilleux & osterbrock 1997 ) .
therefore , based on the optical spectra alone , objects _ 2 - 251.0 _ ( blagn ) , possibly _ 4 - 976.1 _ ( liner / seyfert 2 ) , and possibly _ 4 - 254.0 _ ( through the weak presence of nev ) show evidence of agn .
knowledge of the space density of faint agn at higher redshifts is critical to our understanding of agn evolution and the agn phenomenon in general .
our sample of 16 variable sources spans a redshift range of z=0.1 to 1.8 with most sources lying between z=0.4 and 1.1 .
if we make the assumption that these variables are indeed agn , it is interesting to compare their number density with local samples of agn .
we must first estimate the luminosity of the agn candidates in our survey .
due to the faint magnitudes and small angular sizes of many of the variables , 2-dimensional modeling to separate the components ( disk+bulge+agn ) of the galaxies is difficult and would not yield consistent results for all of the selected galaxies .
a simpler approach is to estimate the agn luminosities by assuming that the entire nuclear flux is attributed to the agn .
this value is an overestimate in one sense since the nuclear light will also include light from the underlying galaxy , though the small nuclear aperture minimizes the contribution from the host .
based on the expected level of variability over this time interval predicted from agn structure functions , we estimate the variable component for our agn candidates to be no less than 15% of the total nuclear light . in the most extreme case , the true agn magnitude could be @xmath32 magnitudes fainter than that measured within the nuclear aperture . in another sense
however , we underestimate the agn flux since we have not applied aperture corrections to include light from the wings of the psf . our aperture of [email protected]@xmath11 encircles about 50% of the flux from a point source ( holtzman @xmath7 1995 ) .
if the point source makes up the entire flux within the aperture , an aperture correction would increase the magnitude by up to @xmath30.8 mag .
if the point source is some fraction of the total nuclear light , the aperture correction will be smaller .
determining the true aperture correction is not possible without knowing the relative fluxes of the point source and host within the fixed aperture .
figure 8 is the absolute magnitude based on the nuclear flux _ vs _ the source redshift ( h@xmath24=75 km / s / mpc ; q@xmath24=0.5 ) .
the absolute b magnitudes are calculated assuming a power - law index of @xmath17=1.0 resulting in a k - correction of zero and v@xmath25b@xmath26=0.32 .
we find that most sources lie between -15@xmath15m@xmath6@xmath15 - 17 in the redshift range 0.4@[email protected] .
we conduct the luminosity - volume test , or v / v@xmath27 test , of schmidt ( 1968 ) to determine if significant evolution is present in our sample of variable nuclei . for this calculation
, we determine for each source in our survey the maximum redshift at which it could exist and still be included in our survey .
the volume of space enclosed by a sphere with a radius equal to this z@xmath27 is the value v@xmath27 .
the volume enclosed at the true redshift of the source is @xmath28 .
spectroscopic redshifts have been determined for 14 of our objects .
photometric redshifts are estimated for one other ( fly ) .
there is no spectroscopic or photometric redshift published for _ 2 - 456.22_. based on the v
i color for this object , we estimate its photometric redshift to be z=1.8 which corresponds to those of other hdf sources with similar colors . to define z@xmath27 , we determine the limiting redshift for each source to be detected in our survey .
the brightest limiting magnitude in our survey is defined by the 3@xmath9 line in figure 4a , which indicates the faintest magnitude at which a source varying by a given @xmath14 m would be considered a significant variable . below this magnitude
, a source varying by @xmath14 m would not be detected as variable .
therefore , z@xmath27 is the redshift where the apparent magnitude of a nucleus varying by @xmath14 m falls below the 3@xmath9 variability significance limit shown in figure 4a .
the mean v / v@xmath27 for the 16 nuclei in our survey is [email protected] . for a population of objects uniformly distributed in space , @xmath4v / v@xmath27@xmath16=0.5 .
our value of [email protected] indicates that sources are found more often at larger distances than at nearby distances , indicating some evolution in the population of galaxies with variable nuclei .
our survey incompleteness , as discussed in section 3 , affects our ability to detect sources at higher redshifts and fainter magnitudes .
thus , the value of @xmath4v / v@xmath27@xmath16 would actually be higher if corrected for this incompleteness .
an increase in the number density would be consistent with the picture of agn / qso evolution at higher luminosities , where the population increases rapidly out to z@xmath03 .
even though evolution may be present among our sample , the small number of sources spread over a range in redshift makes it difficult to construct the luminosity function in more than one redshift bin . to accurately compare the density of our sources to local samples
, we construct the lf in one redshift bin at z=0.41.1 .
we choose these limits to minimize the redshift range of our sources while maximizing the number of sources in the bin .
twelve of our 16 sources lie within this redshift range , which excludes the 2 lowest and highest redshift galaxies .
the mean redshift for these sources is @xmath4z@xmath16=0.69 .
we calculate the luminosity function using a technique similar in nature to the v / v@xmath27 test described above .
the lf is created by summing 1/v@xmath27 for sources in discrete magnitude bins . in this case , we are actually determining v@xmath29 , or the _ accessible _ volume in which each source could be detected .
this volume is limited not only by the magnitude limit for detection , but also by the lower and upper redshift limits of the shell in which we are summing .
figure 9 is the lf for hdf variable nuclei in the redshift range z=0.41.1 .
incompleteness due to the fact that our variability selection is based on only 2 epochs is expected to produce an underestimate in the number of agn by a factor of @xmath31.3 , or a decrease in log@xmath30 of @xmath30.12 . for comparison ,
we plot the local seyfert lfs of huchra & burg ( 1992 ; hb92 ) from the cfa survey and ulvestad & ho ( 2001 ; uh01 ) from the palomar survey .
both local surveys consist of spectroscopically selected seyfert 1s and 2s .
our lf for variable sources at 0.4@[email protected] extends to fainter absolute magnitudes than the local lfs , largely due to the fact that the local lfs are calculated for the total galaxy light ( agn+host ) rather than the magnitude of the agn alone .
nonetheless , many of the agn candidates detected in our survey may be truly fainter than those of hb92 .
based on the spectroscopic selection criterion employed by hb92 , most of our sources would not be detected as agn ( see section 6 ) in the hb92 survey . therefore , we assume that the agn component in our galaxies do not comprise as much of the total galaxy light as the hb92 sources and are likely to be intrinsically fainter .
the uh01 lf covers the same integrated galaxy magnitude range as hb92 but shows an overall higher density for local seyferts .
ulvestad & ho explain that this is due to the fact that their lf includes intrinsically fainter seyfert nuclei than does the hb92 sample . for the seyfert 1 galaxies , ho & peng ( 2001 ) decompose the nuclei from the host galaxies in both local samples and confirm this assertion , finding a median nuclear magnitude of m@xmath6=-14.6 for uh01 versus -17.4 for the hb92 seyfert 1 galaxies . from their study
, we also find that the nuclear magnitudes of the local seyferts in both samples can be extremely faint when compared to the total galaxy magnitude , in some cases up to 10 magnitudes fainter .
therefore , an lf consisting of nuclear magnitudes for local seyferts would be a much better comparison to our hdf agn candidates .
since this is not currently available , an alternative is to produce the lf of our sources using the total integrated galaxy magnitude rather than the nuclear magnitude alone ( dashed lf ) .
the integrated galaxy magnitudes for our variable sources cover the same range as the faint end of the local seyfert lfs but have a number density 610@xmath10 greater than the local seyfert density of uh01 .
if the agn in these two samples cover a similar magnitude range , this is evidence for a significant increase in number density from z=0 to @xmath30.7 .
the magnitudes for the uh01 agn determined in ho & peng ( 2001 ) for the seyfert 1 population cover the range -9@xmath15m@xmath6@xmath15 - 22 mag with most lying within -12@xmath15m@xmath6@xmath15 - 20 .
our nuclear aperture lf ( open circles in figure 9 ) extends from -14.5@xmath15m@xmath6@xmath15 - 19 . in the most extreme case , based on the over and
underestimates of the true agn magnitudes discussed previously , the variable agn in our survey could cover the magnitude range -12.5@xmath15m@xmath6@xmath15 - 20 .
this magnitude range is consistent with that estimated for the uh01 sample .
it is therefore unlikely that the increase in number density between the uh01 lf and our lf is due to the inclusion of instrinsically fainter agn as appears to be the case for the difference between the uh01 and hb92 lfs .
finally , we note that the lf for the variable nuclei continues to rise at faint magnitudes , indicating that number counts have not begun to turn over at magnitudes even as faint as m@xmath6@xmath3 - 15 .
the effects of incompleteness in our survey due to agn light dilution would cause us to be less complete at the faint end and corrections in this sense would likely increase rather than decrease our faintest lf bin .
most local agn lfs , such as those shown in figure 9 , show a flattening toward fainter magnitudes .
the local lf of londish @xmath7 ( 2002 ) for type 1 agn reveals a much flatter slope at the faint end as compared to the bright end and can be well fit with a two power - law function .
our increasing numbers at the faint end may be an evolutionary effect or might be attributed to the inclusion of a greater fraction of obscured agn .
more data are needed as well as better classification into agn types to improve the statistics and verify this trend .
we have investigated variability within the hubble deep field north with the aim of identifying low - luminosity agn in the nuclei of field galaxies .
we compare photometry of the nuclei of galaxies in the initial v - band image obtained in 1995 and a second image obtained in 2000 and find evidence for significant variability in 16 galaxies . correcting for possible spurious detections and incompleteness results in a total of @xmath318 sources which make up 8.3% of galaxy nuclei to v@xmath1=27.5 .
the galaxies hosting variable nuclei range in redshift from 0.09@[email protected] with most lying between 0.4@[email protected] .
nuclear colors of the variable nuclei are consistent with those of non - variable nuclei .
seven of our variable sources are coincident ( within @xmath31.2@xmath11 ) with x - ray sources detected in the 2ms chandra exposure ( alexander @xmath7 2003 ) .
we find that 39% of the 2ms chandra detections in the original hdf are significant variables .
the optical variability observed in these x - ray sources provides important evidence to confirm the presence of an agn and could potentially help discriminate between low - luminosity x - ray emitting agn and other x - ray emitting phenomena such as supernova remnants or x - ray binaries .
in addition , 44% of our sources are associated with mid - ir detections at 15@xmath5 and 31% are detected at 1.4ghz .
optical spectra are currently available for 13 of the 16 variable sources .
2 - 251.0 _ is a broad - line agn .
two others show some evidence of harboring an agn . object _ 4 - 976.1
_ , a low redshift spiral , has emission line ratios indicative of a liner / seyfert 2 and _ 4 - 254.0 _ reveals weak nev emission . based on the v / v@xmath27 test , there is evidence for evolution within our sample of variable nuclei in the sense that more variables are found at higher redshifts .
we estimate the luminosity function of variable agn in the redshift range 0.4@[email protected] .
this lf is a lower - limit since our variability selection technique , employing only 2 epochs for detection , could miss @xmath32025% of variable nuclei .
the nuclear magnitudes for our variables extend to m@xmath6@xmath0 - 15.0 with a median magnitude of m@xmath6=-16.3 . computing our lf with total integrated galaxy magnitudes ,
we compare to the lf of ulvestad & ho ( 2001 ) for local seyfert galaxies . if our agn cover a similar magnitude range as those included in their sample , we find evidence for an increase in the number density of faint agn by a factor of 610 from z@xmath30 to 0.7 .
the lf for our selected variable nuclei continues to rise at faint magnitudes .
unlike the lfs of most qsos and local seyferts , we do not see signs of flattening at the faint end .
the results of this paper can be tested and confirmed with a larger sample of variable nuclei detected at these redshifts .
a similar variability survey is being conducted for the groth - westphal survey strip ( sarajedini @xmath7 2003 ) over 7 years and will produce a larger sample of variables to improve the statistical significance of any evolutionary trends .
additional future work includes the analysis of the currently ongoing goods hst treasury program ( dickinson @xmath7 2003 ) , an acs survey of the sky around and including the hdf . the 5-epoch survey , at intervals of @xmath345 days , will allow for the investigation of short - term variability as well as providing confirmation and improved completeness for the variables detected in this study .
support for proposal go-8389.02 - 97a for vs and go-8389.01 - 97a and ar-7984.01 - 96a for rg was provided by nasa through a grant from the space telescope science institute , which is operated by the association of universities for research in astronomy , inc .
, under nasa contract nas5 - 26555 .
we gratefully thank rodrigo ibata who was the pi for obtaining the second epoch v - band image of the hdf . a special thanks to amy barger and judy cohen for supplying the keck spectra in digital form to be analyzed for these sources .
thanks also to david koo and the referee for helpful comments and suggestions for improving this paper .
2 - 860.0 & 54.102 & 13:54.35 & 0.851 & 3 & 0.00 ( 1.25 ) & 27.293 & 0.107 & 3.53 & & & hdf - pm3 - 34 & 2.650 & & + 2 - 456.22 & 50.027 & 13:51.99 & 1.8 & & 0.52 ( 1.08 ) & 26.587 & -0.056 & 3.23 & & & & & & + 2 - 251.0 & 46.344 & 14:04.62 & 0.960 & 2 & 1.00 ( 1.76 ) & 24.134 & 0.097 & 6.48 & 171 & 0.194 & hdf - pm3 - 20 & 1.439 & 192 & 0.048 + 3 - 355.0 & 56.923 & 13:01.56 & 0.474 & 2 & 0.94 ( 1.00 ) & 26.388 & -0.048 & 3.20 & 203 & 0.066 & hdf - ps2 - 4 & 2.116 & 225 & 0.604 + 3 - 863.0 & 58.649 & 12:21.72 & 0.682 & 4 & 0.00 ( 1.14 ) & 26.411 & 0.046 & 3.06 & & & & & & + 3 - 943.0 & 56.432 & 12:09.31 & 0.321 & 3 & 0.00 ( 1.23 ) & 26.972 & 0.095 & 4.04 & & & & & 224 & 2.074 + 3 - 143.0 & 49.644 & 12:57.43 & 0.477 & 2 & 0.00 ( 1.58 ) & 26.148 & 0.070 & 4.65 & & & & & & + 3 - 659.1 & 51.722 & 12:20.18 & 0.401 & 2 & 0.00 ( 1.50 ) & 25.390 & 0.060 & 4.02 & 190 & 1.235 & hdf - pm3 - 29 & 1.501 & 209 & 1.155 + 3 - 386.111 & 50.254 & 12:39.72 & 0.474 & 3 & 0.00 ( 6.23 ) & 25.397 & 0.057 & 3.77 & & & hdf - ps3 - 16 & 0.823 & & + 3 - 777.1 & 52.022 & 12:09.63 & 0.458 & 4 & 0.53 ( 1.14 ) & 26.24 & 0.048 & 3.21 & & & & & & + 3 - 965.111111 & 57.485 & 12:10.55 & 0.663 & 1 & & 25.293 & -0.141 & 9.42 & 206 & 0.456 & & & & + 4 - 254.0 & 46.127 & 12:46.50 & 0.901 & 1 & 0.90 ( 1.03 ) & 26.216 & -0.049 & 3.28 & & & hdf - ps3 - 17 & 2.182 & & + 4 - 516.0 & 45.652 & 11:53.97 & 1.24 & 3 & 0.66 ( 1.08 ) & 26.827 & -0.074 & 3.54 & & & & & & + 4 - 752.1 & 44.377 & 11:33.20 & 1.011 & 1 & 0.87 ( 1.25 ) & 26.250 & -0.046 & 3.07 & 165 & 0.210 & & & 187 & 0.121 + 4 - 801.0 & 39.990 & 12:33.65 & 0.920 & 4 & 0.17 ( 1.07 ) & 27.234 & 0.096 & 3.33 & & & & & & + 4 - 976.1 & 41.643 & 11:31.85 & 0.089 & 4 & 0.00 ( 19.1 ) & 24.998 & 0.054 & 3.60 & 160 & 1.112 & hdf - pm3 - 12 & 3.308 & & + | we present the results of a 2-epoch variability survey in the hubble deep field with the goal of investigating the population of agn to z@xmath01 .
the high resolution and stability of hst allows accurate photometry to be obtained within subarcsecond apertures , resulting in the ability to probe much lower agn / host galaxy luminosity ratios than can be done from the ground .
the primary data sets analyzed for galactic variability are the original hdf observations obtained in december 1995 and a second v - band ( f606w ) image obtained almost exactly 5 years later in 2000 .
we find evidence for nuclear variability in 16 of 217 galaxies brighter than v@[email protected] .
correcting for incompleteness and spurious detections , variable nuclei make up @xmath38% of the surveyed galaxies .
these sources have a redshift range of 0.09@[email protected] and cover the full range of galaxy nuclear v
i colors .
seven of our variable sources are coincident with x - ray sources detected in the 2ms chandra survey ; six from the main catalog and one from the supplementary catalog .
we find that 44% of the variable nuclei are associated with mid - ir detections at 15@xmath5 and 31% are detected at 1.4ghz .
optical spectra are available for 13 of the 16 variables .
one is a broad - line agn and 2 others show weak evidence of type 2 agns . with the assumption that these variables are all active nuclei
, we estimate the agn lf at 0.4@[email protected] extending to m@xmath6@xmath0 - 15 .
we find evidence for an increase in the number density of faint agn when comparing to the local seyfert luminosity function .
the lf for optically varying nuclei appears to rise in number density with no evidence of turning over at these faint magnitudes . |
the temperature anisotropy in the cosmic microwave background is well established as a powerful constraint on theories of the early universe . a related observable
, the polarization anisotropy of the cmb , gives us a new window into the physical conditions of that era . at large angular scales
the polarization has the potential to be a direct probe of the universe at an age of @xmath26 s as well as to inform us about the ionization history of the universe .
this paper reports on the direct detection of cmb polarization at large angular scales and helps set a foundation for future observations .
it is one of four related papers on the three - year _ wmap _ analysis : @xcite report on systematic errors and mapmaking , @xcite on the temperature anisotropy and basic results , and @xcite on the parameter estimation and cosmological significance .
the polarization of the cmb was predicted soon after the discovery of the cmb @xcite . since then
, considerable advances have been made on both theoretical and observational fronts .
the theoretical development @xcite has evolved to where there are precise predictions and a common language to describe the polarization signal .
@xcite give a pedagogical overview .
the first limits on the polarization were placed by @xcite , followed by @xcite and @xcite . in 2002 ,
the dasi team announced a detection of cmb polarization at sub - degree angular scales based on 9 months of data from a 13 element 30 ghz interferometer @xcite .
the signal level was consistent with that expected from measurements of the temperature spectrum .
the dasi results were confirmed and extended @xcite almost contemporaneously with the release of the cbi @xcite and capmap @xcite results .
more recently , the boomerang team has released its measurement of cmb polarization @xcite .
all of these measurements were made at small angular scales ( @xmath27 ) .
of the experiments that measure the polarization , the dasi , cbi , and boomerang @xcite teams also report detections of the temperature - polarization cross correlation .
the cmb polarization probes the evolution of the decoupling and reionization epochs .
the polarization signal is generated by thompson scattering of a local quadrupolar radiation pattern by free electrons .
the scattering of the same quadrupolar pattern in a direction perpendicular to the line of sight to the observer has the effect of isotropizing the quadrupolar radiation field .
the net polarization results from a competition between these two effects .
we estimate the magnitude of the signal following basko and polnarev ( 1980 ) . by integrating the boltzmann equation for the photon distribution
they show that the ratio of the polarization anisotropy ( @xmath28 ) to the temperature ( @xmath29 ) signal in a flat cosmology is given by @xmath30\sqrt{1+z'}dz'\over \int_0^\infty [ 6e^{-\tau(z ' ) } + e^{-0.3\tau(z')}]\sqrt{1+z'}dz ' } , \label{eq : ih}\ ] ] where @xmath31 is the optical depth . here , @xmath32 is the thompson cross section , @xmath33 is the speed of light , and @xmath34 is the free electron density .
the difference in brackets in the numerator sets the range in @xmath35 over which polarization is generated .
for example , if the decoupling epoch entailed an instantaneous transition from an extremely high optical depth ( @xmath36 ) to transparency ( @xmath37 ) , there would be no polarization signal . to estimate the polarization fraction we compute the optical depth using ordinary atomic physics and the thermal history of the universe @xcite .
the result is shown in figure 1 . from inserting @xmath38 in equation [ eq : ih ]
, we find that the expected level of polarization anisotropy is @xmath39% ( in @xmath40 ) of the anisotropy .
the polarization producing quadrupole is generated by different mechanisms at different epochs .
near decoupling at @xmath41 @xcite , velocity gradients in the flow of the primordial plasma give rise to the quadrupole .
more specifically , in the rest frame of an electron in such a flow , the radiation background has a quadrupolar pattern proportional to the velocity gradient , @xmath42 , and the mean free path between scatterings , @xmath43 . just before decoupling , @xmath44 , the photons are tightly coupled to the electrons and @xmath43 is small .
thus , the polarization is small . as decoupling proceeds @xmath43 increases and the quadrupole magnitude increases .
the process is cut off at lower redshift because the optical depth drops so rapidly . in the context of inflationary cosmology ,
@xcite show that in fourier space the polarization signal is @xmath45 where @xmath46 is the wavevector and @xmath47 is the width of the last scattering surface .
after decoupling there are no free electrons to scatter the cmb until the first generation of stars ignite and reionize the universe at @xmath48 .
the free electrons then scatter the intrinsic cmb quadrupole , @xmath49 , and produce a polarized signal @xmath50 . as
this process occurs well after decoupling , the effects of the scattering are manifest at comparatively lower values of @xmath51 .
we expect the maximum value of the signal to be at @xmath52 where @xmath53 is the current angular size of the horizon at reionization . for @xmath54
a simple fit gives @xmath55 , so that for @xmath56 , @xmath57 . thus , the signature of reionization in polarization is cleanly separable from the signature of decoupling .
in the first data release the _ wmap _ team published a measurement of the temperature - polarization ( te ) cross spectrum for @xmath58 @xcite with distinctive anti - peak and peak structure @xcite .
the @xmath59 part of the spectrum was consistent with the prediction from the temperature power spectrum , while the @xmath11 part showed an excess that was interpreted as reionization at @xmath60 ( 95% cl ) .
this paper builds on and extends these results .
not only are there three times as much data , but the analysis has improved significantly : 1 ) the polarization mapmaking pipeline now self - consistently includes almost all known effects and correlations due to instrumental systematics , gain and offset drifts , unequal weighting , and masking @xcite . for example , the noise matrix is no longer taken to be diagonal in pixel space , leading to new estimates of the uncertainties .
2 ) the polarization power spectrum estimate now consistently includes the temperature , e and b modes ( defined below ) , and the coupling between them ( see also * ? ? ?
3 ) the polarized foreground emission is now modeled and subtracted in pixel space ( [ sec : fgr ] ) .
potential residual contamination is examined @xmath51 by @xmath51 as a function of frequency .
in addition to enabling the production of full sky maps of the polarization and their power spectra , the combination of these three improvements has led to a new measure of the @xmath11 te and ee spectra , and therefore a new evaluation of the optical depth based primarily on ee .
the rest of the paper is organized as follows : we discuss the measurement in [ sec : inst ] and consider systematic errors and maps in [ sec : syst ] . in [ sec : foregrounds ] we discuss foreground emission .
we then consider , in [ sec : pspec ] and [ sec : cosmo ] , the polarization power spectra and their cosmological implications .
we conclude in [ sec : conclude ] .
_ wmap _ measures the difference in intensity between two beams separated by @xmath61 in five frequency bands centered on 23 , 33 , 41 , 61 , and 94 ghz @xcite .
these are called k , ka , q , v , and w bands respectively .
corrugated feeds @xcite couple radiation from back - to - back telescopes to the differential radiometers .
each feed supports two orthogonal polarizations aligned so that the unit vectors along the direction of maximum electric field for an a - side feed follow @xmath62 in spacecraft coordinates @xcite . for a b - side feed ,
the directions are @xmath63 .
the @xmath64 axis points toward the sun along the spacecraft spin axis ; the @xmath65 plane bisects the telescopes and is perpendicular to the radiator panels ( * ? ? ?
* figure 2 ) ( * ? ? ?
* figure 1 ) .
the angle between the spacecraft spin axis and the optical axes is @xmath66 .
thus the two polarization axes on one side are oriented roughly @xmath67 with respect to the spin axis .
the polarization maps are derived from the difference of two differential measurements @xcite .
one half of one differencing assembly ( da ) @xcite measures the difference between two similarly oriented polarizations , @xmath68 , from one feed on the a side and one feed on the b side ( e.g. , w41 : polarization 1 of the 4th w - band da corresponding to @xmath69 in both expressions above ) .
the other half of the da measures the difference between the other polarizations in the same pair of feeds , @xmath70 ( e.g. , w42 : polarization 2 of the 4th w - band da corresponding to @xmath71 in both the expression above ) . the polarization signal is proportional to @xmath72 . in other words ,
_ wmap _ measures a double difference in polarized intensity , not the intensity of the difference of electric fields as with interferometers and correlation receivers ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . with these conventions , the total intensity and polarization signals as measured at the output of the detectors
are ( * ? ? ?
3&4 ) : @xmath73 where @xmath74 and @xmath75 are the unit vectors for the a and b sides ; @xmath76 , @xmath77 , and @xmath78 are the stokes parameters stokes parameter . ] , and @xmath79 is the angle between the polarization direction of the electric field and the galactic meridian @xcite . in the mapmaking algorithm @xcite , @xmath76 , @xmath77 , and @xmath78
maps of the sky are produced from the time - ordered differential measurements , @xmath80 and @xmath81 . from these , we form maps of polarization intensity , @xmath82 , and direction , @xmath83 .
this convention has @xmath79 positive for north through west and follows the convention in @xcite and healpix @xcite .
however , it differs from the standard astronomical position angle ( pa ) which has @xmath84 with @xmath85 positive for north through east .
the choice of convention does not affect the plots . for linear polarization in a given pixel ,
the @xmath77 and @xmath78 quantities are related to the @xmath86 and @xmath87 components of the electric field , @xmath88 , through the coherency matrix @xcite : @xmath89 where we have set stokes @xmath90 .
the polarized component of the coherency matrix is a spin - two field on a sphere ; the total power is the trace of the coherency matrix .
the crab nebula [ tau a , 3c144 , ra = @xmath91 , dec=@xmath92 ( j2000 ) ] is the brightest polarized point source in the sky and provides a useful end - to - end check of the sign conventions and mapmaking pipeline .
figure [ fig : crab ] shows our measurement of the crab in q band ( 41 ghz ) in @xmath76 , @xmath77 , @xmath78 , @xmath93 , and @xmath79 .
note that its polarization direction ( @xmath94negative ) , is perpendicular to the polarization of the galaxy ( @xmath94positive ) .
the _ wmap _ polarization direction and intensity are in general agreement with previous measurements .
table [ tab : crab ] summarizes the results in all five frequency bands and previous measurements in our frequency range .
a second check is needed to fully resolve the sign convention because with @xmath95 , @xmath96 . in figure
[ fig : crab ] we show that the polarization direction of the centaurus a galaxy [ cen a , ngc5128 , ra=@xmath97 , dec@xmath98 ( j2000 ) ] is consistent with that measured by @xcite .
figures [ fig : pgmaps1 ] and [ fig : pgmaps2 ] show the @xmath93 and @xmath79 maps of the full sky for all five frequency bands in galactic coordinates .
figure [ fig : kstereo ] shows a lambert equal area projection of the galactic polar region in k band .
a number of features are immediately apparent to the eye .
k band is strongly polarized over a large fraction of the sky , including the polar region .
the north polar spur and its southern extension are clearly evident .
the polarization has a coherent structure over large swaths of sky which translates into significant emission at low @xmath51 .
the polarization intensity decreases with increasing frequency but follows the same pattern .
k band is a good monitor of polarized foreground emission as discussed below . though not immediately apparent to the eye , there is somewhat more polarized emission at w band than v band .
the uneven weighting due to the scan strategy is also evident as increased noise in the ecliptic plane ( * ? ? ?
* figure 4 ) . figure [ fig : qumaps ] shows the k and ka bands in stokes @xmath77 and @xmath78 . while foreground emission is visible with a high signal to noise ratio , the cmb polarization anisotropy is not , a situation unlike that for the temperature anisotropy .
detection of the cmb polarization requires tight control of systematic errors , as small couplings to the temperature field or instrument will dominate the polarization signal .
_ wmap _ s differential nature and interlocked scan strategy suppress potential polarization systematics in ways similar to the suppression for temperature systematics .
the details are different however , and more complex because of the tensorial nature of the polarization field and the double difference required to measure the polarization . throughout our analyses , the overall level of systematic contamination
is assessed with null tests as described here and in @xcite & @xcite .
the mapmaking procedure is described in @xcite .
end - to - end simulations of the instrument and scan strategy , incorporating realistic models of the frequency response , foreground emission , and detector noise characteristics , are used to assess the possible levels of contamination .
interactions between the slow @xmath99 % drifts in the gain , non - uniform weighting across the sky , the 0.2% correlation due to the oppositely directed beams , the time series masking of the planets , and the @xmath100 noise are accounted for in the map solution . in the following we discuss how the instrumental offset , gain / calibration uncertainty , passband mismatch , main beam mismatch , polarization isolation and cross polarization , loss imbalance , and sidelobes affect the polarization maps . _
offset and baseline drift- _ the instrumental offset is the output of the detector in the absence of celestial signal .
the average polarization offset in the q , v , and w bands is 250 mk .
changes in this offset on time scales of minutes to hours arise from spacecraft temperature changes and from @xmath100 drifts in the amplifier gain acting on the 250 mk . to measure polarization at the level of @xmath101k
, we require that changes in the baseline be suppressed by roughly a factor of @xmath102 . the first step in achieving
this is maintaining a stable instrument and environment .
the physical temperature of the das averaged over a spin period changes by less than 5 parts in @xmath102 @xcite , suppressing changes in the baseline by a similar factor .
the second step in achieving this is through the baseline removal in the mapmaking algorithm @xcite .
if the precession of the satellite were stopped , the temperature data for @xmath103 would repeat in the time stream at the spin period ( 2.16 m ) .
the offset , though , would change sign relative to the celestial signal at half the spin period enabling the differentiation of celestial and instrumental signals .
( alternatively , one may imagine observing a planet in which case the temperature data would change sign at half the spin period and an offset would be constant . )
by contrast , with our choice of polarization orientations , the polarization data @xmath104 , would repeat at half the spin period for some orientations of the satellite .
consequently , an instrumental offset would not change sign relative to a celestial signal upon a @xmath105 spacecraft rotation .
thus the polarization data are more sensitive to instrumental offsets than are the temperature data . in general
, the polarization data enters the time stream in a more complex manner than does the temperature data . _
calibration _ an incorrect calibration between channels leads to a leakage of the temperature signal into @xmath104 , contaminating the polarization map .
calibration drifts cause a leakage that varies across the sky .
@xcite show that calibration drifts on @xmath106 day time scales are the result of sub - kelvin changes in the amplifier s physical temperature .
the calibration can be faithfully modeled by fitting to the physical temperature of each da with a three parameter model . here again _ wmap _
s stability plays a key role .
the residual calibration errors are at the @xmath107% level .
these errors do not limit the polarization maps because the bright galactic plane is masked in the time ordered data when producing the high galactic latitude maps @xcite .
the overall absolute calibration uncertainty is still the first - year value , 0.5% @xcite .
_ passband mismatch _ the effective central frequencies @xcite for @xmath68 and @xmath70 are not the same .
this affects both the beam patterns , treated below , and the detected flux from a celestial source , treated in the following .
the passbands for the a and b sides of one polarization channel in a da may be treated as the same because the dominant contributions to the passband definition , the amplifiers and band defining filters , are common to both sides . since _ wmap
_ is calibrated on the cmb dipole , the presence of a passband mismatch means that the response to radiation with a non - thermal spectrum is different from the response to radiation with a cmb spectrum @xcite .
this would be true even if the sky were unpolarized , the polarization offset zero , and the beams identical .
the effect produces a response in the polarization data of the form : @xmath108 where @xmath109 is the unpolarized temperature difference observed in radiometer one , and similarly for @xmath110 .
if these differ due to passband differences , the polarization data will have an output component that is independent of parallactic angle .
given sufficient paralactic coverage , such a term can be separated from stokes @xmath77 and @xmath78 in the mapmaking process .
we model the polarized signal as @xmath111 where the constant , @xmath112 , absorbs the signal due to passband mismatch .
we solve for the mismatch term simultaneously with @xmath77 and @xmath78 as outlined in @xcite .
note that we do not need to know the magnitude of the passband mismatch , it is fit for in the mapmaking process .
the @xmath112 map resembles a temperature map of the galaxy but at a reduced amplitude of @xmath113 in k band , 2.5% in the v1 band , and on average @xmath106% for the other bands .
the maps of @xmath112 agree with the expectations based on the measured passband mismatch . _ beamwidth mismatch _ the beamwidths of each polarization on each of the a and b sides are different .
the difference between the a and b side beam shapes is due to the difference in shapes of the primary mirrors and is self consistently treated in the window function @xcite .
the difference in beam shapes between @xmath68 and @xmath70 is due to the mismatch in central frequencies . and
@xmath70 would be the same to @xmath114% accuracy . ]
this effect is most easily seen in the k - band observations of jupiter .
we denote the brightness temperature and solid angle of jupiter with @xmath115 and @xmath116 , and the measured quantities as @xmath117 and @xmath118 .
although the product @xmath119 is the same for the two polarizations ( because jupiter is almost a thermal source in k band ) , the beam solid angles differ by 8.1% on the a - side and 6.5% on the b - side @xcite .
the primary effect of the beamwidth mismatch is to complicate the determination of the intrinsic polarization of point sources .
the difference in beams also leads to a small difference in window functions between @xmath68 and @xmath70 .
the signature would be leakage of power from the temperature anisotropy into the polarization signal at high @xmath51 .
we have analyzed the data for evidence of this effect and found it to be negligible .
additionally , as most of the cmb and foreground polarization signal comes from angular scales much larger than the beam , the difference in window functions can safely be ignored in this data set
. _ polarization isolation and cross polarization _ polarization isolation , @xmath120 , and cross polarization are measures of the leakage of electric field from one polarization into the measurement of the orthogonal polarization .
for example , if a source were fully polarized in the vertical direction with intensity @xmath121 and was measured to have intensity @xmath122 with a horizontally polarized detector , one would say that the cross polar response ( or isolation ) is @xmath123 or @xmath124 db .
the term `` polarization isolation '' is usually applied to devices whereas `` cross polarization '' is applied to the optical response of the telescope .
we treat these together as a cross - polar response . for _ wmap
_ , the off - axis design and imperfections in the orthomode transducers ( omt ) lead to a small cross - polar response .
the ratio of the maximum of the modeled crosspolar beam to the maximum of the modeled copolar beam is @xmath125 , @xmath126 , @xmath127 , @xmath127 , & @xmath128 db in k through w bands respectively .
the determination of the feed and omt polarization isolation is limited by component measurement .
the maximum values we find are : @xmath129 @xmath130 , @xmath127 , @xmath127 , @xmath126 , & @xmath125 db for k through w bands respectively @xcite .
we consider the combination of beams plus components below . because _ wmap _ measures only the difference in power from two polarizations , it measures only stokes @xmath77 in a reference frame fixed to the radiometers , @xmath131 .
the sensitivity to celestial stokes @xmath77 and @xmath78 comes through multiple observations of a single pixel with different orientations of the satellite .
the formalism that describes how cross polarization interacts with the observations is given in appendix [ app : jones ] . to leading order
, the effect of a simple cross polarization of the form @xmath132 is to rotate some of the radiometer @xmath78 into a @xmath77 component .
the measured quantity becomes : @xmath133 where @xmath134 and @xmath135 are the stokes @xmath77 components for the a and b sides in the radiometer frame , similarly with @xmath136 and @xmath137 .
note that in the frame of the radiometers @xmath135 ( stokes @xmath77 in the b - side coordinate system ) is @xmath138 .
this leads to the difference in sign conventions between the above and equation [ eq : diff_defs ] .
system measurements limit the magnitude of @xmath139 but do not directly give the phase , @xmath140 .
laboratory measurements of selected omts show @xmath141 , indicating the effective cross polar contamination is negligible .
we limit the net effect of the reflectors and omt with measurements in the gemac antenna range @xcite .
we find that for a linearly polarized input , the ratios of the maximum to minimum responses of the omts are 1 ) @xmath125 , @xmath126 , @xmath125 , @xmath125 , @xmath142 db for k through w band respectively ; 2 ) @xmath143 apart ; and 3 ) within @xmath144 of the design orientation .
thus , we can limit any rotation of one component into another to @xmath145 .
the comparison of @xmath79 derived from tau a to the measurement by @xcite in table [ tab : crab ] gives further evidence that any possible rotation of the stokes components is minimal .
based on these multiple checks , we treat the effects of optical cross polarization and incomplete polarization isolation as negligible . _
loss imbalance _ a certain amount of celestial radiation is lost to absorption by the optics and waveguide components .
if the losses were equal for each of the four radiometer inputs their effect would be indistinguishable from a change in the gain calibration .
however , small differences exist that produce a residual common - mode signal that is separable from the gain drifts @xcite .
the mean loss difference ( @xmath146 ) between the a- and b - sides is accounted for in the mapmaking algorithm @xcite .
in addition , the imbalance between the two polarizations on a single side , the `` loss imbalance imbalance , '' is also included @xcite .
it contributes a term @xmath147 to @xmath104 . here
@xmath148 is the sky temperature observed by the a , b side , and @xmath149 is the loss imbalance between the two polarizations on the @xmath150 side ( see appendix [ app : jones ] ) .
the magnitude of @xmath149 is @xmath151 @xcite .
a change in the loss across the bandpass due to , for example , the feed horns is a potential systematic error that we do not quantify with the radiometer passband measurements @xcite .
the magnitude of the effect is second order to the loss imbalance which is 1% .
we do not have a measurement of the effect .
nevertheless , as the effect mimics a passband mismatch , it is accounted for in the map solution . _
sidelobes _ when the sidelobes corresponding to @xmath104 are measured , there are two terms @xcite .
the largest term is due to the passband mismatch and is consistently treated in the mapmaking process .
the second smaller term is due to the intrinsic polarization .
we assess the contribution of both terms by simulating the effects of scan pattern of the sidelobes on the @xmath77 and @xmath78 polarization maps .
the results are reported in @xcite for the first - year polarization maps . in k band ,
the net rms contamination is @xmath152 outside of the kp0 mask region @xcite .
the intrinsic polarized sidelobe pickup is @xmath153 and is not accounted for in this three - year data release .
the contamination is more than an order of magnitude smaller in the other bands .
the microwave sky is polarized at all frequencies measured by _
wmap_. in k band the polarized flux exceeds the level of cmb polarization over the full sky .
by contrast , unpolarized foreground emission dominates over the cmb only over @xmath154 of the sky .
near 60 ghz and @xmath155 , the foreground emission temperature is roughly a factor of two larger than the cmb polarization signal .
thus , the foreground emission must be subtracted before a cosmological analysis is done .
while it is possible to make significant progress working with angular power spectra , we find that due to the correlations between foreground components , a pixel space subtraction is required .
table [ tab : gcent ] gives the foreground emission levels in a region around the galactic center .
the two dominant components of diffuse polarized foreground emission in the @xmath156ghz range are synchrotron emission and thermal dust emission @xcite .
free - free emission is unpolarized and spinning dust grains are expected to have polarization fractions of 1 - 2% @xcite .
the signal from polarized radio sources is negligible ( table 9 , * ? ? ?
the detected polarized sources are all well known , and among the brightest objects in the temperature source catalog .
they include 3c273 , 3c274 ( m87 , vir a ) , 3c279 , fornax a , pictor a , [ hb93]2255 - 282 , and [ hb93 ] 0637 - 752 and are masked as discussed below .
the potential impact of polarized foreground emission on the detection of the cmb polarization has been discussed by many authors including @xcite .
synchrotron emission is produced by cosmic - ray electrons orbiting in the @xmath157 g total galactic magnetic field .
the unpolarized synchrotron component has been well measured by _ wmap _ in the 23 to 94 ghz range @xcite .
the brightness temperature of the radiation is characterized by @xmath158 where the index @xmath159 varies considerably across the sky @xcite . in the microwave range ,
the spectrum reddens ( @xmath160 tends to more negative values ) as the frequency increases @xcite .
synchrotron radiation can be strongly polarized in the direction perpendicular to the galactic magnetic field @xcite .
the polarization has been measured at a number of frequencies [ from leiden between 408 mhz to 1.4 ghz @xcite , from parkes at 2.4 ghz @xcite , and by the medium galactic latitude survey at 1.4 ghz @xcite ] . at these low frequencies
, faraday rotation alters the polarization .
electrons in the galactic magnetic field rotate the plane of polarization because the constituent left and right circular polarizations propagate with different velocities in the medium . in the interstellar medium ,
the rotation is a function of electron density , @xmath34 , and the component of the galactic magnetic field along the line of sight , @xmath161 , @xmath162 where the integral is over the line of sight . with @xmath163 , @xmath164 , and @xmath165 ,
the net rotation is @xmath166 , with @xmath167 in ghz . at _ wmap
_ frequencies the rotation is negligible , though the extrapolation of low frequency polarization measurements to _ wmap _ frequencies can be problematic .
in addition there may be both observational and astrophysical depolarization effects that are different at lower frequencies @xcite .
thus , our method for subtracting the foreground emission is based , to the extent possible , on the polarization directions measured by _ wmap _ .
the other dominant component of polarized foreground emission comes from thermal dust .
nonspherical dust grains align their long axes perpendicularly to the galactic magnetic field through the davis - greenstein mechanism @xcite .
the aligned grains preferentially absorb the component of starlight polarized along their longest axis .
thus , when we observe starlight we see it polarized in the same direction as the magnetic field .
these same grains emit thermal radiation preferentially polarized along their longest axis , perpendicular to the galactic magnetic field .
thus we expect to observe thermal dust emission and synchrotron emission polarized in the same direction , while starlight is polarized perpendicularly to both . in section
[ sec : mag ] , we describe a physical model of the polarized microwave emission from our galaxy that explains the general features of the _ wmap _ polarization maps .
_ however this model is not directly used to define the polarization mask or to clean the polarization maps .
_ we go on to define the polarization masks in [ sec : mask ] and in [ sec : fgr ] we describe how we subtract the polarized foreground emission . in the following , we present a general model of polarized foreground emission based on _ wmap _ observations .
we view this as a starting point aimed at understanding the gross features of the _ wmap _ data . a more detailed model that includes the wide variety of external data sets that relate to polarization is beyond the scope of this paper . for both synchrotron and dust emission ,
the galactic magnetic field breaks the spatial isotropy thereby leading to polarization .
thus , to physically model the polarized foreground emission we need a model of the galactic magnetic field . as a first step
, we note that the k - band polarization maps suggest a large coherence scale for the galactic magnetic field , as shown in figure [ fig : pgmaps1 ] .
we can fit the large - scale field structure seen in the k - band maps with a gas of cosmic ray electrons interacting with a magnetic field that follows the spiral arms .
the galactic magnetic field can be quite complicated @xcite : there are field direction reversals in the galactic plane ; the field strength depends on length scale , appearing turbulent on scales @xmath168 kpc ( e.g. , * ? ? ?
* ) ; and the field strength of the large - scale field depends on the galactocentric radius ( e.g. , * ? ? ?
. nevertheless , most external spiral galaxies have magnetic fields that follow the spiral arm pattern ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
inspired by this , we model the field in cylindrical coordinates as : @xmath169\end{aligned}\ ] ] where @xmath170 , @xmath171 , @xmath20 and @xmath35 are measured in kpc with respect to the center of the galaxy , @xmath20 ranges from @xmath172 to @xmath173 , and the angles are in degrees .
the coordinates follow those in @xcite . for a fixed radius
, |*b*| has the same value at all azimuths .
we term the expression the logarithmic spiral arm ( lsa ) model to distinguish it from previous forms .
we take 8 kpc as the distance to the center of the galaxy @xcite .
the values are determined by fitting to the k - band field directions .
while the tilt , @xmath174 with @xmath175 , and the radial dependence , @xmath176 with @xmath177 , optimize the fit , the key parameter is @xmath178 , the opening angle of the spiral arms .
we find that the magnetic field is a loosely wound spiral with @xmath179 . to model the cosmic ray electrons , we assume they have a power - law distribution with slope in place of @xmath180 . ]
@xmath181 @xcite and are distributed in a exponential disk with a scale height of @xmath182 kpc and a radial scale length of @xmath183 kpc ( e.g. , * ? ? ?
* ) as @xmath184 while the amplitude of the signal is sensitive to the details of the cosmic ray distribution and the magnetic field structure , we may estimate its overall structure with the smooth field model ( eq .
[ eq : bfield ] ) and cosmic ray distribution .
the model predictions are not very sensitive to the assumed scale height and scale length .
we compute the polarization direction in this simple model as : @xmath185dx}\end{aligned}\ ] ] where @xmath186 is the line - of - sight direction , @xmath86 is the distance along that direction , @xmath34 is the electron distribution described above , and @xmath187 and @xmath188 are orthogonal components of the field perpendicular to the line of sight , with @xmath187 the component perpendicular to the @xmath35 axis of the galactic plane .
the parameters of the lsa model are determined by fitting the predicted directions , equation [ eq : tanqu ] , to the measured the k - band field directions .
figure [ fig : coherence ] shows the predicted magnetic field orientation for the lsa model .
the actual direction has a @xmath105 ambiguity . in the plane ,
the field lines are parallel to the galactic plane and the polarization projects into positive stokes @xmath77 . near the galactic pole , the field lines point along the spiral arm direction .
when projected into @xmath77 and @xmath78 , this leads to @xmath79 rotating around the pole .
we assess the agreement between the model field orientation and the orientation inferred from the k - band polarization with the correlation coefficient @xmath189 , and take the rms average over 74.3% of the sky ( outside the p06 mask described below ) . for our simple model
the agreement is clear : @xmath190 for k band .
using using rotation measures derived from pulsar observations in the galactic plane , one finds instead a spiral arm opening angle of @xmath191 as reviewed in @xcite .
our method is more sensitive to the fields above and below the plane ; and , unlike the case with pulsars , we have no depth information .
it has been suggested by r. beck and others that the north polar spur may drive our best fit value to @xmath179 . though the agreement between our simple model and the k - band polarization directions indicate that we understand the basic mechanism , more modeling is needed to connect the _ wmap _ observations to other measures of the magnetic field . for a power law distribution of electrons moving in a homogeneous magnetic field ,
the polarization fraction is @xmath192 @xcite . because the field direction changes as one integrates along the line of sight
, there is a geometric suppression of the amplitude of the polarization signal .
we estimate this geometric suppression as @xmath193 where all quantities are determined from the model : @xmath194 and @xmath76 is found by integrating the perpendicular component of the magnetic field , @xmath195 , and cosmic ray distribution along the line of sight .
the result is shown in figure [ fig : supgeom ] .
similar results have been found by @xcite .
this geometric reduction factor ranges from unity to zero .
the polarization of a number of spiral galaxies similar to the milky way has been measured by @xcite .
the observations are at 10.55 ghz and thus probe primarily synchrotron emission . for one of the best measured edge - on spirals ,
ngc 891 , they find : ( 1 ) at distances @xmath196kpc off the galactic plane the polarization fraction can be @xmath19710% ; and ( 2 ) in the plane , at heights @xmath198kpc , the polarization fraction drops to @xmath1995% .
similar behavior is seen by @xcite at 5 ghz .
in addition , @xcite show that ( 3 ) between 0.66 ghz and 1.5 ghz the spectral index ranges from @xmath200 in the plane to @xmath201 well off the plane .
_ wmap _ observes qualitatively similar behavior in k band . at 408 mhz ,
@xcite have surveyed the galactic plane in intensity . at this frequency
, synchrotron emission dominates maps .
we test the magnetic field model by extrapolating the 408 mhz measurements to 22 ghz ( an extrapolation of 40 in frequency and over 10,000 in amplitude ) : @xmath202 where @xmath203 is the ratio of the homogeneous field strength to the total field strength .
note that the model effectively has only one free parameter : an overall amplitude , which is described by a degenerate combination of the spectral index , @xmath160 and @xmath203 .
for @xmath204 , the best fit value for @xmath203 is @xmath205 .
this implies that the energy in the large scale field is roughly the same as the energy in small scale fields , consistent with other results for the milky way ( * ? ? ?
* ; * ? ? ?
* and references therein ) .
figure [ fig : haslam ] compares the k - band polarization signal to the extrapolated 408 mhz maps .
given the simplicity of the model ( uniform cosmic ray spectral index , @xmath180 , and a uniform lsa field ) , the agreement is remarkably good .
the largest deviations are seen near spiral arms .
recent observations @xcite suggest that cosmic rays are accelerated in star - forming regions .
if most cosmic rays are accelerated in spiral arms and then diffuse away from the arms , we would expect a flatter spectral index in the arms , consistent with the observations . in figure
[ fig : loops ] we show that the radio loops @xcite seen at 408 mhz , probably from supernovae or `` blowouts , '' are also seen in the _ wmap _ data .
measurements of starlight polarization serve as a template for the analysis of polarized microwave dust emission @xcite .
we have combined several catalogs of optical dust polarization measurements @xcite to construct a template for the magnetic field direction in dusty environments .
since there are significant variations in the dust column density , we only use the measured direction to construct the dust template .
the dust layer has a scale height of 100 pc @xcite .
observations toward the galactic poles suggest that most of the dust absorption occurs within 200 pc . to select stars outside the dust column for @xmath206 , we limit the sample to the 1578 stars with heliocentric distances greater than 500 pc . for @xmath207 ,
the model is problematic because there is ample dust emission from distances further away than the stars sample .
we represent the starlight polarization data , @xmath208 , in terms of a polarization amplitude , @xmath209 and direction , @xmath210 : @xmath211 we then smooth the starlight data by convolving @xmath212 and @xmath213 with a gaussian window with a fwhm of @xmath214 . the smoothing is required because the measurements are coarsely distributed . as a result , this dust model is applicable only for @xmath215 and @xmath206 . above
, @xmath216 describes the direction of this smoothed starlight polarization field .
we can quantify the agreement between the starlight and _ wmap _ k - band polarization measurements by computing their correlation in each pixel , @xmath217 where @xmath218 is the direction in k band .
figure [ fig : zzz ] shows a plot of the correlation as a function of position .
the median correlation coefficient is 0.72 implying that the dust and k - band directions typically agree to 20@xmath219 .
because of noise in both the k - band and starlight maps , this is an underestimate of the correlation . nevertheless , the correlation tells us that the basic model relating the starlight , the dust , synchrotron emission , and the magnetic field agrees with observations .
based on the detection of starlight polarization , thermal dust emission is expected to be polarized at millimeter and sub - millimeter wavelengths .
archaeops has detected polarized thermal emission at 353 ghz @xcite .
an extrapolation from this high frequency suggests that _ wmap _ should see polarized thermal dust emission at 94 ghz . here
, we report on the _ wmap _ detection of dust polarization at 94 ghz .
we generate a template for the dust polarization by using the maximum entropy method ( mem ) map @xcite ] .
dust intensity map @xcite , the smoothed polarization direction from the starlight , and the model geometric factor for the dust layer : @xmath220 where @xmath221 is the smoothed starlight polarization direction .
the geometric suppression factor for the dust , @xmath222 , is computed along the same lines as @xmath223 in equation [ eq : geom ] and is shown in figure [ fig : supgeom ] . to compute @xmath224 for the analog to equation [ eq : geom ] , we assume the dust has a scale height of 100 pc and a radial scale length of 3 kpc . to find @xmath225 we use the lsa magnetic field model .
the fractional polarization , @xmath226 , is found with a best fit of the model to the data .
similar results are found using the fds dust map @xcite instead of the mem dust map .
the uncertainty is estimated to be 50% .
figure [ fig : dust_template ] compares this predicted pattern of polarization to the cleaned w - band observations .
we use the k - band synchrotron template to clean q , v and w bands and then use the q and v band maps to remove the cmb polarization signal from the w - band maps ; though removing the cmb component is not necessary .
the w - band map is then smoothed with a 10@xmath219 beam for plotting .
the appearance of the dust polarization signal pattern is similar to that found by archeops ( * ? ? ?
* figures 2 & 3 ) . however , the signal to noise ratio is low due to the low level of polarized dust emission at 94 ghz .
the predominant feature is that the plane is dominated by positive stokes @xmath77 emission . a visual comparison to the model is less robust .
one must keep in mind that since stars are heavily obscured in the plane , the model is not expected to be accurate in the plane . nevertheless
, since stokes @xmath77 emission corresponds to the dominant horizontal magnetic field , one does not have to sample too deeply to pick it up .
similarly , we interpret the poor correlation between the model @xmath78 and the observed @xmath78 as due to the insufficient sampling of other magnetic field directions by rather limited depth of the stars .
some common features between the model and w - band data are seen for @xmath206 .
fits of the data to the model are given in section [ sec : fgr ] .
clearly , more integration time and more stellar polarization measurements are needed to fill out the model .
electric dipole emission from rapidly spinning dust grains is potentially a significant source of emission at _ wmap _ frequencies @xcite .
thermal fluctuations in the magnetization of magnetized grains may also be a potentially significant source of emission at microwave wavelengths @xcite . both have been proposed as an explanation for the correlations seen between thermal dust emission at 140 @xmath227 m and microwave emission in many cosmic background experiments : cobe @xcite , ovro @xcite , saskatoon @xcite , the 19 ghz survey @xcite , tenerife @xcite , python v @xcite , and cosmosomas @xcite .
the spectral shape of spinning dust emission can be similar to synchrotron emission in the 20 - 40 ghz range .
thus models with either variable synchrotron spectral index @xcite or with a spinning dust spectrum with a suitably fit cutoff frequency @xcite can give reasonable fits to the data .
however , at @xmath228ghz there is a considerable body of evidence , reviewed in @xcite and @xcite , that shows ( 1 ) that the synchrotron index varies across the sky steepening with increasing galactic latitude ( as is also seen in _ wmap _ ) and ( 2 ) that in other galaxies and our galaxy there is a strong correlation between 5 ghz synchrotron emission and 100 @xmath227 m ( 3000 ghz ) dust emission .
the combination of these two observations imply that the @xmath229 ghz _ wmap _ foreground emission is dominated by synchrotron emission as discussed in @xcite .
nevertheless , we must consider spinning dust as a possible emission source .
while on a galactic scale it appears to be sub - dominant , it may be dominant or a significant fraction of the emission in some regions or clouds .
spinning dust models predict an unambiguous signature in intensity maps : at 5 - 15 ghz , the dust emission should be significantly less than the synchrotron emission . @xcite and
@xcite argue that the tenerife and green bank data show evidence for a rising spectrum between 10 and 15 ghz , suggesting the presence of spinning dust .
observations of individual compact clouds also show evidence for spinning dust emission @xcite though the signature is not ubiquitous .
the status of the observations is discussed further in @xcite .
the _ wmap _ polarization measurements potentially give us a new way to distinguish between synchrotron and dust emission at microwave frequencies .
while synchrotron emission is expected to be highly polarized , emission from spinning dust grains is thought to be weakly polarized .
while promising , the signature is not unique as a tangle of magnetic field lines can also lead to a low polarization component @xcite as seen at 5 ghz where spinning dust emission is expected to be negligible . using a model for the polarization fraction of the synchrotron emission based on the lsa structure , we separate the microwave intensity emission into a high and low polarization component : @xmath230 where @xmath231 and @xmath232 are the intensity and polarization maps at frequency @xmath167 . for notational convenience , we use @xmath233 .
@xmath234 is the foreground - free cmb map made with a linear combination of _ wmap _ bands , and @xmath235 is the mem free - free map for band @xmath167 @xcite . in effect
, we use the _ wmap _ polarization maps to extract the intensity map of the low - polarization component in the data .
figure [ fig : lowpol ] compares the morphology of the low polarization k - band map to the w - band mem dust map @xcite . even in this simple model based on a number of assumptions ,
the agreement in morphology is striking .
we quantify this by computing the rms deviation between the two scaled maps , @xmath236 where w is the w - band map , the scale factor is @xmath237 , and the sum is taken over pixels .
we find @xmath238 .
in other words , we can `` predict '' the distribution of dust in w band from just the k band intensity and polarization maps . the low polarization fraction component has a spectral index of @xmath239 between k and q bands .
this correlation between the low polarization emission regions at 22 - 45 ghz and the thermal emission at 90 ghz and higher may be interpreted as either a very tight correlation between tangled field lines in star forming ( dusty ) regions or as evidence for spinning dust emission .
more polarization data , @xmath240 ghz observations , and extensive modeling are needed to conclusively delineate the magnitude and morphology of the various components . to compute the cmb power spectrum
, we must mask the regions with the brightest foreground emission . for polarization
we create a set of masks with a process that is somewhat analogous to the creation of the temperature masks @xcite .
first , the k - band @xmath77 and @xmath78 polarization maps are used to compute a positive - definite healpix r4 where @xmath241 for r4 , or resolution 4 @xcite .
see notation in @xcite .
] p map . from this a noise - bias variance map @xcite
is subtracted .
the rectified noise - bias correction is small because of the coarse resolution at r4 .
a histogram of pixel polarization amplitudes in this noise - bias - corrected map approximates a power law .
the peak is near the zero pixel value , there are just a few negative pixels ( due to the noise bias correction ) , and there is a long positive tail . unlike the process in which the temperature masks were created , there is no natural cut level based on the histogram peak . instead
, the cuts are given in terms of the mean of the noise bias corrected map of p at k band .
the cut level at the mean is denoted `` p10 '' . the cut level at 0.2 times
the mean is `` p02 '' , etc . for each cut level ,
a preliminary mask is made by setting r4 pixels greater than the cut level to 1 , and all others to 0 .
this mask is expanded to r9 and smoothed by a @xmath242 fwhm gaussian .
this mask map is set back to all 0s and 1s using the 0.5 level as a cut - off and the sense of the mask is reversed , so that the masked - out parts of the sky have zeros ( the _ wmap _ convention ) .
the above process results in a synchrotron polarization mask . in the case of temperature masks
, we found that additional masking based on the higher frequency bands was redundant .
this is not the case with polarization .
thus we make a dust polarization mask in a similar manner .
we begin with the first - year mem dust model box - averaged to r7 .
half the maximum value found in a subset of pixels in the polar caps ( @xmath243 ) is adopted as the cut - off level .
a preliminary mask is made by setting r7 pixels greater than the cut - off level to unity , and all others to zero .
this mask is then resolution expanded to a r9 map , smoothed by a @xmath244 gaussian , and set back to digital levels with a 0.5 cut - off .
the sense of the mask is then reversed to fit the _ wmap _ convention .
each synchrotron polarization mask is anded with the ( constant ) dust polarization mask and a constant polarized source mask .
we find , in general , that the extragalactic point sources are minimally polarized in the _ wmap _ bands , as discussed in @xcite .
we construct a source mask based on the exceptions . the most significant exception ( not already covered by the synchrotron or dust polarization masks ) is centaurus a , an extended and polarized source .
we found excellent agreement between _ wmap _ and previously published maps of cen a ( figure [ fig : crab ] ) .
based on this information , we custom - masked the full extent of cen a. six other bright polarized sources that we masked are fornax a , pictor a , 3c273 , 3c274 , 3c279 , pks 1209 - 52 .
( some bright polarized sources already covered by the synchrotron and dust mask regions include : 3c58 , orion a , taurus a , ic443 , 1209 - 52 , w51 , w63 , hb21 , ctb104a ) .
we have determined that , for most applications , the mask that we call `` p06 '' is the best compromise between maximizing usable sky area while minimizing foreground contamination . with the above considerations , the p06 mask masks 25.7% of the sky , mostly near the galactic plane .
we use the terminology `` outside the p06 mask '' to refer to data in the 74.3% of the sky left for cosmological analysis .
various masks are shown in figure [ fig : polmasks ] . based on our analysis of the galactic foreground emission ,
we have generated synchrotron and dust template maps for the purposes of foreground removal .
the template maps are fit and subtracted from the ka through w band data to generate cleaned maps that are used for cmb analysis .
we assess the efficacy of the subtraction with @xmath245 and by examining the residuals as a function of frequency and multipole @xmath51 , as described in
[ sec : psfgm ] .
we use the k - band data to trace the synchrotron emission , taking care to account for the ( relatively weak ) cmb signal in the k - band map when fitting and subtracting the template . for dust emission ,
we construct a template following equation ( [ eq : dust_template ] ) that is based on the starlight - derived polarization directions and the fds dust model eight @xcite evaluated at 94 ghz to trace the dust intensity , @xmath246 .
we call this combination of foreground templates `` kd3pol '' .
the synchrotron and dust templates are fit simultaneously in stokes @xmath77 and @xmath78 to three - years maps in ka through w bands .
the three - year maps are constructed by optimally combining the single - year maps for each da in a frequency band . specifically @xmath247 = \big ( \sum_i { \bf n}_i^{-1 } \big)^{-1 } \sum_i { \bf n}_i^{-1 } [ q_i , u_i]\ ] ] where @xmath248 is a combined year and da index , @xmath249 $ ] is a polarization map degraded to r4 @xcite , and @xmath250 is the inverse noise matrix for polarization map @xmath248 .
the fit coefficients , @xmath251 and @xmath252 are obtained by minimizing @xmath245 , defined as @xmath253 - \alpha_{s,\nu}[q_s , u_s ] - \alpha_{d,\nu}[q_d , u_d])^2}{[\sigma_q^2,\sigma_u^2]},\ ] ] where @xmath254 $ ] is the k - band polarization map ( the synchrotron template ) , @xmath255 $ ] is the dust template , and @xmath256 $ ] is the noise per pixel per stokes parameter in the three - year combined maps .
we have tried using optimal ( @xmath257 ) weighting for the fits as well , and found similar results for the coefficients .
the results reported here are based on the simpler diagonal weighting .
the fit is evaluated for all pixels outside the processing mask @xcite .
the fit coefficients are given in the top half of table [ tab : fgfit ] .
for each emission component we also report the effective spectral index derived from the fit : @xmath258 for synchrotron emission , and @xmath259 for dust .
these results indicate that the spectrum of the component traced by k - band is systematically flattening with increasing frequency , which is unexpected for synchrotron emission .
this behavior is statistically significant , and is robust to variations in the dust model and the data weighting .
we do not have a definitive explanation for this behavior . to guard against the possibility of subtracting cmb signal , we modified the template model as follows .
we take the four synchrotron coefficients in table [ tab : fgfit ] and fit them to a spectrum model of the form @xmath260 where @xmath261 , @xmath160 , and @xmath262 are model parameters that are fit to the @xmath263 , and @xmath264 is the conversion from antenna temperature to thermodynamic temperature at frequency @xmath167 .
this results in a modified set of synchrotron coefficients that are forced to follow a power - law that is largely determined by the ka and q - band results .
specifically , the modified coefficients are given in table [ tab : fgfit ] .
the implied synchrotron spectrum is @xmath265 .
this results in a 12% reduction in the synchrotron coefficients at q - band , and a 33% reduction at v - band .
however , because the k - band template is dominated by an @xmath266 e - mode signal ( see [ sec : psfgp06 ] ) , this change has a negligible effect on our cosmological conclusions , which are dominated by e - mode signal at @xmath267 .
a comparison of selected `` before and after '' cleaning maps is shown in figure [ fig : inout ] .
we also account for the cleaning in the map error bars .
since the k - band data are a combination of synchrotron and cmb emission , subtracting a scaled version of k band from a higher frequency channel also subtracts some cmb signal . if the fit coefficient to the higher frequency channel is @xmath268 , then the cleaned map is @xmath269 , where @xmath270 is the map and @xmath167 denotes the frequency band .
the maps we use for cosmological analysis were cleaned using the coefficients in the bottom half of table [ tab : fgfit ] .
the factor of @xmath271 dilates the noise in the cleaned map . to account for this in the error budget
we scale the covariance matrix of the cleaned map by a factor of @xmath272 .
additionally , we modify the form of the inverse covariance matrix by projecting from it any mode that has the k - band polarization pattern : @xmath273 , where @xmath274 = 0 $ ] .
this ensures that any residual signal traced by k - band ( due , for example , to errors in the form of the spectrum ) will not contribute to cosmological parameter constraints .
one measure of the efficacy of the foreground removal is the change in @xmath245 , relative to a null signal , between pre - cleaned and cleaned maps .
table [ tab : fgchi ] gives the values for the full sky and the p06 cut . in both cases
the full pixel covariance matrix was used to compute @xmath245 for stokes @xmath77 and @xmath78 simultaneously . for the full sky
the number of degrees of freedom , @xmath167 , is 6144 ( twice the number of pixels in an r4 map ) and outside the p06 mask @xmath275 .
note the large @xmath276 achieved with just a two parameter fit . by comparing the full sky to the p06 @xmath245 , we find that the starlight - based dust template is insufficient in the plane as discussed in [ sec : thde ] .
we also see that outside the p06 mask , that q and v bands are the cleanest maps and that they are cleaned to similar levels . since @xmath277 for q and v bands is so close to unity for the cleaned maps , it is no longer an effective measure of cleaning .
instead , we examine the power spectra @xmath51 by @xmath51 to assess the cleaning , and then test the sensitivity of the cosmological conclusions to cleaning by including ka and w band data . we have tried a number of variants on the kd3pol cleaning .
we find , for example , that setting @xmath278 across the sky has negligible effect on the fits or the derived optical depth . alternatively , when one uses the k - band polarization direction to trace the dust direction , @xmath279 in equation [ eq : dust_template ] , the cleaning is not as effective .
outside the p06 mask , the reduced @xmath245 in the q and v band maps is 1.022 and 1.016 as compared to 1.014 and 1.018 for the starlight - based directions .
thus the latter are used .
regardless of template , we find that our cosmological conclusions are relatively insensitive to the details as indicated in figure [ fig : tau_likes ] .
the @xmath77 & @xmath78 maps are well suited to analyzing foreground emission , are useful for comparing to other polarization maps , and have straightforward noise properties . however , they are not well suited to quantifying the cmb polarization anisotropy because their definition is coordinate dependent .
the @xmath77 and @xmath78 maps may be transformed into scalar and pseudo - scalar quantities called e and b modes @xcite .
e and b are so named because they comprise a curl - free and divergence - free decomposition of the spin-2 polarization field , analogous to static electric and magnetic fields .
the problem of separating e and b modes with an unevenly sampled and cut sky has been considered by a number of authors ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . in our analysis
, we work directly with @xmath77 and @xmath78 maps to produce the e and b angular power spectra . the conventions follow appendix a of @xcite . and @xmath280 of @xcite .
] fundamental symmetries in the production and growth of the polarized signal select the possible configurations for the cmb polarization .
scalar ( density ) perturbations to the matter power spectrum give rise to t and e modes .
tensor perturbations ( gravitational waves ) give rise to t , e , and b modes primarily at @xmath281 and @xmath282 , primordial b modes are dominated by the gravitational lensing of e modes . ] .
both scalar and tensor perturbations can produce polarization patterns in both the decoupling and reionization epochs .
vector perturbationscdm model to _ wmap _ tt data , the contribution of such modes is not large @xcite .
however , a formal search for them has not been done . ]
( both inside and outside the horizon ) are redshifted away with the expansion of the universe , unless there are active sources creating the vector modes , such as topological defects .
we do not consider these modes here . at the noise levels achievable with _
, the standard cosmological model predicts that only the e mode of the cmb polarization and its correlation with t will be detected .
the b - mode polarization signal is expected to be too weak for _ wmap _ to detect , while the correlations of t and e with b is zero by parity .
thus the tb and eb signals serve as a useful null check for systematic effects .
the polarization of foreground emission is produced by different mechanisms .
foreground emission can have any mixture of e and b modes , it can be circularly polarized ( unlike the cmb ) , and e and b can be correlated with t. we quantify the cmb polarization anisotropy with the @xmath283 , @xmath284 , and @xmath285 angular power spectra , where @xmath286 here the `` @xmath287 '' denote an ensemble average , @xmath288 are the multipoles of the temperature map , and @xmath289 are related to the spin-2 decomposition of the polarization maps @xmath290(\hat{x } ) = \sum_{\ell>0 } \sum_{m=-\ell}^{\ell}{{}_{\mp2}a_{\ell m}{}_{\mp2}y_{\ell m}(\hat{x})}\ ] ] via @xmath291 @xcite .
the remaining polarization spectrum combinations ( tb , eb ) have no expected cosmological signal because of the statistical isotropy of the universe .
we compute the angular power spectrum after applying the p06 polarization mask using two methods depending on the @xmath51 range .
all power spectra are initially based on the single - year r9 @xmath77 and @xmath78 maps @xcite . for @xmath292 , is the nyquist limit on @xmath51 .
for some analysis methods ( [ app : ektau ] ) we use healpix r3 for which @xmath293 we compute the power spectrum following the method outlined in @xcite , and ( * ? ? ?
* appendix a ) as updated in @xcite and appendix [ sec : masterp ] .
the statistical weight per pixel is @xmath294 where @xmath295 is the noise per observation @xcite . here
@xmath296 is a 2x2 weight matrix that multiplies the vector @xmath297 $ ] in each pixel @xmath298 where @xmath299 , @xmath300 , and @xmath301 are the elements of the weight arrays provided with the sky map data . note that the correlation between @xmath77 and @xmath78 within each pixel is accounted for .
we refer to this as `` @xmath296 weighting . '' from these maps , only cross power spectra between das and years are used .
the cross spectra have the advantage that only signals common to two independent maps contribute and there are no noise biases to subtract as there are for the auto power spectra .
the covariance matrices for the various @xmath302 are given in appendix [ app : covar ] .
for @xmath303 we mask and degrade the r9 maps to r4 ( see the last paragraph of appendix d and * ? ? ?
* ) so that we may use the full r4 inverse pixel noise matrix , @xmath304 , to optimally weight the maps prior to evaluating the pseudo-@xmath302 .
this is necessary because the maps have correlated noise that is significant compared to the faint cmb signal . by `` @xmath304 weighting '' the maps , we efficiently suppress modes in the sky that are poorly measured given the _ wmap _ beam separation and scan strategy ( mostly modes with structure in the ecliptic plane ) .
we propagate the full noise errors through to the fisher matrix of the power spectrum . for the spectrum plots in this section
, the errors are based on the diagonal elements of the covariance matrix which is evaluated in appendix [ app : master ] . figure [ fig : nobsninv ] shows the effect that correlated noise has on the low @xmath51 errors in the ee and bb spectra .
the curves show the diagonal elements of the inverse fisher matrix ( the @xmath302 errors ) computed in two ways : ( 1 ) assuming the noise is uncorrelated in pixel space and described by @xmath296 ( red ) and ( 2 ) assuming it is correlated and correctly described by @xmath304 ( black).the smooth rise in both curves toward low @xmath51 is due to the effects of @xmath100 noise and is most pronounced in the w4 da , which has the highest @xmath100 noise .
the structure in the black trace is primarily due to the scan strategy .
note in particular , that we expect relatively larger error bars on @xmath305 in ee and on @xmath306 in bb .
_ we caution those analyzing maps that to obtain accurate results , the @xmath304 weighting must be used when working with the @xmath303 power spectra .
_ for the monte carlo markov chains ( mcmc ) and cosmological parameter evaluation , we do not use the power spectrum but find the exact likelihood of the temperature and polarization maps given the cosmological parameters ( appendix [ app : ektau ] & * ? ? ?
* ) . for both r4 and r9 maps
there are 15 master cross power spectra ( see table [ tab : eebb ] ) . for the full three - year result , we form @xmath307 omitting the @xmath308 , @xmath309 , and @xmath310 auto power spectra . in this expression ,
@xmath311 and @xmath312 denote the frequency band ( k - w ) and @xmath248 and @xmath313 denote the year .
the noise per @xmath51 in the limit of no celestial signal , @xmath314 , is determined from analytical models that are informed by full simulations for r9 ( including @xmath100 noise ) , and from the full map solution for r4 .
figure [ fig : rawps_lspace ] shows the ee and bb power spectra for the region outside the p06 mask , 74.3% of the sky , before any cleaning .
the 15 cross spectra have been frequency averaged into four groups ( table [ tab : eebb ] ) by weighting with the diagonal elements of the covariance matrix .
data are similarly binned over the indicated ranges of @xmath51 .
it is clear that even on the cut sky the foreground emission is non negligible . in k band ,
we find @xmath315 and @xmath316 , where @xmath317 denotes the weighted average over multipoles two through six . the emission drops by roughly a factor of 200 in @xmath302 by 61 ghz resulting in @xmath318 for both ee and bb .
there is a `` window '' between @xmath319 and @xmath320 in the ee where the emission is comparable to , though larger than , the detector noise .
unfortunately , bb foreground emission dominates a fiducial @xmath8 , @xmath7 model by roughly an order of magnitude at @xmath321 . in general , the power spectrum of the foreground emission scales approximately as @xmath322 in @xmath323 .
figure [ fig : rawps_fspace ] shows the power spectra as a function of frequency for a few @xmath51 bands .
the spectrum of the emission follows that of synchrotron with @xmath324 with @xmath325 for both ee and bb . there is some evidence for another component at @xmath326 as seen in the flattening of the ee @xmath266 term .
we interpret this as due to dust emission . in the foreground model
, we explicitly fit to a dust template and detect polarized dust emission . however , there is not yet a sufficiently high signal to noise ratio to strongly constrain the dust index or amplitude outside the p06 mask .
a simple parameterization of the foreground emission outside the p06 mask region is given by @xmath327 we have introduced the notation @xmath328 to simplify the expression .
the `` d '' and `` s '' subscripts stand for `` dust '' and `` synchrotron . '' from an unweighted fit to all the raw @xmath329 data with the dust index fixed at @xmath330 , we find for ee @xmath331 , @xmath332 , @xmath333 and @xmath334 ; and for bb @xmath335 , @xmath336 , @xmath337 and @xmath334 . this model is given as an approximate guide .
its @xmath51 dependence is shown in figure [ fig : rawps_lspace ] for @xmath338ghz and its frequency dependence is shown in figure [ fig : rawps_fspace ] for bb @xmath266 .
one can see that this scaling model picks up the general trends but not the details of the foreground emission .
for example , it ignores correlations between dust and synchrotron emission .
it predicts an average foreground emission of @xmath106 ( @xmath227k)@xmath339 at 30 ghz and @xmath340 . @xcite
give an upper bound of @xmath341 for synchrotron emission in this range .
as dasi observes a relatively synchrotron - free region and at @xmath51s beyond where this simple parametrization can be tested , there is not a conflict with their results .
the same is true for the cbi experiment @xcite which also observed at 30 ghz but at a predominantly higher @xmath51 and in a predetermined clean region of sky .
the dust amplitude in the model is especially uncertain .
depending on the region of sky within the p06 cut , and the @xmath51 of interest , it may be an order of magnitude off . for a more complete model of the power spectra of foreground emission
, one must take into account the correlations or anticorrelations between various foreground components and between the foreground components and the cmb .
for example , a reasonable fit to the @xmath266 ee spectrum , which is dominated by foreground emission , is given by @xmath342 where @xmath343 is the dust synchrotron correlation coefficient , @xmath344 and @xmath345 are the frequencies of the two spectra that are correlated , the @xmath346 and @xmath160 are the dust and synchrotron spectral indices , and @xmath347 .
this fit is shown in figure [ fig : rawps_fspace ] .
after normalizing the frequency to 65 ghz , the following coefficients were found to reasonably represent the data : @xmath348 , @xmath325 , @xmath349 , @xmath330 , and @xmath350 . in order to produce the kv , kw , and kaw features
, there must be significant correlations between dust and synchrotron emission .
for the @xmath319 ee spectrum a similar expression fits the data if @xmath343 is negative .
some care is needed in interpreting the statistical significance of power spectra that include foreground emission and a cut sky .
the lack of statistical isotropy of the foreground emission means that it must be treated separately from the cmb when assessing the net noise . in the presence of foregrounds ,
the random uncertainty becomes @xmath351\ ] ] where @xmath352 is the foreground emission at each @xmath51 .
we plot only the first term in figures [ fig : rawps_lspace ] & [ fig : rawps_fspace ] to indicate the size of the statistical error .
additionally , with the sky cut there is a noise - foreground coupling between @xmath353 and @xmath354 , and between @xmath353 and @xmath355 .
this is analogous to the noise coupling shown in appendix [ app : fisher ] .
we next discuss the power spectrum after removing the foreground emission from the _
maps_. cleaning foregrounds not only changes the mean of @xmath302 , but it reduces @xmath356 because of the couplings .
the choice of model makes little difference to the conclusions .
for all the following we have subtracted the best fit kd3pol @xmath77 and @xmath78 templates from the ka through w maps ( both r4 and r9 versions ) as described in section [ sec : foregrounds ] .
table [ tab : eebb ] shows the ee @xmath266 and bb @xmath357 , the multipoles with the largest foreground contributions , for both before and after the subtraction . where the foreground signal is dominant , the subtraction can reduce its level by a factor of 6 - 10 in temperature .
when we fit and subtract the foreground templates , we use essentially all of the available data on polarized foreground emission .
the error bar on the power spectrum of the cleaned maps is dilated in the cleaning process as discussed above .
we do not include an additional error for systematic uncertainty in the model .
rather , by comparing spectra of pre - cleaned to cleaned maps , we estimate that the model removes at least 85% of the synchrotron .
this is demonstrated , for example , in the kka and kaka combinations for @xmath266 ee in table [ tab : eebb ] , in the subtraction shown in figure [ fig : inout ] , and to a lesser degree by the null eb and bb power spectra .
we also note that to a good approximation foreground emission adds only in quadrature to cmb emission .
figure [ fig : lbyleebb ] shows the power spectra of the foreground cleaned maps as a function of frequency for @xmath358 .
it also shows what we estimate to be the maximum levels of residual foreground contamination in the power spectrum . in the figure , we plot the synchrotron spectrum scaled to 0.15 of the pre - cleaned ka band value ( in temperature ) .
this shows that there is negligible residual synchrotron from 40 to 60 ghz with the possible exception of @xmath266 at 40 ghz .
given the size of the @xmath266 error bar , this potential contribution to the determination of the optical depth is negligible as discussed in section [ sec : tau ] . constraining the residual dust contamination is more difficult . in figure
[ fig : lbyleebb ] , we also show the mem temperature dust model scaled by 5% , a typical dust polarization value .
a similarly scaled fds model is almost identical .
this shows that even if we did not model and subtract dust , the contamination from it would not be large in q and v bands .
a more detailed model might have to take into account the possibility that the electrons and dust grains are in regions at different line of sight distances with different magnetic fields or that variations in the magnetic field could alias power from low multipoles to higher ones .
the cross power spectra of the cleaned maps are combined by frequency band for testing cosmological models .
the 10 cross spectra ( since k - band is used in the model , there is no k - band cleaned spectrum ) are assessed @xmath51 by @xmath51 with a least squares fit to a flat line in figure [ fig : lbyleebb ] .
the results are shown for the qq+qv+vv ( denoted `` qv combination '' ) and qv+kav combinations in ee , we find @xmath359 for all @xmath11 , where pte is for `` probability to exceed '' and is the probability that a random variable drawn from the same distribution exceeds the measured value of @xmath245 .
when w band is added to the mix , we find @xmath360 for @xmath361 , though all other values of @xmath51 give reasonable values . for bb ,
all frequency combinations yield reasonable ptes for all @xmath51 .
thus , there is a residual signal in our power spectra that we do not yet understand .
it is evident in w band in ee at @xmath362 and to a lesser degree at @xmath357 and @xmath363 .
we see no clear evidence of it anywhere else .
null tests are critical for assessing the quality of the data . we have examined the data in a wide variety of ways based on differencing assembly , frequency band , @xmath51 range , and year .
we present selected , though typical , results in the following .
a particularly important test is the null measurement of the bb , tb , and eb signals as shown in table [ tab : chisq ] and figure [ fig : chisq ] .
these data combinations are derived from the same processing as the ee , te , and tt combinations , where a signal is detected .
thus , the null result highlights the stability of the _ wmap _ data , the mapmaking , the foreground cleaning , and the power spectrum estimation .
the power spectrum of the difference of the individual yearly maps is another significant test .
table [ tab : chisq2 ] shows the results for all the yearly differences for @xmath364 , the critical region of @xmath51-space for the cosmological analysis .
we have also used the @xmath365 cross spectra to similar effect . here again the @xmath311 and @xmath312 denote different frequency bands .
this combination is equivalent to forming the power spectrum of the difference between year one and year two maps . in principle
it does not contain any signal .
the cross - spectrum method treats the noise in a slightly different way from the straight map method , where one must use the error bars from one of the maps .
it has been checked with simulations .
similar combinations are used for the other years .
using a variant of cross - spectrum method , we have also tested combinations of das for multiple ranges in @xmath51 within each frequency band . for all null tests ,
we find the expected null measurements , apart from the previously mentioned residuals at @xmath366 in w band . table [ tab : chisq ] gives the reduced @xmath245 for all combinations of t , e , and b data for a number of data combinations . from figure
[ fig : lbyleebb ] , it is clear that the large signal in w band is not residual dust contamination because the dust would not fit measurements in vw .
additionally , if one assumes that the polarized emission at a particular @xmath51 is a fraction times the intensity at the same @xmath51 , it would require @xmath367% dust polarization , which is unreasonable . though this simple picture does not take into account the aliasing of intensity from a lower @xmath51
, we do not observe a similar effect with the synchrotron emission , which in the simplest case is polarized by the same magnetic fields .
the w - band ee @xmath362 value is essentially unchanged by cleaning , removing a @xmath368 radius around the galactic caps , or by additionally masking @xmath369 in the ecliptic plane .
a number of tests have been done to identify this artifact of the data .
we are not yet certain if it is due to an ersatz signal or an incorrect noise term .
the error bars on the individual year differences are too large to clearly see if the effect is the same from year to year .
simulations show that @xmath100 alone can not explain the signal . the scan pattern in combination with the change in polarization is directly related to the large error bars at @xmath361 and is well understood .
we have not identified a mechanism that leads to a further increase in these uncertainties .
we know that different treatments of the noise , for example using @xmath296 weighting , decreases the magnitude of the discrepancy , though we are confident that the @xmath304 treatment of the pixel noise is the correct approach .
the discrepancy can be made smaller by eliminating the w1 data simply because the error bars increase .
the w1 radiometer has the lowest noise but also the largest number of `` glitches '' .
however , since we could not identify any correlation between the glitch rate ( assuming that unmasked glitches are responsible ) and the magnitude of the signal , we do not have a basis for eliminating this channel .
we believe there is an as yet unknown coupling in the w - band data that is driving the signal but more simulations and more sensitivity are needed to understand it .
we can not rule out similar lower - level problems in other bands , but we see no evidence of systematic effects in bb , eb , or other values of @xmath51 and other frequencies in ee . to avoid biasing the result by this residual artifact
which also possibly masks some unmodeled dust and synchrotron contamination , we limit the cosmological analysis to the qv combination .
we also show that including w band ee does not alter our conclusions . a comparison of the raw spectra and foreground cleaned spectra is shown in figure [ fig : correebb ] .
we start with the weighted sum of the 8 cross spectra with @xmath370 ghz ( without kw ) .
this is the upper - level line ( green ) in the figure .
the individual maps are then cleaned and the power spectra remade and coadded .
this is shown in violet .
similar comparisons are repeated for the qvw and qv combinations .
a simple visual inspection shows that even at the @xmath51s with the highest foreground contamination , the cleaning is effective .
from the bottom left panel in figure [ fig : correebb ] one sees that there is a clear signal above the noise in ee at @xmath371 . for the qv combination , @xmath372 .
the signal has persisted through a number of different analyses .
we can not rule out that this signal might find explanation in an unmodeled foreground component ; however , we find this explanation unlikely since the emission would have to be strikingly different from the measured spatial and frequency characteristics of the polarized foreground emission .
additionally , when different bands are coadded , the signal level is consistent : for qvw @xmath373 and for all channels with @xmath374ghz except kw , @xmath375 .
we have searched for systematic effects in the ee @xmath376 range and have not been able to identify any , other than the one discussed above .
we can not find a more plausible explanation than that the signal is in the sky .
we are thus led to interpret it cosmologically .
this is done in the next section .
we show the ee signal for @xmath377 in table [ tab : highee ] and in figure [ fig : highee ] along with a comparison to other recent measurements @xcite .
based on the best fit to the tt spectra , we produce a template for the predicted ee spectrum @xmath378 and form : @xmath379 where @xmath380 , @xmath381 is the fit amplitude , and @xmath382 is the diagonal fisher matrix in appendix [ app : covar ] . off
diagonal elements in @xmath383 have a negligible effect on the results .
the results of the fit are plotted in figure [ fig : eeteampphase ] for various frequency combinations .
we plot @xmath276 from the minimum value and find that @xmath384 for the pre - cleaned qvw combination , where the uncertainty is determined from the bounds at @xmath385 .
the reduced @xmath245 at the minima are 1.34 , 1.34 , 1.24 , and 1.30 for qv , vw , qvw , and kaqvw combinations respectively , most likely indicating residual foreground contamination . at a relative amplitude of zero ,
@xmath386 respectively for the same frequency combinations .
it is clear that the noise is not yet low enough to use just qv as was done at low multipoles .
in addition , cleaning the maps with the kd3pol is problematic because the k - band window function is reduced to 0.1 by @xmath387 .
when the same code is used to analyze eb and bb data , the fitted amplitude is always consistent with zero .
to summarize , the _ wmap _ ee data are consistent with a model of adiabatic fluctuations based on the temperature maps at greater than the @xmath388 level for the qvw and kaqvw combinations .
figure [ fig : tespec ] ( left panel ) shows the te spectrum for @xmath11 .
we use v band for temperature and the qv combination for polarization .
several aspects of the new processing led to increased errors and a reduced low-@xmath51 signal estimate relative to the first - year result ( figure 8 , * ? ? ? * ) .
these include : improvements in mapmaking and power spectrum estimation ( especially accounting for correlated noise and applying @xmath304 weighting ) ; limiting the bands to just q and v instead of ka - w ; increasing the cut from kp0 to p06 ; and improvements in foreground modeling , including a new estimate of dust polarization .
recall also that the first - year result was based on the combinations of ka , q , v , and w bands and did not include a dust polarization template in contrast to the new prescription .
furthermore , if the year - two data are processed in the same way as the first - year data , we obtain a spectrum similar to that in @xcite indicating that the major difference between first - year and three - year results rests on new knowledge of how to make and clean polarization maps .
the new spectrum is fully consistent with the first - year results and prefers a model based just on tt and ee data to a null signal at the @xmath388 level .
however , the new spectrum is also consistent with the absence of a te signal .
thus , it will take greater signal - to - noise to clearly identify the te signal with our new analysis methods .
figure [ fig : tespec ] ( right panel ) shows the te signal over the full range in @xmath51 .
other detections of te at @xmath27 have been reported by dasi ( @xmath389 ) @xcite , boomerang ( @xmath390 ) @xcite , and cbi ( @xmath391 ) @xcite .
the _ wmap _
data have had foreground models subtracted from both the temperature and polarization maps prior to forming the cross correlation .
the expected anticorrelation between the polarization and temperature is clearly evident . to quantify the consistency with the tt data we make a te template based on the model fit to tt .
next , a fit is made to the te data for @xmath392 with the following : @xmath393 where @xmath394 , @xmath395 is the predicted power spectrum shifted by @xmath396 , @xmath397 is the fit amplitude , and @xmath398 is the diagonal fisher matrix in appendix [ app : covar ] . off diagonal elements in @xmath398 have a negligible effect on the results .
similar 2d fits were done in @xcite .
we show the combination that uses v and w bands for t and q and v bands for e. the result , shown in figure [ fig : eeteampphase ] , is @xmath399 and @xmath400 with @xmath401 ( pte=0.66 ) .
similar results are obtained with other band combinations .
thus the te data are consistent with the tt data to within the limits of measurement .
figure [ fig : grandspec ] shows a summary of the various components of the cmb anisotropy .
the @xmath329 region of the cmb polarization spectra is rich with new tests of cosmology .
the ee spectrum gives us a new measure of the optical depth .
the same free electrons from reionization that lead to the @xmath402 ee signal act as test particles that scatter the quadrupolar temperature anisotropy produced by gravitational waves ( tensor modes ) originating at the birth of the universe .
the scatter results in polarization b modes .
tensor modes also affect the tt spectrum in this region .
a combination of these and related observations leads to direct tests of models of inflation .
the detection of the te anticorrelation near @xmath403 is a fundamental measurement of the physics of the formation of cosmological perturbations @xcite .
it requires some mechanism like inflation to produce and shows that superhorizon fluctuations must exist .
@xcite showed that with enough free parameters one could in principle make a model based on post - inflation causal physics that reproduced the tt spectrum .
@xcite show that the te anticorrelation is characteristic of models with superhorizon fluctuations .
the reason is that the anticorrelation is observed on angular scales larger than the acoustic horizon at decoupling .
thus , the observed velocity - density correlations implied by the te data must have existed on scales larger than the horizon and were not produced by post - inflation causal processes .
although multiple distinct physical mechanisms affect the @xmath329 spectra , their effects can be disentangled through an analysis of the full data complement @xcite . the separation , though , is not perfect and there remain degeneracies .
in particular , to some degree , the values of the scalar spectral index , @xmath404 , optical depth , and the tensor to scalar ratio , @xmath20 , may be traded against each other , although far less than in the first - year _ wmap _ results . as the data improve , or as more data sets are added , the degeneracy is broken further . in the following
we take a step back from the full mcmc analysis @xcite and estimate @xmath14 and @xmath20 from analyses of just the @xmath402 polarization spectra .
this approach aids our intuition in understanding what it is in the data that constrains the cosmological parameters .
our knowledge of the optical depth ripples through the assessment of all the cosmic parameters .
free electrons scatter the cmb photons thereby reducing the amplitude of the cmb spectrum .
this in turn directly impacts the determination of other parameters .
the distinctive signature of reionization is at @xmath402 in ee .
the only known contamination is from foreground emission which has been modeled and subtracted .
the amplitude of the reionization signal is proportional to @xmath14 in te and is proportional to @xmath405 in ee and bb . in the first year analysis
, we imposed a prior that @xmath406 @xcite .
such a high value would produce a signal @xmath407 times the model in figure [ fig : correebb ] and is clearly inconsistent with the ee data .
thus this new analysis is a significant improvement over the previously assumed prior .
we assess @xmath14 using three methods : ( 1 ) with template fits to the ee power spectra ; ( 2 ) with an exact likelihood technique based directly on the maps as described in appendix [ app : ektau ] ; and ( 3 ) with a multiparameter mcmc fit to all the data as reported in @xcite .
the first method is based directly on the master spectrum ( * ? ? ? * and appendix [ app : master ] ) of ee data and serves as a simple check of the other two .
additionally , the simplicity allows us to examine the robustness of the ee and te detections to cuts of the data .
the second method is robust and takes into account the phases of the ee and te signals .
it is run either as a stand alone method , as reported here , or as part of the full mcmc chain as reported in @xcite .
the best estimate of the optical depth comes from the full chains . for the template
fits , @xmath6cdm power spectra were generated for @xmath408 , with the remaining parameters fixed to @xmath409 , @xmath410 , @xmath411 , and @xmath412 .
for each spectrum , the scalar amplitude @xmath413 is fixed by requiring that @xmath414 .
we then form : @xmath415\ ] ] where @xmath416 is the data as shown in figure [ fig : correebb ] , @xmath417 is the model @xmath6cdm spectrum , @xmath418 as in c.14 , and @xmath419 is the uncertainty shown in figure [ fig : correebb ] and is derived from the master spectrum determination .
we use the symbol @xmath420 in this context because the likelihood function we obtain is not the full likelihood for @xmath14 .
uncertainties in other parameters , especially @xmath404 , have been ignored and the @xmath302 distribution is taken to be gaussian . thus @xmath421 does not give a good estimate of the uncertainty .
its primary use is as a simple parametrization of the data .
we call this method `` simple tau .
'' table [ tab : taucomp ] shows that simple tau is stable with data selection .
one can also see that if the qq component is removed from the qv combination , @xmath14 increases slightly .
this is another indication that foreground emission is not biasing the result .
additionally , one can see that removing @xmath422 for all band combinations does not greatly affect @xmath14 .
the optimal method for computing the optical depth is with the exact likelihood ( as in appendix [ app : ektau ] ) .
the primary benefits are : it makes no assumptions about the distribution of @xmath302 at each @xmath51 but does assume that the polarization signal and noise in the maps are normally distributed ; it works directly in pixel space , taking advantage of the phase relations between the t and e modes both together and separately ; and it is unbiased .
the only disadvantages are that it is computationally intensive and that it is not easy to excise individual values of @xmath51 such as @xmath422 . in the exact likelihood method
we take into account errors in our foreground model by marginalizing the likelihood function given in ( appendix [ app : ektau ] ) over the errors in the fitting coefficient for synchrotron emission , @xmath423 .
we ignore foreground errors in dust emission , as polarized dust emission is negligible in any of combinations of frequencies reported in table 9 .
when errors in @xmath423 are gaussian , the marginalization simply yields an additional term in the covariance matrix , @xmath424 where @xmath425 and @xmath426 and @xmath427 is a template map of polarized foreground ( i.e. , q and u maps in k band ) . here , the mean values , @xmath428 , are given in the second column of table 3 .
we find @xmath429 in q and v band .
to be conservative , we adopt @xmath430 as our foreground error , which is the 2-@xmath431 bound on the foreground error in qv combination . as the foreground marginalization yields a new positive term in the covariance matrix , a fraction of the signal that was attributed to cmb before is now attributed to foreground , when the spatial distribution of the signal is the same as that of k - band maps .
the values of @xmath14 with the foreground marginalization are tabulated table 9 .
the marginalization reduces @xmath14 by 0.0017 in qv .
the largest effect is seen in q band , for which @xmath14 drops by 0.0027 .
thus , the foreground error does not significantly affect our determination of @xmath14 .
table [ tab : taucomp ] shows that similar vales of @xmath14 are obtained for a wide variety of band combinations .
this is another indication that foreground emission is not significant .
we conservatively select the qv combination .
table [ tab : taucomp ] also compares the exact likelihood for the ee qv combination to the simple tau method .
one can see that simple tau is slightly biased high when compared to the exact likelihood and underestimates the likelihood at @xmath37 .
one source of the bias is the assumption of a gaussian likelihood .
nevertheless , it is reassuring that a variety of combinations of data give consistent values of @xmath14 .
the values given here are just for the ee and te data considered alone , with the first peak tt amplitude fixed .
when the exact likelihood is used in the full mcmc analysis @xcite yielding the best estimate , we find @xmath432 , slightly lower than the values reported here but with the same uncertainty , indicating that the simple analysis has exhausted most of the information on the optical depth contained in the polarization data .
as discussed in @xcite there is a degeneracy between the scalar spectral index , @xmath404 and @xmath14 .
if we had instead selected the k - band directions for the dust polarization template , we would have found @xmath433 and an increase in @xmath404 of 0.004 .
a similar shift would have been found using the kaqvw combination shown in table [ tab : taucomp ] and figure [ fig : tau_likes ] .
this is another indication of the relative insensitivity of the results to the cleaning method .
the @xmath434 spectrum directly probes the primordial gravitational wave background produced by tensor fluctuations in the early universe .
the existence of these gravitational waves was proposed by @xcite .
modern treatments may be found in , for example , @xcite . while scalar and tensor fluctuations both contribute to the tt and ee spectra , only tensors produce b modes .
inflation models generally predict similar scalar spectra , but differ in their prediction of the tensor component .
for example , ekpyrotic / cyclic models @xcite predict no observable tensor modes .
the tensor contribution is quantified with the tensor to scalar ratio @xmath20 .
we follow the convention in the camb code ( * ? ? ?
* version , june 2004 ) , in cmbfast v4.5.1 @xcite and in @xcite : @xmath435 here , @xmath436 and @xmath437 are the variance due to scalar and tensor modes respectively .
they are defined through @xmath438 where @xmath439 is the primordial tensor metric perturbation in real space that was generated during inflation and stretched to outside the horizon for slow - roll inflationary models with a single scalar field .
here , @xmath440 is the slow roll parameter related to the square of the slope of the inflaton potential . ] .
@xcite shows the @xmath46-dependence of these expressions .
the expression for @xmath20 is evaluated at @xmath441 corresponding to @xmath442 with the distance to the decoupling surface @xmath443mpc .
following @xcite , we use @xmath444 with @xmath445 in microkelvins .
some of the simple models of inflation in a @xmath6cdm cosmology predict @xmath446 ( e.g. , * ? ? ?
* ; * ? ? ?
for example , near this range inflationary models with a massive scalar field , @xmath447 , predict @xmath448 @xcite and models with a self coupling , @xmath449 , predict @xmath450 for @xmath451 . here , @xmath452 is the number of e - foldings before the end of inflation .
however , some variants produce @xmath453 ( e.g. * ? ? ?
* ) while many other have @xmath454 . for the best fit _
wmap_-only @xmath6cdm plus tensor model , the optical depth is @xmath455 . if we add to this model a tensor component with @xmath8 , then @xmath456 .
a simple average of the @xmath434 data gives @xmath457 , @xmath458 , @xmath459 , for qv , qvw , and @xmath374ghz ( no kw ) combinations respectively . to detect a signal at the upper range of the predictions would require maps with @xmath460 times smaller error bars .
we constrain @xmath20 by directly fitting a template of @xmath461 to the bb data . with the above definition ,
@xmath20 directly scales the @xmath462 power spectrum .
additionally , the amplitude of @xmath462 for @xmath11 scales as @xmath405 .
we set the template to be the standard @xmath6cdm model @xcite and use the single field inflation consistency relation , @xmath463 , to fix the tensor spectral index .
we assume the spectral index does not run and set @xmath409 .
we distinguish the @xmath20 in the template fit by the @xmath464 notation .
the sum is over @xmath465 .
the results of the fit are plotted in figure [ fig : simple_rtau ] .
when we consider just the limit on @xmath464 from the polarization spectra , ignoring the tensor contribution to tt , we find @xmath466 ( 95% cl ) after marginalizing over @xmath420 .
it is clear that the bb spectrum is not driving the limit on @xmath20 . after including the tt data
, the limit drops to @xmath467 ( 95% cl ) .
this shows that the tt data in combination with the limits on @xmath14 from ee and te are leading to the limit on @xmath20 .
the full mcmc analysis gives @xmath468 ( 95% cl ) with just the _ wmap _ data .
the increase in the error over the simple method given above is the result of the marginalization over the other parameters , particularly @xmath404 .
additionally , when @xmath404 is allowed to depend on @xmath46 , the error in @xmath20 increases dramatically , allowing @xmath469 ( 95% cl ) . we can relate @xmath20 to the current energy density in primordial gravitational radiation @xcite , @xmath470 where @xmath471 is conformal time and the transfer function , @xmath472 , is given in equation [ eq : tffinal ] . the approximation given in equation [ eq : ogw ] evaluated for @xmath473 and @xmath474 yields @xmath475 ( 95% cl ) and for @xmath23 , @xmath476 ( 95% cl ) .
_ wmap _ detects significant levels of polarized foreground emission over much of the sky .
the minimum in contamination is near 60 ghz outside the p06 mask . to detect the polarization in the cmb at @xmath402 a model of the foreground emission
must be subtracted from the data .
this situation differs from that of the analysis of the temperature anisotropy for which the foreground emission may be simply masked as a first approximation .
_ wmap _ has detected the primary temperature anisotropy , the temperature polarization cross correlation , and the e - mode polarization of the cmb .
we detect the optical depth with @xmath432 in a full fit to all _ wmap _ data .
this result is supported by stand - alone analyses of the polarization data .
using primarily the tt spectrum , along with the optical depth established with the te and ee spectra , the tensor to scalar ratio is limited to @xmath477 .
when the large scale structure power spectrum is added to the mix @xcite , the limit tightens to @xmath478 .
these values are approaching the predictions of the simplest inflation models .
a clear detection of the b modes at @xmath329 would give a direct handle on the physics of the early universe at energy scales of @xmath479gev .
this paper shows that care will be required to unambiguously separate the intrinsic signal from the foreground emission .
however , the bb spectrum is particularly clean in _ wmap _ and , at least for @xmath480 , the foreground contamination is relatively low . in the noise dominated regime , the error bar on @xmath434 decreases in proportion to time
. continued _ wmap _ operations combined with other experimental efforts are nearing a range of great interest .
these new results involve a complete reevaluation of all the components of our previous analyses , from the beams and gain models through to the mapmaking and foreground modeling . the data and most of the derived data products are available through the lambda website , http://lambda.gsfc.nasa.gov/. _ wmap
_ continues to operate nominally . in the future
we will address a number of the open issues raised above .
in particular , we can anticipate a better understanding of systematic errors and foreground emission , and therefore improved constraints on @xmath14 and @xmath20 .
it is remarkable that our understanding of the cosmos has reached the point where we have begun to quantitatively distinguish between different models of the birth of the universe .
we wish to thank uros seljak , suzanne staggs , and paul steinhardt for enlightening conversations on polarization and inflation , and jim peebles for his explanation of reionization .
we thank bruce draine for helpful discussions on dust polarization and rainer beck for many suggestions on modeling the galactic magnetic field .
we thank andr waelkens and torsten enlin for pointing out a mistake in the original calculation of @xmath223 .
( the fix affected figures 8 and 9 but not the conclusions . )
we thank bruce winstein for a detailed set of comments on the submitted draft .
ek acknowledges support from an alfred p. sloan research fellowship .
hvp wishes to acknowledge useful discussions with r. easther , s. larson , d. mortlock , and a. lewis .
the _ wmap _
mission is made possible by the support of the office of space sciences at nasa headquarters and by the hard and capable work of scores of scientists , engineers , technicians , machinists , data analysts , budget analysts , managers , administrative staff , and reviewers .
hvp is supported by nasa through hubble fellowship grant # hf-01177.01-a awarded by the space telescope science institute which is operated by the association of universities for research in astronomy , inc .
, for nasa under contract nas 5 - 26555 . this research has made use of nasa s astrophysics data system bibliographic services , the healpix software , camb software , and the cmbfast software .
this research was additionally supported by nasa ltsa03 - 000 - 0090 , nasa atpnng04gk55 g , and nasa adp03 - 0000 - 092 awards .
in this section we develop a simple model for the _ wmap _ instrument using jones matrices @xcite . in the following
we assume that all circuit elements are matched and ignore additive noise terms .
the jones matrix @xmath481 models the instrumental response to polarization , @xmath482 linearly relating the output electric field to the input .
wmap is a differential instrument , so the input radiation vector @xmath483 has four elements , @xmath484 , corresponding to the electric field seen by the a- and b - side feed pair .
the outputs @xmath485 are the inputs to the detectors .
the first link in the chain is to model the optics , feeds , and orthomode transducers ( omts ) .
we consider them as a single unit , because ascribing effects to the individual components is difficult and not well defined in terms of observations .
we include two effects , loss imbalance and polarization leakage : @xmath486 here @xmath487 is the loss for the particular polarization and @xmath488 quantifies the level of cross - polarization ( or polarization isolation ) leakage , which we model as a small rotation error .
the matrix @xmath489 is the first term in the expansion of a general unitary matrix but is not unitary itself .
the subscripts `` 1 '' and `` 2 '' refer to the two orthogonally polarized radiometers which are differenced to form @xmath490 .
the matrix @xmath489 is the first term in the expansion of a general unitary matrix , though it is not unitary itself .
the cross - polarization terms are allowed to have arbitrary phases @xmath491 .
it is possible for cross polarization to produce circular polarization but _ wmap_cannot detect it in @xmath492 or @xmath490 .
while in general there are four loss terms , two of them are calibrated out . the two that remain are the radiometer loss imbalances , @xmath493 and @xmath494 .
* table 3 ) measured the loss imbalances by fitting the response to the common mode cmb dipole signal , and found them to be @xmath151 .
the mean imbalance , @xmath495 , is corrected for by the map - making algorithm , while the `` imbalance in the imbalance '' , @xmath496 , is not @xcite . to connect the different notations , @xmath497 , @xmath498 , @xmath499 , and @xmath500 .
the @xmath501 and @xmath502 are calibrated out .
the next step is to model the radiometers .
they are described in detail in @xcite , so we simply present the jones representation of the radiometer and refer the reader to the paper for more details .
@xmath503 @xmath504 here , @xmath505 are the amplifier gains in the two legs of the radiometer , and @xmath506 is the instantaneous phase of the ( unjammed ) phase switch . we have lumped the warm and cold amplifiers together .
@xmath507 @xmath508 the detector outputs in counts @xmath509 are the diagonal elements of @xmath510 , multiplied by the responsivities @xmath511 .
@xmath512 is a 4x4 matrix with @xmath513 ( equation [ eq : a2 ] ) filling the upper left 2x2 entries and @xmath514 filling the lower right 2x2 entries .
@xmath515 @xmath516 @xmath517 in this expression , stokes @xmath77 , @xmath78 , and @xmath518 refer to the quantities measured in the radiometer reference frame ; we drop the `` rad '' notation used in [ sec : syst ] for notational convenience . before
the outputs are recorded they are demodulated in phase with the phase switch .
we model this process as @xmath519\ ] ] where @xmath520 is the phase difference between the two radiometer legs , and @xmath521 is the error between the two switch states .
since the input radiation is incoherent , @xmath522 since @xmath523 is the only frequency dependent component in the model , we make the substitution @xmath524 , where @xmath525 the calibrated detector outputs are @xmath526 , where @xmath527 is a gain for the temperature difference , @xmath528 here @xmath529 is the calibration uncertainty .
the radiometer signal channels are @xmath530 , from which are formed the temperature and polarization signal channels @xmath531 .
then to first order in the systematic uncertainties , @xmath532 here @xmath533 , @xmath534 are similarly defined , @xmath535 encodes the influence of the crosspol effects , and @xmath536 .
the dominant @xmath104 component is @xmath537 , not @xmath538 , because @xmath539 when the spacecraft rotates 180 . in the limit of no loss imbalance or calibration error , and similar cross polarization for all components , @xmath540
the wmap polarization power spectra at @xmath541 incorporate an extension of the master quadratic estimator @xcite , which is used to account for mode coupling .
the original method assumes that observations of every point on the sky give statistically independent noise .
however , wmap has a significant component of the noise that is correlated between pointings due to its scan pattern and the @xmath100 noise , and thus the method needs to be modified as described here to accommodate a full covariance matrix .
the most conspicuous mathematical feature of the original method is wigner 3-@xmath313 symbols , whereas in the extended method , these objects are not used . for more details of the original method , as well as the application to polarization ,
see appendix a of @xcite , together with the references therein .
the original method is derived by modeling the sky brightness as a continuous function of pointing .
for example , the observed cut - sky spherical harmonic coefficients for stokes @xmath76 , denoted as @xmath542 , are defined as follows : @xmath543 here , @xmath186 is the unit vector of the pointing , @xmath544 is the weighting function , @xmath545 is the sky brightness , and @xmath546 is a spherical harmonic basis function . expanding @xmath545 and @xmath547 in spherical harmonics gives a series .
each term of the series includes an integral of a product of three spherical harmonic basis functions : @xmath548
these distinctive integrals are what give rise to the 3-@xmath313 symbols .
the orthogonality relations of 3-@xmath313 symbols eliminate many terms in the expression for the observed power spectrum . when there is noise covariance , the weight is a function of two pointings rather than just one , and the 3-@xmath313 symbols are not used .
this case is most easily treated by modeling the sky as a set of discrete pixels .
the goal of the derivation is to form a mode - coupling matrix @xmath549 , where @xmath550 and @xmath551 are each chosen from the nine correlations @xmath552 , @xmath553 , @xmath554 , @xmath555 , @xmath556 , @xmath557 , @xmath558 , @xmath559 , and @xmath560 . in order to introduce the formalism
, we first discuss the @xmath552 correlation , which is the simplest .
because there is no coupling between @xmath552 and the other eight correlations , only @xmath561 needs to be considered .
we note here that we do not actually use this formalism for @xmath552 but only for the others , as the temperature power spectrum at low-@xmath562 is dominated by the signal and the noise correlation is not important .
we use @xmath552 here to illustrate the main point of the method .
the extension to the polarization power spectra that follows @xmath552 ( [ sec : masterp ] ) is what we use for the actual analysis .
the weighting is computed initially as the inverse of the covariance matrix of the pixels .
the sky cut is expressed by setting the appropriate rows and columns to a very large number in the noise covariance matrix before inverting it [ eq . ( [ eq : maskninv ] ) ] .
we call the resulting weight matrix @xmath563 .
further , let @xmath564 be a matrix containing ( appropriately normalized ) values of a spherical harmonic basis function evaluated at each pixel , @xmath180 .
index @xmath565 .
the number of rows of @xmath564 is @xmath566 , which is the number of pixels in each sky map .
the observed stokes @xmath76 sky map is @xmath567 . in this notation ,
the observed spherical harmonic coefficients are expressed as @xmath568 if the matrix @xmath563 is diagonal , this expression is simply the discrete version of eq .
[ eq-3j ] above .
expanding @xmath569 in spherical harmonics gives @xmath570 t_{l'm'}.\ ] ] this expression suggests the utility of defining @xmath571 so that @xmath572 the value of the observed power spectrum at @xmath562 is expressed as follows : @xmath573 in order to get the true , underlying cmb power spectrum into the equation , the next step is to take the expectation of eq .
[ eq - zz ] : @xmath574 therefore , we obtain the unbiased estimator of the underlying power spectrum as @xmath575 where @xmath576 in order to apply this method to cross - correlations between das , one of the @xmath577 matrices in eq .
[ eq - zsum ] is computed from the noise matrix of the first da , and the other from that of the second da .
the same formalism accommodates polarization . in what follows , uppercase @xmath578 or @xmath140
indicates one of the three harmonic transforms @xmath579 , @xmath580 , or @xmath581 , and lowercase @xmath582 or @xmath583 denotes the stokes parameter label @xmath76 , @xmath77 , or @xmath78 .
the following substitutions are made in the above derivation : @xmath584 where the non - zero elements of @xmath585 are @xmath586 @xmath587 are spin-2 spherical harmonics in the same matrix form as @xmath564 .
for each pair of das , a @xmath577 matrix is computed by analogy with eq .
[ eq - zdef ] .
the derivation follows the general steps above .
the analog of eq .
[ eq : ttrelation ] is @xmath588 where @xmath589 where @xmath590 for each da pair , the 81 coupling submatrices @xmath591 are combined in a grand coupling matrix that takes into account all the coupling among the nine correlation types .
the expressions for the coupling matrices greatly simplify when @xmath592 is diagonal in pixel space , @xmath593 .
this limit is a good approximation to the wmap data at high @xmath562 , where noise is approximately uncorrelated ( diagonal in pixel space ) . in this limit
, one can evaluate the coupling matrices analytically .
it is convenient to write the @xmath296 matrix as @xmath594 where @xmath595 one can show that under a rotation of basis by an angle @xmath596 , these quantities transform as @xmath597 therefore , we expand them into spin harmonics as follows : @xmath598 we obtain @xmath599 { \left(\begin{array}{ccc}{l}&{l}&{l'}\\ { 0}&{2}&{-2}\end{array}\right)}\right.\\ & & \left.+ \left [ { { } _ { + 4}{n}_{lm } } + ( -)^{l+l+l'}{{}_{-4}{n}_{lm } } \right]{\left(\begin{array}{ccc}{l}&{l}&{l'}\\ { -4}&{2}&{2}\end{array}\right)}\right\},\\ \nonumber z_{lm , l'm'}^{bb } & = & \frac12\sum_{lm}i^{lm}_{lm , l'm'}\left\ { n^+_{lm}\left[1+(-)^{l+l+l'}\right ] { \left(\begin{array}{ccc}{l}&{l}&{l'}\\ { 0}&{2}&{-2}\end{array}\right)}\right.\\ & & \left.- \left [ { { } _ { + 4}{n}_{lm } } + ( -)^{l+l+l'}{{}_{-4}{n}_{lm } } \right]{\left(\begin{array}{ccc}{l}&{l}&{l'}\\ { -4}&{2}&{2}\end{array}\right)}\right\},\\ \nonumber z_{lm , l'm'}^{eb } & = & \frac{i}2\sum_{lm}i^{lm}_{lm , l'm'}\left\ { n^+_{lm}\left[1-(-)^{l+l+l'}\right ] { \left(\begin{array}{ccc}{l}&{l}&{l'}\\ { 0}&{2}&{-2}\end{array}\right)}\right.\\ & & \left.- \left [ { { } _ { + 4}{n}_{lm } } - ( -)^{l+l+l'}{{}_{-4}{n}_{lm } } \right]{\left(\begin{array}{ccc}{l}&{l}&{l'}\\ { -4}&{2}&{2}\end{array}\right)}\right\},\end{aligned}\ ] ] where @xmath600 using the identity @xmath601 it is straightforward to evaluate all the relevant coupling matrices analytically : @xmath602 { \left(\begin{array}{ccc}{l}&{l}&{l'}\\ { 0}&{2}&{-2}\end{array}\right)}\right.\\ & & + \left .
\left[{}_4n_{lm}+(-)^{l+l+l'}{}_{-4}n_{lm}\right ] { \left(\begin{array}{ccc}{l}&{l}&{l'}\\ { -4}&{2}&{2}\end{array}\right)}\right|^2,\\ \nonumber m_{l l'}^{bb , bb } & = & \frac1{2l+1}\sum_{mm'}|z_{lm , l'm'}^{bb}|^2\\ \nonumber & = & \frac{2l'+1}{16\pi}\sum_{lm } \left|n_{lm}^+\left[1+(-)^{l+l+l'}\right ] { \left(\begin{array}{ccc}{l}&{l}&{l'}\\ { 0}&{2}&{-2}\end{array}\right)}\right.\\ & & -\left . \left[{}_4n_{lm}+(-)^{l+l+l'}{}_{-4}n_{lm}\right ] { \left(\begin{array}{ccc}{l}&{l}&{l'}\\ { -4}&{2}&{2}\end{array}\right)}\right|^2,\\ \nonumber m_{l l'}^{ee , bb } & = & \frac1{2l+1}\sum_{mm'}|z_{lm , l'm'}^{eb}|^2\\ \nonumber & = & \frac{2l'+1}{16\pi}\sum_{lm } \left|n_{lm}^+\left[1-(-)^{l+l+l'}\right ] { \left(\begin{array}{ccc}{l}&{l}&{l'}\\ { 0}&{2}&{-2}\end{array}\right)}\right.\\ & & -\left .
\left[{}_4n_{lm}-(-)^{l+l+l'}{}_{-4}n_{lm}\right ] { \left(\begin{array}{ccc}{l}&{l}&{l'}\\ { -4}&{2}&{2}\end{array}\right)}\right|^2,\\ \nonumber m_{l l'}^{eb , eb } & = & \frac1{2l+1}\sum_{mm'}z_{lm , l'm'}^{ee*}z_{lm , l'm'}^{bb}\\ \nonumber & = & \frac{2l'+1}{16\pi}\sum_{lm } \left\{\left|n_{lm}^+\left[1+(-)^{l+l+l'}\right ] { \left(\begin{array}{ccc}{l}&{l}&{l'}\\ { 0}&{2}&{-2}\end{array}\right)}\right|^2\right.\\ & & -\left .
\left|\left[{}_4n_{lm}+(-)^{l+l+l'}{}_{-4}n_{lm}\right ] { \left(\begin{array}{ccc}{l}&{l}&{l'}\\ { -4}&{2}&{2}\end{array}\right)}\right|^2\right\}.\end{aligned}\ ] ]
in this appendix , we derive expressions for the fisher and covariance matrices of the temperature and polarization power spectra .
our derivation extends the derivation of the tt matrices given in @xcite to all combinations of polarization power spectra .
note that we do _ not _ use these results for evaluating the likelihood that is used in the cosmological analysis . at low multipoles , @xmath603
, we evaluate the likelihood of polarization data directly from the maps using the exact method described in appendix [ app : ektau ] .
why do we not use the fisher or covariance matrix for the cosmological analysis , except for tt and te spectra at @xmath292 ?
the reason is because the form of the likelihood function for the power spectra is not a gaussian at low multipoles , and therefore the fisher or covariance matrix , which only characterize the second - order moment of the power spectrum , is not sufficient to fully specify the likelihood function .
this was pointed out after the first year release by @xcite and @xcite and is discussed in @xcite . as we do not know the precise form of the likelihood for the power spectra
, we evaluate the likelihood of the temperature and polarization maps directly , which is a gaussian , at low multipoles , @xmath603 . for high multipoles , @xmath604 , the likelihood function
may be approximated as a gaussian and therefore we use a gaussian likelihood with the fisher or covariance matrices .
while we do not use the ee or bb power spectra at @xmath604 , as they contain very little signal compared to noise , we do use the covariance matrix of the te power spectrum at @xmath604 in the likelihood code , for which we adopt the analytical _ ansatz _ given in equation [ eq : teansatz ] , which was also used in the first - year analysis of the te power spectrum @xcite .
for the evaluation of the tt likelihood , see @xcite . the fisher matrix , @xmath605 , is given by @xmath606 , \label{eq : fisher}\ ] ] where the covariance matrix @xmath607 consists of the covariance matrices of all the bilinear combinations of @xmath579 , @xmath77 , and @xmath78 : @xmath608 and the covariance includes the signal and noise , @xmath609 .
here @xmath610 is the angular ( cross ) power spectrum of the signal where @xmath578 and @xmath140 denote @xmath579 , @xmath580 , or @xmath581 .
the inverse covariance matrix in harmonic space is then given by the harmonic transform of @xmath611 : @xmath612 where @xmath585 is given by the equations following ( [ eq : b8 ] ) .
using these quantities , each term of the fisher matrix ( eq .
[ [ eq : fisher ] ] ) evaluates to @xmath613 ^ 2,\\ f_{ll'}^{xx , xy } & = & \sum_{mm'}\left[(c^{-1})^{xx}_{lm , l'm'}(c^{-1})^{xy}_{lm , l'm'}\right],\\ f_{ll'}^{xx , yy } & = & \frac12 \sum_{mm'}\left[(c^{-1})^{xy}_{lm , l'm'}\right]^2,\\ \label{eq : fxy } f_{ll'}^{xy , xy } & = & \sum_{mm'}\left[(c^{-1})^{xy}_{lm , l'm'}\right]^2 + \sum_{mm'}\left[(c^{-1})^{xx}_{lm , l'm'}(c^{-1})^{yy}_{lm , l'm'}\right],\end{aligned}\ ] ] where @xmath614 . in general cases where @xmath615 or @xmath616 ( or both ) are non - diagonal , one must calculate @xmath611 by directly inverting the covariance matrix given by equation ( [ eq : fullcovariance ] ) .
in reality , however , the matrix inversion requires @xmath617 operations and thus it become computationally too expensive to evaluate for the full wmap resolution . on the other hand ,
if one considers only large scale anisotropies at low @xmath562 , then the matrix inversion can be done in a reasonable computational time .
we use eq .
( [ eq : fxx])([eq : fxy ] ) for computing the fisher matrices for @xmath618 , @xmath619 , @xmath620 , @xmath621 , @xmath622 , and @xmath462 , at low multipoles , @xmath623 .
the expressions for the fisher matrices can be evaluated analytically when @xmath607 is diagonal in pixel space .
this limit is a good approximation to the _ wmap _ data at high @xmath562 , where @xmath607 is dominated by noise and noise is approximately uncorrelated ( diagonal in pixel space ) . in this limit ,
one obtains the following analytical formulae : @xmath624 { \left(\begin{array}{ccc}{l}&{l}&{l'}\\ { 0}&{2}&{-2}\end{array}\right)}\right.\\ & & + \left .
\left[{}_4n_{lm}+(-)^{l+l+l'}{}_{-4}n_{lm}\right ] { \left(\begin{array}{ccc}{l}&{l}&{l'}\\ { -4}&{2}&{2}\end{array}\right)}\right|^2,\\ \nonumber f_{l l'}^{bb , bb } & = & \frac1{2}\sum_{mm'}|(n^{-1})_{lm , l'm'}^{bb}|^2\\ \nonumber & = & \frac{(2l+1)(2l'+1)}{32\pi}\sum_{lm } \left|n_{lm}^+\left[1+(-)^{l+l+l'}\right ] { \left(\begin{array}{ccc}{l}&{l}&{l'}\\ { 0}&{2}&{-2}\end{array}\right)}\right.\\ & & -\left .
\left[{}_4n_{lm}+(-)^{l+l+l'}{}_{-4}n_{lm}\right ] { \left(\begin{array}{ccc}{l}&{l}&{l'}\\ { -4}&{2}&{2}\end{array}\right)}\right|^2,\\ \nonumber f_{l l'}^{ee , bb } & = & \frac1{2}\sum_{mm'}|(n^{-1})_{lm , l'm'}^{eb}|^2\\ \nonumber & = & \frac{(2l+1)(2l'+1)}{32\pi}\sum_{lm } \left|n_{lm}^+\left[1-(-)^{l+l+l'}\right ] { \left(\begin{array}{ccc}{l}&{l}&{l'}\\ { 0}&{2}&{-2}\end{array}\right)}\right.\\ & & -\left .
\left[{}_4n_{lm}-(-)^{l+l+l'}{}_{-4}n_{lm}\right ] { \left(\begin{array}{ccc}{l}&{l}&{l'}\\ { -4}&{2}&{2}\end{array}\right)}\right|^2,\\ \nonumber m_{l l'}^{eb , eb } & = & \sum_{mm'}|(n^{-1})_{lm , l'm'}^{eb}|^2 + \sum_{mm'}(n^{-1})_{lm , l'm'}^{ee*}(n^{-1})_{lm , l'm'}^{bb}\\ \nonumber & = & \frac{(2l+1)(2l'+1)}{32\pi}\sum_{lm } \left\ { \left|n_{lm}^+\left[1-(-)^{l+l+l'}\right ] { \left(\begin{array}{ccc}{l}&{l}&{l'}\\ { 0}&{2}&{-2}\end{array}\right)}\right.\right.\\ \nonumber & & -\left .
\left[{}_4n_{lm}-(-)^{l+l+l'}{}_{-4}n_{lm}\right ] { \left(\begin{array}{ccc}{l}&{l}&{l'}\\ { -4}&{2}&{2}\end{array}\right)}\right|^2\\ \nonumber & + & \left|n_{lm}^+\left[1+(-)^{l+l+l'}\right ] { \left(\begin{array}{ccc}{l}&{l}&{l'}\\ { 0}&{2}&{-2}\end{array}\right)}\right|^2\\ & & -\left .
\left|\left[{}_4n_{lm}+(-)^{l+l+l'}{}_{-4}n_{lm}\right ] { \left(\begin{array}{ccc}{l}&{l}&{l'}\\ { -4}&{2}&{2}\end{array}\right)}\right|^2\right\}.\end{aligned}\ ] ] the inverse of the fisher matrix gives the covariance matrix , @xmath625 .
while we use the map - based exact likelihood described in appendix [ app : ektau ] for the cosmological analysis , it is still useful to have an approximate method to evaluate the likelihood of the data given theory and noise model from the power spectra . for this purpose , we use the following _ ansatz _ : @xmath626 ^ 2 } \label{eq : teansatz}\ ] ] @xmath627 ^ 2}\ ] ] @xmath628 ^ 2}\ ] ] @xmath629 ^ 2}\ ] ] @xmath630 ^ 2}\ ] ] in these expressions @xmath631 denotes the effective noise as a function of @xmath51 and @xmath632 denotes the effective fraction of the sky observed .
these are obtained from comparing the _ ansatz _ to the inverse of the fisher matrices derived in the previous sections .
we have found that @xmath633 to a very good approximation .
see also @xcite for the evaluation of @xmath634 and @xcite for the evaluation of @xmath635 .
at low multipoles , @xmath603 , we evaluate the likelihood of the data for a given theoretical model exactly from the temperature and polarization maps .
the standard likelihood is given by @xmath636}{|s+n|^{1/2 } } \frac{d\vec{m}}{(2\pi)^{3n_p/2 } } , \label{eq : like}\ ] ] where @xmath637 is the data vector containing the temperature map , @xmath638 , as well as the polarization maps , @xmath639 , and @xmath640 , @xmath566 is the number of pixels of each map , and @xmath112 and @xmath641 are the signal and noise covariance matrix ( @xmath642 ) , respectively . as the temperature data are completely dominated by the signal at such low multipoles , noise in temperature
may be ignored .
this simplifies the form of likelihood as @xmath643}{|\tilde{s}_p+n_p|^{1/2 } } \frac{d\vec{\tilde{m}}}{(2\pi)^{n_p}}~ \frac{\exp\left(-\frac12\vec{t}^ts_{t}^{-1}\vec{t}\right)}{|s_{t}|^{1/2 } } \frac{d\vec{t}}{(2\pi)^{n_p/2 } } , \label{eq : dnslike}\ ] ] where @xmath644 is the temperature signal matrix ( @xmath645 ) , the new polarization data vector , @xmath646 , is given by @xmath647 and @xmath648 is the signal matrix for the new polarization vector with the size of @xmath649 .
as @xmath650 is totally signal dominated , the noise matrix for @xmath651 equals that for @xmath652 , @xmath566 . to estimate @xmath650
, we used the full - sky internal linear combination ( ilc ) temperature map @xcite .
one can show that equation ( [ eq : like ] ) and ( [ eq : dnslike ] ) are mathematically equivalent when the temperature noise is ignored .
the new form , equation ( [ eq : dnslike ] ) , allows us to factorize the likelihood of temperature and polarization , with the information in their cross - correlation , @xmath653 , fully retained .
we further rewrite the polarization part of the likelihood as @xmath654}{|n_p^{-1 } \tilde{s}_pn_p^{-1}+n_p^{-1}|^{1/2 } } \frac{|n_p^{-1}|d\vec{\tilde{m}}}{(2\pi)^{n_p}}. \label{eq : eklike}\ ] ] this form is operationally more useful , as it contains only @xmath655 .
@xcite describes the method to evaluate the temperature part of the likelihood . the effect of _
p06 _ mask is included in @xmath655 .
suppose that the structure of @xmath655 is given by @xmath656 where @xmath413 is the noise matrix for unmasked pixels , @xmath657 is for masked pixels , and @xmath581 is for their correlations .
we assign infinite noise to the masked pixels such that @xmath658 , where @xmath270 is the diagonal matrix whose elements are zero for masked pixels and unity otherwise . in the limit of @xmath659 , the inverse of @xmath660 is given by @xmath661 we have checked that this form of @xmath655 yields the unbiased estimates of the signal matrix from simulated realizations of the _ wmap _ data .
when the masked pixels were simply ignored ( i.e. , @xmath662 ) , on the other hand , the estimated signal matrix was found to be biased high .
as the likelihood form is sensitive to the precise form of @xmath655 , it is important to treat the mask in this way so that the estimated signal matrix from the data is unbiased .
we mask the polarization maps as follows .
we first mask the maps at the full resolution , @xmath663 , and then degrade the masked maps using the weight that is diagonal in pixel space , @xmath664 , to a lower resolution , @xmath665 .
( note that while the weight is diagonal in pixel space , it contains noise covariance between @xmath666 and @xmath667 .
the spurious polarization term , @xmath112 , is ignored in this process . )
the degraded mask is redefined such that it takes on 1 when the lower resolution pixel contains more than half of the original full resolution pixels , and 0 otherwise .
we degrade these maps further to the resolution of @xmath668 using the full noise matrix , and also degrade the mask and the noise matrix .
( the noise matrix has been masked using eq [ [ eq : maskninv ] ] . )
we use the resulting maps and noise matrix in the likelihood function given in equation ( [ eq : eklike ] ) .
tensor perturbations generated by inflation are stochastic in nature , so the gravity wave perturbation can be expanded in plane waves @xmath670,\ ] ] where @xmath671 is the polarization tensor , and @xmath672 are the two polarizations in the transverse traceless ( tt ) gauge ( in which @xmath673 ; we also set @xmath674 ) .
the stress - energy tensor for gravity waves is defined as @xmath675 and in the tt gauge , we have @xmath676 thus , @xmath677 \end{aligned}\ ] ] the variance of the perturbations in the @xmath678 fields can be written as @xmath679 and since @xmath680 , we obtain @xmath681.\ ] ] writing @xmath682 where @xmath579 is the transfer function , we have @xmath683 \dot{t}^2(\eta , k ) \nonumber \\ & = & \int \frac{dk}{k } \frac{2k^3}{2\pi^2 } \left[\langle|h_+ ( 0,\mathrm{\bf k})|^2\rangle + \langle |h_\times ( 0,\mathrm{\bf k})|^2\rangle \right ] \dot{t}^2(\eta , k).\end{aligned}\ ] ] from the definition of the primordial tensor power spectrum , @xmath684,\ ] ] we obtain @xmath685 now @xmath686 thus we have @xmath687 the transfer function @xmath579 and its time derivative @xmath691 can be calculated easily by numerically integrating the evolution equation for the polarization states , which , neglecting the neutrino anisotropic stress , is given by @xmath692 where prime denotes derivatives with respect to conformal time @xmath471 , related to the time derivative by @xmath693 .
this expression may be numerically integrated . in the following , however , we derive an analytic estimate relating a given limit on the tensor - to - scalar ratio , @xmath20 , and the measured amplitude of the primordial scalar power spectrum , @xmath413 , to a limit on the current energy density in primordial gravitational radiation .
there are several approaches taken in the literature to derive analytic expressions for the tensor transfer function , though these results are obtained in almost all cases for a universe containing only matter and radiation .
these include using ( 1 ) an instantaneous transition from radiation to matter domination ( * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* e.g. , ) ( 2 ) a `` transfer function '' to account for the smooth transition from radiation domination to matter domination ( * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* e.g. , ) , and ( 3 ) wkb methods ( * ? ? ?
* ; * ? ? ?
* e.g. , ) . in the following derivation
, we will apply the sudden transition approximation to a @xmath6cdm universe ( * ? ? ?
* see also ) , which is a good approximation for gravitational waves with wavelengths much longer than the time taken for the transition to happen . in a universe which undergoes a set of piecewise instantaneous transitions in the scale - factor , given by @xmath694 , the solution to eq .
[ eq : kgeqn4h ] is given by @xmath695,\ ] ] where @xmath696 and @xmath697 are spherical bessel functions of order @xmath167 of the first and second kinds , respectively .
here , @xmath698 for radiation - domination ( rd , @xmath699 ) , @xmath700 for matter - domination ( md , @xmath701 ) , and @xmath702 for @xmath6-domination ( ld , @xmath703 ) .
@xmath704 is the conformal time at radiation - matter equality , with a scale - factor corresponding to @xmath705 , and @xmath706 is the conformal time at matter-@xmath6 equality , with a scale - factor corresponding to @xmath707 . for a concordance cosmology with @xmath708 , @xmath709 mpc@xmath710 and @xmath711 mpc@xmath710 ( 115 and 12030 mpc@xmath710 respectively in the instantaneous approximation ) . to obtain the coefficients @xmath712 and @xmath657 , we require @xmath678 and @xmath713 to be continuous at each of the transitions , @xmath704 and @xmath706 . thus , denoting @xmath714 and making use of special properties of spherical bessel functions , we obtain the transfer function and its derivative at present : @xmath715 , \\ \label{eq : tffinal } \dot{t}(k , x ) & = & k x^2 \left[c\ j_0(x ) + d\ y_0(x ) \right ] .
\label{eq : tdotfinal}\end{aligned}\ ] ] the coefficients are given by @xmath716\ ] ] @xmath717\ ] ] @xmath718 where @xmath719 and @xmath720 .
combining these equations , and evaluating them at the present conformal time @xmath724 ( with @xmath725 ) for modes within our current horizon , we are left with @xmath726 ^ 2,\ ] ] where @xmath46 and @xmath724 are to be evaluated in units of 1/mpc and @xmath727 mpc@xmath710 .
we can now change to the dimensionless variable @xmath714 and obtain @xmath728 ^ 2,\ ] ] where @xmath729 .
we also have the result @xmath730 ^ 2 \label{eq : domdlnk } \\ &
\simeq & 2.21 \times 10^{-3 } ( ra ) \left ( \frac{k}{k_0 } \right)^{-r/8 } \left\{k x^2 \left[c\ j_0(x ) + d\ y_0(x)\right]\right\}^2.\end{aligned}\ ] ] now @xmath731 and @xmath732 for the concordance @xmath6cdm model .
taking the concordance model and @xmath727 mpc@xmath710 , for given upper limits on @xmath20 and @xmath413 , the upper limit on @xmath669 is given by @xcite , @xmath733 , \label{eq : ogw}\ ] ] where the logarithm is taken in base ten .
, s. & lazarian , a. 1999 , in microwave foregrounds .
sloan summit , institute for advanced study , princeton , new jersey 14 - 15 november 1998 .
eds : a. rde oliveira - costa and m. tegmark .
181 ( astronomical society of the pacific ) , 113 , m. & brunthaler , a. 2005 , in future directions in high resolution astronomy : the 10th anniversary of the vlba , asp conference proceedings , eds : j. romney and m. reid .
, vol . 340
( astronomical society of the pacific ) , 253 ccccccc _
wmap_&22.5 ( k ) & @xmath734 & @xmath735 & @xmath736 & @xmath737 & @xmath738 ( @xmath739 ) + _ wmap_&32.8 ( ka ) & @xmath740 & @xmath741 & @xmath742 & @xmath743 & @xmath744 ( @xmath745 ) + _ wmap_&40.4 ( q ) & @xmath746 & @xmath747 & @xmath748 & @xmath749 & @xmath750 ( @xmath751 ) + _ wmap_&60.2 ( v ) & @xmath752 & @xmath753 & @xmath754 & @xmath755 & @xmath756 ( @xmath757 ) + _ wmap_&92.9 ( w ) & @xmath758 & @xmath759 & @xmath760 & @xmath761 & @xmath762 ( @xmath763 ) + @xcite & 19 & & & & 6.6 [ 15.5]&(@xmath764 ) + @xcite & 23 & & & & 9 & ( @xmath765 ) + @xcite & 31 & & & & 8.1 [ 17 ] & ( @xmath766 ) + @xcite & 33 & & & & [ 16 ] & ( [ @xmath767 ) + @xcite & 86 & & & & ( [ @xmath768 ) & + @xcite & 88 & & & & 13 & ( @xmath765 ) + @xcite & 99 & & & & [ @xmath769 & ( [ @xmath770 ) + @xcite & 273 & & & & [ @xmath771 & ( [ @xmath772 ) + @xcite & 363 & & & & @xmath773 & ( @xmath774 ) + ccccc ka & 0.3103 & @xmath776 & 0.0148 & 1.54 + q & 0.1691 & @xmath777 & 0.0154 & 1.89 + v & 0.0610 & @xmath778 & 0.0343 & 1.92 + w & 0.0358 & @xmath779 & 0.0891 & + ka & 0.2973 & @xmath780 & 0.0148 & 1.54 + q & 0.1492 & @xmath780 & 0.0154 & 1.89 + v & 0.0414 & @xmath780 & 0.0343 & 1.92 + w & 0.0112 & @xmath780 & 0.0891 & + ccccc ka & 10.65 & 1.20 & 6144 &
58061 + q & 3.91 & 1.09 & 6144 & 17326 + v & 1.36 & 1.19 & 6144 & 1045 + w & 1.38 & 1.58 & 6144 & -1229 + ka & 2.142 & 1.096 & 4534 & 4743 + q & 1.289 & 1.018 & 4534 & 1229 + v & 1.048 & 1.016 & 4534 & 145 + w & 1.061 & 1.050 & 4534 & 50 + ccccccccc kk & 1 & 22.8&@xmath781 & & @xmath782 & & & + kka & 2 & 27.4&@xmath783 & @xmath784 & @xmath785 & @xmath786 & @xmath787 & @xmath788 + kq & 2 & 30.5&@xmath789 & @xmath790 & @xmath791 & @xmath792 & @xmath793 & @xmath794 + kv & 3 & 37.2&@xmath795 & @xmath796 & @xmath797 & @xmath798 & @xmath799 & @xmath800 + kw & & 46.2&@xmath801 & @xmath802 & @xmath803 & @xmath804 & @xmath805 & @xmath806 + kaka & 2 & 33.0&@xmath807 & & @xmath808 & & @xmath809 & @xmath810 + kaq & 3 & 36.6&@xmath811 & @xmath812 & @xmath813 & @xmath814 & @xmath815 & @xmath816 + kav & 4 & 44.8&@xmath817 & @xmath818 & @xmath819 & @xmath820 & @xmath821 & @xmath822 + kaw & 4 & 55.5&@xmath823 & @xmath824 & @xmath825 & @xmath826 & @xmath827 & @xmath828 + qq & 4 & 40.7&@xmath829 & @xmath830 & @xmath831 & @xmath832 & @xmath833 & @xmath834 + qv & 4 & 49.7&@xmath835 & @xmath836 & @xmath837 & @xmath838 & @xmath839 & @xmath840 + qw & 4 & 61.7&@xmath841 & @xmath842 & @xmath843 & @xmath844 & @xmath845 & @xmath846 + vv & 4 & 60.8&@xmath847 & @xmath848 & @xmath846 & @xmath849 & @xmath850 & @xmath851 + vw & 4 & 75.4&@xmath852 & @xmath853 & @xmath846 & @xmath849 & @xmath854 & @xmath855 + ww & 4 & 93.5&@xmath856 & @xmath857 & @xmath858 & @xmath859 & @xmath860 & @xmath861 + ccccc te & 0.31 ( 0.99 ) & 1.01 ( 0.46 ) & 1.20(@xmath862 ) & 1.08(0.06 ) + te ( 2003 ) & 1.88 ( 0.03 ) & 1.18 ( 0.25 ) & 2.06 ( 0 ) & + tb & 0.57 ( 0.90 ) & 0.72 ( 0.97 ) & 0.97 ( 0.70 ) & 0.97 ( 0.74 ) + ee & 1.34 ( 0.17 ) & 1.06 ( 0.33 ) & 0.98 ( 0.59 ) & 0.96 ( 0.76 ) + bb & 0.72 ( 0.77 ) & 1.28 ( 0.04 ) & 0.96 ( 0.73 ) & 0.95 ( 0.81 ) + eb & 0.41 ( 0.98 ) & 1.21 ( 0.09 ) & 1.03 ( 0.34 ) & 0.96 ( 0.76 ) + cccc te & 1.70 ( 0.04 ) & 1.05 ( 0.40 ) & 1.87 ( 0.02 ) + tb & 1.95 ( 0.02 ) & 1.20 ( 0.26 ) & 1.08 ( 0.37 ) + ee & 1.55 ( 0.08 ) & 0.89 ( 0.58 ) & 0.55 ( 0.91 ) + bb & 0.56 ( 0.90 ) & 1.50 ( 0.09 ) & 0.76 ( 0.72 ) + eb & 0.62 ( 0.86 ) & 1.04 ( 0.41 ) & 0.84 ( 0.63 ) + cccccccc qv & @xmath863 & @xmath864 & @xmath865 & @xmath866 & @xmath867 & @xmath868 & @xmath869 + vw & @xmath870 & @xmath871 & @xmath872 & @xmath873 & @xmath874 & @xmath875 & @xmath876 + qvw & @xmath877 & @xmath871 & @xmath872 & @xmath873 & @xmath874 & @xmath875 & @xmath876 + kaqvw & @xmath878 & @xmath879 & @xmath880 & @xmath881 & @xmath882 & @xmath883 & @xmath884 + qv@xmath885 & @xmath886 & @xmath887 & & & & & + vw@xmath885 & @xmath870 & @xmath888 & & & & & + qvw@xmath885 & @xmath880 & @xmath889 & & & & & + kaqvw@xmath885 & @xmath890 & @xmath891 & & & & & + ccccc kaqv & @xmath892 & @xmath892 & & + q & @xmath893 & @xmath894 & @xmath895 & @xmath896 + qv & @xmath897 & @xmath898 & @xmath899 & @xmath900 + qv+vv & & & @xmath901 & @xmath902 + v & @xmath903 & @xmath904 & @xmath905 & @xmath906 + qvw & @xmath907 & @xmath908 & @xmath909 & @xmath910 + kaqvw & @xmath911 & @xmath912 & @xmath913 & @xmath913 + | the wilkinson microwave anisotropy probe ( _ wmap _ ) has mapped the entire sky in five frequency bands between 23 and 94 ghz with polarization sensitive radiometers .
we present three - year full - sky maps of the polarization and analyze them for foreground emission and cosmological implications .
these observations open up a new window for understanding how the universe began and help set a foundation for future observations . _ wmap _ observes significant levels of polarized foreground emission due to both galactic synchrotron radiation and thermal dust emission .
synchrotron radiation is the dominant signal at @xmath0 and @xmath1 ghz , while thermal dust emission is evident at 94 ghz .
the least contaminated channel is at @xmath2 ghz .
we present a model of polarized foreground emission that captures the large angular scale characteristics of the microwave sky . after applying a galactic mask that cuts 25.7% of the sky ,
we show that the high galactic latitude rms polarized foreground emission , averaged over @xmath3 , ranges from @xmath4k at 22 ghz to @xmath5k at 61 ghz . by comparison
, the levels of intrinsic cmb polarization for a @xmath6cdm model with an optical depth of @xmath7 and assumed tensor to scalar ratio @xmath8 are @xmath9k for e - mode polarization and @xmath10k for b - mode polarization . to analyze the maps for cmb polarization at @xmath11 ,
we subtract a model of the foreground emission that is based primarily on a scaling _ wmap _ s 23 ghz map
. in the foreground corrected maps , we detect @xmath12 .
this is interpreted as the result of rescattering of the cmb by free electrons released during reionization at @xmath13 for a model with instantaneous reionization . by computing the likelihood of just the ee data as a function of @xmath14 we find @xmath15 .
when the same ee data are used in the full six parameter fit to all _ wmap _ data ( tt , te , ee ) , we find @xmath16 .
marginalization over the foreground subtraction affects this value by @xmath17 .
we see no evidence for b - modes , limiting them to @xmath18 .
we perform a template fit to the e - mode and b - mode data with an approximate model for the tensor scalar ratio .
we find that the limit from the polarization signals alone is @xmath19 ( 95% cl ) where @xmath20 is evaluated at @xmath21 .
this corresponds to a limit on the cosmic density of gravitational waves of @xmath22 . from the full _ wmap _ analysis
, we find @xmath23 ( 95% cl ) corresponding to a limit of @xmath24 ( 95% cl ) .
the limit on @xmath20 is approaching the upper bound of predictions for some of the simplest models of inflation , @xmath25 . |
a well - known open problem in fourier analysis is the bochner - riesz operator conjecture , which asserts the @xmath2-boundedness of the fourier multipliers @xmath3 on @xmath4 , so long as @xmath5 where @xmath6 and @xmath7 denotes the fourier transform in @xmath8 .
the problem is well understood in dimensions @xmath9 and @xmath10 ( see @xcite , @xcite , @xcite , @xcite ) .
but in higher dimensions , although there are several interesting results by many authors , it remains open .
its relevance is due , on the one hand , to the very natural question being asked , but also because of its close connection with some other basic objects , namely the so - called kakeya maximal function , the restriction properties of the fourier transform , or the covering properties satisfied by parallelepipeds in @xmath8 having arbitrary directions and eccentricities .
there is also the hope that obtaining deep understanding of the bochner - riesz operators could be a first step in the project of extending the classical caldern - zygmund theory of singular integrals , or pseudodifferential operators , going beyond kernels whose singularities are located only at the origin or at infinity , as is demanded in several areas of number theory or pdes . in the extreme case , @xmath11 , the multiplier @xmath12 is given by the indicator function of the unit ball . by a remarkable result of c. fefferman @xcite we know that it is bounded only in the obvious case @xmath13 , disproving the conjecture about the boundedness of @xmath14 in the range @xmath15 . in its proof fefferman made use of the properties of the kakeya sets in the plane ( for every @xmath16 there is a set whose measure is less than @xmath17 but containing a rectangle of dimensions @xmath18 on every direction ) , but also of a previous result due to y. meyer , who observed that the @xmath2-boundedness of @xmath14 implies a vector - valued control for hilbert transforms in different directions of the space . more concretely : let @xmath19 then @xmath14 bounded on @xmath4 implies @xmath20 ( throughout this paper the symbol @xmath21 means that there exists @xmath22 independent of @xmath23 and @xmath24 such that @xmath25 for all @xmath26 .
we will also write @xmath27 to denote that the involved constant depends upon the parameter @xmath28 ) .
it was precisely that vector - valued inequality , the one disproved by fefferman with the help of an appropriated kakeya set , which one may now describe as the main enemy of the bochner - riesz conjecture .
one way of defeating the enemy was introduced in @xcite ( see also @xcite for an independent proof ) : @xmath29 where the norm in @xmath0 is given by the integral @xmath30 with @xmath31 the uniform measure in the unit sphere and @xmath32 , @xmath33 are the polar coordinates in @xmath8 . throughout this paper the proof given in @xcite will be revisited , improving some of the arguments and estimates .
in particular we will show that meyer s lemma holds in the expected range @xmath15 so long as we substitute @xmath34 by @xmath0 . for every @xmath35 , singular integrals ( i.e. integral operators whose
kernels are given by @xmath36 , with @xmath37 ) yield bounded operators on @xmath0 .
however , it is somehow surprising to realize that directional hilbert transforms @xmath38 defined by are bounded on @xmath0 if and only if @xmath15 . to see that @xmath39 must necessarily lie in this range for the boundedness of @xmath38 it suffices to check the action of @xmath38 over the indicator function of a unit cube . however , the other implication is more involved and we will present two different proofs .
one of them is on the spirit of meyer s lemma and makes use of the main result in @xcite .
the second is based on properties of the _ universal kakeya maximal function _ : @xmath40 with the help of kakeya sets it is easy to see that @xmath41 can not be bounded on any space @xmath4 , @xmath42 .
however we will present a geometric argument to show that , acting on radial functions @xmath23 , we have the estimate @xmath43 this was first proved in @xcite . here
we will present a new proof using the old bush `` and brush '' methods of @xcite , @xcite and @xcite . throughout this paper several weighted estimates for singular integrals play a crucial role .
in particular the fact , discovered in @xcite , that for any @xmath44 , and any locally integrable function @xmath23 , we have that @xmath45 is a weight in the class @xmath46 such that @xmath47 where @xmath48 denotes the hardy - littlewood maximal function .
however , this should not be a big surprise , because the main reason to ask the question raised in @xcite ( circa 1973 ) was to produce a functional machine to reduce the bochner - riesz multiplier problem to the boundedness properties of the kakeya maximal function , that is to the covering properties of parallelepipeds in @xmath8 ( see @xcite , @xcite ) .
in fact the formula @xmath49 yields perhaps the most interesting examples of @xmath50 weights .
the other known cases are the radial powers @xmath51 , and in this paper we will make use of both classes . in particular we use that @xmath52 if and only if @xmath53 .
but we also present an argument showing how the theory for the weights @xmath51 can also be deduced from the estimate @xmath54 * acknowledgement .
* i am grateful to diego crdoba , fernando chamizo and keith rogers for their useful comments and help in the preparation of the manuscript .
consider the singular integral operator @xmath56 where @xmath57 is smooth enough and has zero mean value @xmath58 . [ singular ]
@xmath14 extends to a bounded operator from @xmath0 to itself when @xmath35 .
it suffices to prove the theorem when @xmath59 because then the case @xmath60 will follow by duality .
given @xmath61 and @xmath62 , let us consider @xmath63 by @xcite , we have @xmath64 for @xmath65 , where @xmath66 denotes the hardy - littlewood maximal operator applied to the function @xmath67 . but
@xmath68 is also radial .
therefore @xmath69 so long as @xmath70 , where @xmath71 . taking the supremum over all such @xmath24 so that @xmath72 , we obtain the proof of theorem [ singular ] . the theorem above is somehow the result that one may expect , but
the following is , perhaps , more surprising .
let us consider the hilbert transform @xmath38 in the direction of @xmath73 , defined in .
[ hbounded ] @xmath38 is bounded on @xmath55 if and only if @xmath15 . because the rotational symmetry , it suffices to show the conclusion for @xmath74 . to see the only
if " part we take @xmath23 to be the indicator function of the unit cube and observe that @xmath75 for @xmath76 , @xmath77 , @xmath78 . then an elementary computation yields the inequality @xmath79 therefore the boundedness of @xmath80 implies that @xmath81 .
again the other bound @xmath82 follows by duality . to prove the if part let us observe first that the @xmath55 norm is well - behaved under dilations and it is preserved under multiplication by characters @xmath83 : if @xmath84 then @xmath85 therefore if @xmath86 , where @xmath87 is the ball of radius @xmath88 , we have @xmath89 which produces the estimate @xmath90 uniformly on @xmath91 , so long as @xmath15 .
next we consider the multiplier , @xmath92 , corresponding to the hilbert transform @xmath38 and we observe that its boundedness properties are equivalent to those of the fourier multiplier operator associates to the indicator function of hyperplanes @xmath93 . given a function @xmath23 ( chosen so that its fourier transform @xmath94 has compact support ) the expression @xmath95 corresponds to the fourier multiplier given by the indicator function of the ball of radius @xmath96 centered at @xmath97 . choosing a convenient sequence of points @xmath97 and radius @xmath96 we can write @xmath98 then dominated convergence , together with the uniform bounds of @xmath99 in @xmath55 , @xmath15 , allows to finish the proof .
a different proof of theorem [ hbounded ] is the following : we start with the integral @xmath100 we have @xmath101 where @xmath102 denotes the one - dimensional hardy - littlewood maximal function in the direction of @xmath103 and @xmath44 .
as @xmath104 is radial one may consider its universal kakeya maximal function , defined in , to get the upper bound @xmath105 that is @xmath106 where @xmath71 .
to finish we make use of the fact that @xmath41 is bounded on @xmath107 , @xmath108 , when restricted to radial functions @xcite : observe that @xmath15 yields @xmath109 therefore given @xmath110 we choose @xmath44 so that @xmath111 to obtain @xmath112 which coincides with @xmath113 .
[ meyer ] given @xmath114 , a countable family of directions in @xmath8 , and the corresponding directional hilbert transforms , @xmath115 , the following inequality holds : @xmath116 where @xmath15 .
once more it is enough to prove it when @xmath117 and then use duality to cover the other cases . given @xmath62 we consider the integrals @xmath118 that is @xmath119 and @xmath120 and the proof follows for the same reasons given in theorem [ hbounded ] .
we finnish this section by presenting a new proof , using old techniques , of the boundedness of the general kakeya maximal function acting on radial functions .
related with these problems are maximal functions associated to vector fields in @xmath8 . given a continuous field of directions , @xmath121 , and a positive valued real function @xmath122 , let us consider @xmath123 as it is usual in differentiation theory , the behavior of those maximal operators produces quantitative versions of the lebesgue s differentation theorem . in general
the problem is rather difficult but e. stein and s. wainger @xcite , and also several other authors @xcite , have created a theory which yields sufficient conditions for boundedness .
however , for our discussion the following two examples are of special relevance : 1 .
@xmath124 , @xmath125 . in this case
@xmath48 is bounded on @xmath4 if and only if @xmath126 .
1 . @xmath124 , @xmath127
. then @xmath48 is of weak - type ( 1,1 ) .
2 . in @xmath128 , @xmath129 , @xmath125 . then @xmath48 is bounded on @xmath130 , @xmath35 .
the proof of * 1 ) * uses polar coordinates together with the fact that @xmath131 is an @xmath50 weight in @xmath8 for @xmath108 : @xmath132 so long as @xmath108
. the proof of * 1 * ) * is left as an exercise to the reader .
regarding * 2 ) * let us consider in @xmath128 identified with @xmath133 , the change of variables @xmath134 mapping the straight line @xmath135 , @xmath136 , into the curve @xmath137 .
but @xmath138 is the result of translating the fixed curve @xmath139 to the point @xmath140 .
therefore , in the @xmath141-plane , the maximal function is realized as @xmath142 where @xmath143 . in this
setting we can apply the results of e. stein and s. wainger @xcite to finish the proof . the result for the vector field * 2 ) * can be extended to higher dimensions if we restrict our attention to radial functions , because , in that case , we can take any direction perpendicular to the position vector @xmath26 without changing the value of the maximal function there .
that is , considering the product structure @xmath144 the evaluation of the maximal function at the point @xmath145 can be obtained as the evaluation of its @xmath146-dimensional version acting on the slice @xmath147 .
then an induction argument together with fubini s theorem gives the boundedness for every @xmath148 .
our proof for @xmath41 will be then modeled upon these two extreme examples * 1 ) * and * 2 ) * , and it will make use of the geometry of parallelepipeds in @xmath8 to treat the intermediate case : in @xmath8 , @xmath149 , the universal kakeya maximal function is unbounded on every @xmath4 , @xmath42 , as an appropriated kakeya set argument easily shows . however , acting on radial functions @xmath23 we have [ urad ] @xmath150 if @xmath108 and @xmath23 is radial . fixing a radial function @xmath23 and a positive number @xmath151
let us consider the set @xmath152 given @xmath153 there is a set of directions @xmath154 so that @xmath155 for every @xmath156 . clearly if @xmath157 then @xmath158 and @xmath159 coincide after a rotation .
furthermore , for a given @xmath26 and @xmath160 we will choose @xmath161 satisfying the inequality above and such that its sum @xmath162 is the biggest possible ; then among all such @xmath73 , @xmath163 and @xmath164 , we select those maximizing the projection of the segment @xmath165 into the radial direction . taking advantage of the radial symmetry and after some elementary geometrical considerations , the set @xmath166 is contained in an almost disjoint union of annuli : @xmath167 satisfying the following properties : 1 . for each annulus
@xmath168 there exists a direction @xmath169 so that the average of @xmath170 on the segments of directions @xmath169 , starting at some point of the sphere @xmath171 and ending on the sphere @xmath172 , is the given value @xmath173 .
we have that @xmath174 if @xmath175 .
clearly , in order to estimate the size of @xmath176 it is enough to control , independently , the portion of that set contained in each annulus @xmath168 .
we shall distinguish two classes of annuli : 1 .
thin shells : @xmath177 .
thick shells : @xmath178 . in both cases ( thin or thick )
we fix two poles @xmath179 and @xmath180 ( north and south ) in order to estimate the portion of @xmath166 near the equator . given the radial symmetry we can select the direction @xmath103 in such a way that the ray from each point of the shell meets the @xmath181 axis , but we can also establish the convention that , among the two possible rays , the chosen one is pointing north . for a thick annulus @xmath22 its volume is comparable to @xmath182 . therefore one needs to show the inequality @xmath183 let us observe that for each point in the sphere @xmath184 we are then given a straight line segment @xmath185 starting at @xmath26 ; ending at the inner sphere @xmath186 ; tangent to a certain sphere @xmath187 and such that 1 .
the average of @xmath170 on @xmath185 is @xmath173 .
2 . the straight line @xmath185 intersects the @xmath181 axis . again , given the fixed radial function @xmath23 one can enlarge the segment @xmath185 ( with some small positive number @xmath188 ) to become a tube of radius @xmath188 .
in such a way that the average of @xmath170 in that tube is bigger than , says , @xmath189 .
next we cover @xmath166 ( or a fixed portion of it close to the equator ) with a family of those tubes but keeping them pairwise disjoint at the outer sphere @xmath184 .
that is , we have obtained what is called a _ brush configuration _ of tubes , meeting the @xmath181 axis and being disjoint at the outer sphere of the shell .
let us denote by @xmath190 the collection of those tubes and let us consider their overlapping function @xmath191 .
then some elementary geometric considerations ( brush argument ) yields the following estimate : @xmath192 therefore @xmath193 so long as @xmath194 ( equivalently @xmath108 ) allowing us to finish the proof .
the elementary geometrical considerations are the following : \1 ) for each @xmath195 in the equator @xmath196 of @xmath197 , let us consider the two dimensional plane @xmath198 determined by @xmath195 and the @xmath181 axis , and also the enlarged band @xmath199 .
choosing conveniently @xmath188-spaced points @xmath114 in @xmath196 , we may assume that the sets @xmath200 are pairwise disjoint and their union cover @xmath201 .
then their overlapping where @xmath202 is easily controlled : at distance @xmath203 we have @xmath204 then the collection of tubes @xmath205 is divided into disjoint classes by the inclusion relation @xmath206 .
\2 ) the tubes inside @xmath207 meets at the tangential inner sphere @xmath208 producing a collection of bush configurations ( that is sets of tubes meeting at a common point in @xmath209 ) such that each tube inside @xmath210 belong to , at most , two of those bushes .
then their overlapping function is easily controlled by an elementary calculation ( bush argument ) : @xmath211 when @xmath212 .
those two estimates together imply that in the sphere @xmath213 , @xmath203 the overlapping is bounded by @xmath214 uniformly in @xmath215 .
that is , we get . to treat thin shells we proceed in the same manner , but here
the estimate is reduced easily to each plane @xmath198 .
that is , we only have to consider the two dimensional case and observe that the @xmath188-rectangle @xmath216 arranged in disjoint families of bushes satisfying the overlapping estimate @xmath217 which implies @xmath218
with the same notation used in the introduction , let us consider @xmath219 for rapidly decreasing smooth functions @xmath23 , where @xmath220 is the unit ball in @xmath8 , @xmath149 .
[ disc ] the operator @xmath14 is bounded on @xmath0 if and only if @xmath221 , the proof of this theorem was first given in @xcite ( see also @xcite for an independent proof and @xcite for a weighted version ) . here
we will improve and simplify our previous presentation in order to motivate subsequent results .
the only if part follows easily taking @xmath23 to be the inverse fourier transform of a @xmath222-function , @xmath94 , such that @xmath223 when @xmath224 and @xmath225 if @xmath226 . then @xmath227 which belongs to @xmath0 if and only if @xmath81 . the other bound @xmath82 follows by duality .
given @xmath228 ( the schwartz class of rapidly decreasing smooth functions ) it has a development @xmath229 where the @xmath230 are defined on @xmath231 and @xmath232 is an orthonormal basis of the @xmath233-dimensional space of spherical harmonic polynomials of degree @xmath234 in @xmath8 .
here , the index @xmath234 takes nonnegative integer values and , for each @xmath234 , @xmath235 takes values in the interval @xmath236 .
it is a well - known fact ( see @xcite ) that the fourier transform preserves that development : @xmath237 where @xmath238 here @xmath239 denotes the bessel function of order @xmath240 , which evaluated at a non - negative real number @xmath26 is given by the integral @xmath241 ( @xcite , page 176 , formula 4 ) .
continuing with the proof of the theorem , we note that @xmath242 where @xmath243 are integral operators given by the formula @xmath244 and @xmath245 let us begin decoding the kernels ( see also ref .
@xcite ) . consider the ode verified by the bessel functions @xmath246 which written in terms of the functions @xmath247 becomes @xmath248 then @xmath249 therefore @xmath250 thus we have obtained four families of the integral operators @xmath251 however , taking into account the asymmptotics of bessel functions ( see @xcite ) : @xmath252 where @xmath253 , one may infer heuristically that the kernels @xmath254 and @xmath255 correspond to hilbert transforms , while the cases @xmath256 and @xmath257 , are less singular and produce hardy integral operators . since the order @xmath258 of the bessel functions is increasing , the asymptotic estimate above does not produce uniform bounds , and the more precise analysis of the following lemma is needed in the critical range @xmath259 .
[ vdc ] the following estimates hold uniformly on @xmath260 .
@xmath261 as an indication for the reader we sketch the proof of iii ) .
the stationary phase method , van der corput estimates , and integration by parts allow us to control the size of @xmath262 in the critical interval @xmath263 : we have @xmath264 where @xmath265 .
the critical point corresponds to @xmath266 i.e. @xmath267 .
then @xmath268 therefore stationary phase method yields @xmath269 because @xmath270 , @xmath271 .
the other cases follow by similar arguments ( see ref .
@xcite ) . [ prodj ] in the critical range @xmath259 , uniformly in @xmath258 ,
the following estimate hold : @xmath272 nevertheless , we fortunately have the combination @xmath273 allowing us to make use of the extra decay hidden there to achieve the estimates .
[ kj ] given @xmath274 , there exists a finite constant @xmath275 such that @xmath276 for @xmath277 , any function @xmath278 and every sequence of rapidly decreasing smooth functions @xmath279
. we will consider the operator @xmath280 with kernel @xmath281 the estimate for @xmath282 will follow as a result of duality ; the cases @xmath283 and @xmath284 being less singular are easier to handle and the details will be left to the reader .
fixing @xmath285 , we consider the partition @xmath286 and the corresponding splitting of the kernel .
let us denote by @xmath287 the integral operator with the kernel @xmath288 we consider four cases .
* case 1 : * @xmath289 or @xmath290 .
the cases @xmath289 are the easier , because in @xmath291 both @xmath292 and @xmath293 are uniformly bounded , and the estimate in the proposition is reduced to a well - known inequality for the hilbert transform .
similarly when @xmath11 , @xmath294 or @xmath295 , @xmath296 ; we have that @xmath297 uniformly in @xmath258 ( lemma [ vdc ] ) . * case 2 : * @xmath295 , @xmath298 .
let us consider the partition of the interval @xmath299 through the sets @xmath300 for @xmath301 $ ] .
we have @xmath302 where @xmath303 uniformly in @xmath258 .
* case 3 : * @xmath304 , @xmath294 .
since @xmath305 uniformly in @xmath258 when @xmath306 , the part of the estimate corresponding to that region trivializes . for the remainder
we have , with the notation of the previous case , @xmath307 where as before @xmath308 denotes the hardy - littlewood maximal function .
putting all the estimates together for the cases 1 , 2 and 3 , we get @xmath309 * case 4 : * @xmath295 , @xmath294
. it will be convenient to modify the splitting and for a fixed @xmath258 let us define @xmath310 where @xmath311 $ ] , and @xmath312 .
define also @xmath313 then we have the kernels @xmath314 taking into account once more the estimate of lemma [ vdc ] , it is obvious that the most dangerous is the @xmath315 term . therefore , in the following , we shall present the details of the proof corresponding to that case , leaving the other three as exercises . to estimate the action of the operator with kernel @xmath316 on a function @xmath317 at a point @xmath318
, we divide the integration in three parts : @xmath319 first we estimate @xmath320 .
we have @xmath321 that gives @xmath322 where @xmath28 and @xmath323 are conjugate hlder exponents .
observe that if @xmath324 then @xmath325 and we have @xmath326 and therefore @xmath327 next we estimate @xmath220 : @xmath328 that gives @xmath329 where we have applied several times the estimates of lemma [ vdc ] . finally one needs to control the diagonal terms @xmath22 : @xmath330 we have @xmath331 where @xmath332 uniformly in @xmath258 and @xmath234 . that is , for each @xmath333 , we have obtained the pointwise estimate @xmath334 for @xmath335 . where the functions @xmath336 , @xmath337 are uniformly bounded .
therefore , we can finish the proof of proposition [ kj ] with the help of some well - known estimates for the hilbert transform and the hardy - littlewood maximal functions ( see @xcite , @xcite , @xcite ) , namely @xmath338 so long as the weight @xmath103 belongs to the class @xmath50 , that is @xmath339 furthermore the constant @xmath340 above depends only upon @xmath39 , @xmath28 and @xmath341 .
let us note that in order to prove the theorem it is enough to show the following inequality : @xmath342 so long as @xmath15 . but
this will be a consequence of proposition [ kj ] together with some observations about weights in the class @xmath50 .
since it happens that @xmath343 if and only if @xmath344 , we can conclude that @xmath345 if and only if @xmath15 .
furthermore , if @xmath346 we have that @xmath347 , for every @xmath348 .
however in the interval @xmath349 we have that @xmath350 under the hypothesis @xmath351 , @xmath352 .
the part of the equality which corresponds to the critical intervals : @xmath353 is a direct consequence of proposition [ kj ] for the following reasons : @xmath354 and the result follows taking @xmath355 . the remainder terms can be controlled similarly , except for one of them which needs extra arguments , namely @xmath356 the case @xmath117 is easy because then @xmath357 for @xmath358 and we have the estimate @xmath359 to treat the case @xmath349 let us observe that for @xmath360 we have : @xmath361 then we get the integral @xmath362 and since @xmath363 it happens that @xmath364 allowing us to obtain the estimate needed to finish the proof : @xmath365 with similar methods to those developed above , one can obtain the following restriction theorem due to l. vega @xcite .
[ restriction ] in @xmath8 , @xmath149 , given @xmath366 there exists a finite constant @xmath367 such that @xmath368 for every rapidly decreasing smooth function @xmath23 .
note that the interval @xmath366 is bigger than @xmath369 , which corresponds to the stein - tomas restriction theorem ( @xcite , @xcite ) , @xmath370 the proof of the restriction estimate above will be obtained by duality from the corresponding extension estimate @xmath371 taking @xmath372 it is easy to check that the estimate above can not hold when @xmath373 , which amounts to show that the range @xmath366 is sharp in the statement of the theorem .
given @xmath374 we have an expansion @xmath375 where @xmath376 is a spherical harmonic of degree @xmath234 , normalized as @xmath377 .
then we can invoke the formula ( @xcite , @xcite ) @xmath378 as we have mentioned , any @xmath374 can be written in the form @xmath375 where @xmath376 is a normalized harmonic polynomial of degree @xmath234 and by it suffices to prove the following inequality : @xmath379 for every @xmath380 . since @xmath381
when @xmath382 , the contribution of that interval to the integral above trivializes .
let us then consider for a fixed @xmath383 ( @xmath384 ) the integral @xmath385 as above but restricted to @xmath386 .
we shall first consider the case @xmath10 .
then we have @xmath387 to estimate @xmath388 we observe that @xmath389 uniformly in @xmath234 when @xmath390 .
therefore @xmath391 regarding @xmath392 , we have @xmath393 and @xmath394 which yields @xmath395
finally let us write @xmath396 where @xmath397\ ] ] is an interval in the real line and @xmath398 $ ] .
let us also define @xmath399 then we have @xmath400 and @xmath401 where we have taken @xmath402 and used the fact that @xmath403 .
the term @xmath404 is controlled by the same argument . therefore , when @xmath10 , @xmath403 , adding all the estimates above over the dyadic intervals @xmath405 $ ] , @xmath383 , we get @xmath406 and the theorem is proved in this case . in the general case
@xmath149 , we have to estimate @xmath407 that with the previous argument is @xmath408 and it is enough to note that the resulting exponent in the powers of 2 is negative for @xmath380 .
throughout this paper we have taken advantage of the fact that @xmath409 is a weight in the class @xmath46 , for every integrable function @xmath103 , whose @xmath50 bounds are estimated independently of @xmath103 . this property was discovered ( see @xcite ) with the disc multiplier problem in mind , as an efficient manner of relating functionally the boundedness properties of singular integrals to maximal functions involving different directions of @xmath8 . to our knowledge , the family @xmath410 is the more extended class of known weights ; the other known examples given by powers @xmath51 which have also played an important role in several proofs of this paper .
it is then interesting to realize that those power weights @xmath51 can also be considered particular cases of the construction @xmath410 . in the following
we will present the details in dimension one , leaving the general case as an exercise .
let @xmath412 be a cube containing the point @xmath26 and let us denote by @xmath413 the maximal operator restricted to subcubes of @xmath414 ( the double of @xmath412 ) .
it is well - known that the mapping @xmath415 is bounded from @xmath416 to @xmath417 , @xmath418 .
we have the splitting @xmath419 where @xmath420 i.e. in @xmath421 , @xmath422 for some universal constant @xmath22 .
therefore @xmath423 which is @xmath424 . for @xmath430
, we use the characterization of @xmath50 in terms of the boundedness of the maximal function or the hilbert transform .
we have @xmath431 where the supremum is taken over all @xmath24 such that @xmath432 but we also have @xmath433 this quantity is bounded by @xmath434 then we observe @xmath435 therefore , using the previous case , is @xmath436 and we are done .
a. carbery , e. hernndez and f. soria , the behaviour on radial functions of maximal operators along arbitrary directions and the kakeya maximal operator , tohoku math . j. ( 2 ) * 41 * ( 1989 ) , no . 4 , 647656 .
a. crdoba , maximal functions , covering lemmas and fourier multipliers , in _ harmonic analysis in euclidean spaces ( proc .
pure math .
, williams coll . ,
williamstown , mass . , 1978 ) , part 1 _ , 2950 , proc .
pure math . , xxxv , part amer .
soc . , providence , ri . | several estimates for singular integrals , maximal functions and the spherical summation operator are given in the spaces @xmath0 , @xmath1 .
singular integrals , disk multiplier , fourier restriction theorems 42b20 , 42b25 |
despite its minuteness , the exchange of momentum by light - matter interaction is responsible for observable phenomena ranging from the formation of comet tails to the trapping @xcite and cooling @xcite of small particles down to single atoms . in particular
, the exchange of angular momentum produces mechanical torques . in a pioneering experiment
, beth showed that optical spin can induce rotation in birefringent plates illuminated by circularly polarized ( cp ) light @xcite .
this effect has been subsequently confirmed on the macro- @xcite , micro- @xcite , and nanoscales @xcite .
light can also carry orbital angular momentum , which has been widely used to produce vortex - like motion of microparticles @xcite . additionally , chiral particles undergo rotation even when illuminated by unpolarized plane waves @xcite ( i.e. , light without net angular momentum ) .
the interaction of light with rotating particles raises fundamental questions , such as the possibility of cooling the rotational degrees of freedom down to the quantum regime . on the opposite side ,
the material response of particles rotating at extreme velocities can be largely influenced by spinning forces , eventually leading to centrifugal explosion .
even the kinematical change between lab and rotating frames is known to produce frequency shifts in the light emitted by rotating particles @xcite .
these effects may be occurring in cosmic dust irradiated by polarized light over enormous periods of time .
particle trapping in vacuum @xcite may provide a suitable framework to study these phenomena . in this letter , we study the electromagnetic torque and scattering properties of rotating particles suject to external illumination . like in rotational raman scattering @xcite , the particle produces inelastic scattering at frequencies separated from the incoming light by twice the rotation frequency .
we report two remarkable observations at large rotation velocities : ( i ) inelastic scattering is stronger than elastic scattering ; and ( ii ) net amplification of the incident light takes place via stimulated emission in this purely classical linear system .
we base this conclusions on analytical expressions derived for the torque , the absorbed power , and the inelastic light scattering cross section , from which a complex resonant interplay between the light frequency and the particle rotation frequency is observed .
feasible experimental conditions for the observation of light amplification are discussed . around a transversal direction ( @xmath0 axis ) .
an incident light plane wave of frequency @xmath1 illuminates the rod along a direction normal to its axis , thus producing an electromagnetic torque @xmath2 and scattered light of frequencies @xmath1 and @xmath3 . for left circularly polarized ( lcp ) light , as considered in the figure , only the @xmath4 inelastic component is emitted.,width=226 ]
for simplicity , we consider a spinning rod , the optical response of which can be modeled with an effective spring consisting of a point particle of mass @xmath5 and charge @xmath6 oscillating around a fixed charge @xmath7 with natural frequency @xmath8 and damping rate @xmath9 .
the point - particle motion is constrained to the longitudinal direction of the rod @xmath10 , which is rotating with frequency @xmath11 around a direction @xmath12 perpendicular to @xmath10 ( see fig .
[ fig1 ] ) , so that the instantaneous position of this charge can be written @xmath13 , where @xmath14 is the distance to the central charge .
the force equation of motion is then given by ( see appendix ) @xmath15 where the terms on the right - hand side are ( from left to right ) the restoring force , the intrinsic friction , the external electric - field force , the abraham - lorentz force accounting for radiative damping proportional to @xmath16 @xcite , and the particle reaction force @xmath17 taken to cancel other azimuthal components in order to constrain the motion of the point charge along the rotating @xmath10 axis .
we assume the charge velocity to be small compared to the speed of light , so that the magnetic component of the force can be overlooked , although it could lead to a small precession .
the transversal polarization of the rod is also considered to be negligible .
the resulting torque @xmath2 is directed along @xmath12 and originates in the reaction force according to @xmath18 . for monochromatic incident light of frequency @xmath1 and electric field
@xmath19 , we can solve eq .
( [ force ] ) analytically to find the time - averaged torque ( see appendix ) @xmath20 \frac{|e_+|^2}{|d_+|^2}\nonumber\\ & + \left[\gamma\omega_-+\tau(\omega-2\omega)^3\right ] \frac{|e_-|^2}{|d_-|^2}\bigg\ } , \label{m}\end{aligned}\ ] ] where @xmath21 , @xmath22 , and @xmath23 .
the @xmath9 terms in eq .
( [ m ] ) originate in photon absorption by the particle , whereas the @xmath24 terms describe a radiative reaction torque similar to what happens in rotating dipoles @xcite .
the torque predicted by eq . ( [ m ] ) presents resonant features signaled by the zeros of @xmath25 . in small absorbing particles , for which the @xmath24 terms can be neglected compared to the @xmath9 absorption terms
, this condition reduces to the ellipses @xmath26 as is clear in fig .
[ fig2](a ) for left cp ( lcp ) incident light , corresponding to the lower sign in eq . ( [ cond ] ) .
the figure is actually representing the cross section towards mechanical work , given by @xmath27 , where @xmath28 is the external light intensity and @xmath29 . for light oscillating at the same frequency as the rotation
, the particle sees a frozen incident field , and the torque is therefore zero ( it vanishes along the @xmath30 line ) . for slower rotation ( @xmath31 )
the torque is positive , pointing to a net transfer of momentum from the light field to the particle .
in contrast , the torque and @xmath32 become negative for faster rotation even when the light and the particle rotate in the same direction ( @xmath33 ) .
this suggests that the incident light is actually braking the particle by effectively producing stimulated photon emission with contributions proportional to the ohmic and radiative losses ( terms in @xmath9 and @xmath24 , respectively ) .
the absorption and elastic - scattering cross sections admit the closed - form expressions ( see appendix ) @xmath34 and @xmath35 interestingly , the rotational motion produces inelastic scattering components at frequencies @xmath3 @xcite and responding to the cross sections ( see appendix ) @xmath36 the consistency of this model is corroborated by the fact that the sum of all partial cross sections ( @xmath37 ) equals the total extinction cross section derived from the optical theorem ( see appendix )
. the partial cross sections are also resonant under the condition ( [ cond ] ) ( see fig .
[ fig2 ] ) . in particular ,
light absorption [ fig .
[ fig2](b ) ] produces a positive transfer of intrinsic angular momentum from each absorbed photon to the particle ( @xmath38 ) , while elastic scattering [ fig .
[ fig2](b ) ] dominates the @xmath39 region . for right cp light ,
we obtain similar results , with the ellipsoids of @xmath2 , @xmath40 , and @xmath41 reflected with respect to the @xmath42 axis ( not shown ) . for linear polarization
, we find a superposition of the two orthogonal circular polarizations ( see appendix ) .
two remarkable effects emerge at large rotation velocities ( @xmath43 ) : ( i ) the inelastic emission ( at frequency @xmath44 for lcp light ) becomes a leading process ; and ( ii ) like the mechanical cross section , the total extinction cross section can be negative ( e.g. , for @xmath33 and lcp light ) , thus confirming a net stimulated light emission , whereby mechanical motion is converted into photons ( see below ) .
the model parameters @xmath45 , @xmath8 , and @xmath9 can be easily adjusted to fit the polarizability of a particle at rest ( @xmath46 ) . computing the induced dipole @xmath47 from our model ( see appendix ) ,
we find @xmath48 , where @xmath49 is the polarizability . in the absence of internal friction ( @xmath50 )
, this expression satisfies the property @xmath51 , as expected from the optical theorem for non - absorbing particles @xcite .
this polarizability has the same form as that of a sphere of radius @xmath52 described by the drude dielectric function @xmath53 , which allows us to identify @xmath54 and @xmath55 .
incidentally , at low @xmath11 and within the dipole approximation , we find @xmath56 from an analysis based upon the maxwell stress tensor @xcite for the torque produced by lcp light on a rod under the conditions of fig .
[ fig1 ] , where @xmath57 and @xmath58 are the polarizabilities along directions parallel and perpendicular to the rod , respectively .
this torque vanishes for non - absorbing spheres in virtue of the optical theorem .
in contrast , non - absorbing rods ( @xmath59 ) experience a net torque because the term in @xmath60 can not compensate the first one .
this indicates that the torque acting on small lossy spheres , in which absorption ( the @xmath9 term ) governs @xmath61 , can be approximated with our formalism as twice the torque acting on a rod , but using the above identification of model parameters .
just to give a better idea of the order of magnitude of the mechanical and absorption cross sections for a drude sphere , the units in the color scale of fig .
[ fig2](a)-(b ) are @xmath62 , with typical values of @xmath63 m in noble metals .
we show in fig . [ fig3 ] the partial cross sections of rotating rods in the limits of high and low ohmic losses .
the negative extinction represents an increase in the amplitude of the incident beam after interaction with the rotating particle .
incidentally , our recently reported rotational friction @xcite can be understood as the spontaneous - emission counterpart of the stimulated emission under discussion . stimulated emission takes place at @xmath64 by transferring mechanical energy from the particle to the incident beam ( coherently added photons ) .
this allows us to speculate with the possibility of constructing a laser ( see appendix ) in which an incident cp light beam is exponentially building up as it encounters rotating particles along its path ( the particles are externally driven , for example , by pumping cp light of higher frequency ) .
this scheme is robust because it is insensitive to finite distributions of particle size and rotation velocity , as long as the latter exceeds the light frequency . ) and * ( b ) * non - dissipative ( @xmath65 ) rods rotating with frequency @xmath66 .
the shaded areas for negative extinction ( @xmath67 ) give the excess of photons added to the incident beam as a result of the interaction with the particle .
we take @xmath68.,width=321 ]
for constant incident light intensity , light absorption leads to heating of the particle until it reaches an equilibrium temperature @xmath69 above the temperature of the surrounding vacuum @xmath70 .
this equilibrium is established when the absorbed power @xmath71 equals the radiative cooling rate @xmath72 . in the drude approximation for a metal sphere , and neglecting the effect of @xmath11 , which only enters at large velocities , we have @xcite @xmath73 , which leads to @xmath74 .
this is represented in fig .
[ fig4](b ) for a 10 nm carbon nanotube ( upper solid curve ) for which the polarizability has been obtained in the discrete - dipole approximation @xcite .
there are two different regimes in the dependence of @xmath69 on light intensity : at low @xmath75 , the particle is nearly at the vacuum temperature , but when the intensity increases , the particle heats up , asymptotically approaching a @xmath76 power law . a similar behavior is observed for a gold nanoparticle [ fig .
[ fig4](b ) , lower solid curve ] , incorporating the effect of magnetic polarization and a realistic metal permittivity taken from optical data @xcite .
10-nm - long single - wall carbon nanotube ( cnt ) to a rotational velocity @xmath77mhz as a function of incident cp light intensity . *
( b ) * particle equilibrium temperature .
the particles are rotating in vacuum at an external temperature @xmath78k .
the gold and carbon melting points are indicated by dashed lines in ( b ) .
the light frequency is @xmath79hz ( wavelength @xmath80m).,width=264 ]
the time needed to accelerate a carbon nanotube and a gold nanoparticle to 1mhz is represented in fig . [
fig4](a ) as a function of lcp light intensity , as calculated with the torque of eq . ( [ m ] ) .
this time scales approximately as @xmath81 under the condition @xmath82 , and it plunges well under 0.01s for light intensities below the melting threshold .
this suggests the possibility of achieving extreme rotation velocities in optically trapped nanoparticles , and eventually producing centrifugal explosion , thus introducing an unprecedented physical scenario ( this should happen close to the point at which the centrifugal energy reaches the surface - tension energy , which in a 20-nm liquid gold particle occurs at @xmath83ghz ) .
the present self - consistent oscillator model permits capturing the optical response of particles rotating under extremely high velocities .
this defines a new scenario plagued with exotic phenomena such as strong inelastic scattering , and most notably , the possibility of realizing a laser ( see appendix ) running on a classical linear system by exploiting our prediction of light amplification stimulated by particles rotating faster than the light frequency .
there are several avenues towards potentially practical implementations of these systems , such as , for example , ( 1 ) a diluted gas of linear molecules driven to thz rotational frequencies by cp light and acting as the active medium of a laser operating in that demanded frequency range ; ( 2 ) an extension of these ideas to acoustic lasing based on a similar classical interaction with subwavelength particles rotating faster than the sound frequency ; ( 3 ) microwave waveguide setups such as those used to monitor rotational frequency shifts @xcite ; ( 4 ) dust clouds in cosmic environments exposed to polarized light , in which radio wave amplification might be taking place .
this work has been supported by the spanish micinn ( mat2010 - 14885 and consolider nanolight.es ) and the european commission ( fp7-ict-2009 - 4 - 248855-n4e ) .
a.m. acknowledge financial support through fpu from the spanish me .
rotates with frequency @xmath11 around a static charge @xmath7 .
the moving charge oscillates along the rod axis.,width=302 ] we model a rotating rod by a point charge @xmath6 oscillating around a fixed charge @xmath7 according to the equation of motion @xmath15 where @xmath84 is the external electric field , the dots indicate differentiation with respect to the time , @xmath85 and @xmath5 are the coordinate vector and the mass of the moving charge , respectively , @xmath8 and @xmath9 are the natural frequency and the intrinsic damping rate of the oscillator , @xmath86 gives the coupling to radiation via an abraham - lorentz force @xcite , and @xmath87 is the reaction of the nanoparticle that constraints the point charge to rotate with fixed frequency @xmath11 around the @xmath12 axis .
the point charge moves along the rod axis @xmath10 , which also rotates with frequecy @xmath11 around @xmath12 .
the main elements of the model are graphically illustrated in fig .
[ figsi1 ] .
the moving charge coordinate vector can thus be written as @xmath88 the radial and azimuthal unit vectors depend on time as @xmath89 and @xmath90 , respectively . inserting eq .
( [ r ] ) into eq .
( [ force ] ) and noticing that @xmath91 and @xmath92 , the radial equation of motion reduces to @xmath93 whereas the azimuthal equation is trivially satisfied by our choice of @xmath17 .
we now write an external monochromatic electric field as @xmath19 , which permits us to obtain for the inhomogeneous solution of eq .
( [ forcer ] ) the expression @xmath94 where @xmath95 @xmath96 and @xmath97 notice that the natural oscillation frequency @xmath8 is split into two shifted frequencies by the effect of a finite @xmath11 .
the torque acting on the particle is directed along @xmath12 and produced by the reaction force of eq .
( [ freact ] ) according to @xmath18 . inserting eqs .
( [ freact ] ) and ( [ rho ] ) into this expression and retaining only time - independent terms ( i.e. , the time - averaged contribution ) , we find @xmath98 , \label{m}\end{aligned}\ ] ] where @xmath99 and @xmath100 .
equation ( [ m ] ) here is the same as eq .
( 2 ) of the main paper .
the light exerts a mechanical torque on the particle resulting in a power transfer given by @xmath101 .
the particle is also capable of dissipating energy via light absorption by the material of which it is made .
the total ( mechanical plus absorption ) work exerted by the external light on the particle per unit time has to include the effect of radiation damping , and it is therefore given by @xmath102 .
alternatively , we can write it as the power due to the dissipative force , @xmath103 . using either one of these expressions , the time - averaged absorption power is found to reduce to @xmath104 it is reassuring to observe that our model predicts the absorption power to be proportional to the intrinsic friction rate @xmath9 .
this result can be divided by the external light intensity @xmath28 to obtain the absorption cross section @xmath105 where @xmath29 .
it is interesting to analyze the dipole induced in the rotating particle .
we can write it as @xmath106\xx \nonumber\\&&\;\;\;\;\;-\ii\left[\left(\frac{e_+}{d_+}-\frac{e_-}{d_-}\right)\ee^{-\ii\omega t}-\frac{e_+}{d_+}\ee^{-\ii\omega_{++ } t}+\frac{e_-}{d_-}\ee^{-\ii\omega_{-- } t}\right]\yy\bigg\}+c.c .
\label{dipole}\end{aligned}\ ] ] in the rest frame .
this expression contains terms of frequencies @xmath1 , @xmath107 , and @xmath4 .
the elastic cross section corresponds to the field re - radiated by the component of frequency @xmath1 , @xmath108 as obtained from the time - averaged integral of the far - field poynting vector ( that is , the integral of the squared far - field amplitude ) divided by the external field intensity @xmath75 .
likewise , we can define inelastic cross sections corresponding to the emission of light with frequencies @xmath3 , @xmath109 the quantities @xmath110 describe the effective area of the particles to inelastically scatter the incident light .
these cross sections are the same as eqs .
( 3)-(5 ) of the main paper .
numerical results are given there for left - handed circularly polarized incident light ( @xmath111 and @xmath112 ) and also here in fig .
[ figsi2 ] for linearly polarized light ( @xmath113 ) . for the particle at rest ( @xmath46 ) , with the electric field and
the rod both oriented along @xmath114 , the induced dipole of eq .
( [ dipole ] ) reduces to @xmath115 , where @xmath116 .
the polarizability of the static particle is then given by @xmath117 .
interestingly , this expression yields @xmath118 , in agreement with the optical theorem @xcite , which predicts @xmath119 for a non - absorbing particle ( @xmath50 ) . the static polarizability permits establishing a connection between our simple oscillator model and actual physical parameters of real nanoparticles . for example , for a metallic ellipsoid of volume @xmath120 described by a drude dielectric function of bulk plasmon frequency @xmath121 and damping rate @xmath9 , the electrostatic polarizability , corrected for radiative losses @xcite , reduces to the above expression with @xmath122 and @xmath123 , where @xmath124 is the depolarization factor that depends on the rod aspect ratio @xcite .
the optical theorem allows us to use the @xmath1 component of the induced dipole @xmath125 [ eq . ( [ dipole ] ) ] to obtain the total extinction cross section @xmath126\frac{|e_+/e|^2}{|d_+|^2}+\omega_-\left[\gamma+\tau(\omega_-^2 + 3\omega^2)\right]\frac{|e_-/e|^2}{|d_-|^2}\right ) , \nonumber\end{aligned}\ ] ] where @xmath127 denotes the time average .
furthermore , the mechanical work on the rotational motion can be expressed as the cross section @xmath128 , \nonumber\end{aligned}\ ] ] which is directly obtained from eq .
( [ m ] ) .
then , the total balance of power must be zero , and this leads to the condition @xmath129 which is indeed satisfied by eqs .
( [ sigmaabs ] ) and ( [ sigmaw])-([sigmamech ] ) .
incidentally , @xmath130 can be negative , for example when @xmath33 and @xmath131 ( left circularly polarized light ) , pointing to a net stimulated emission of light from the rotating particle .
in fig . [ figsi2 ] , we provide the equivalent of fig . 2 of the main paper , but for linearly polarized incident light instead of circularly polarized light .
the right and left circularly polarized components of the incident light produce inelastic emission at frequencies @xmath107 and @xmath4 , respectively , as shown in fig .
[ figsi2](d)-(e ) .
the torque and absorbed power are the superposition of the contributions originating from these two components .
interestingly , these terms add up in the absorption but they cancel each other in the torque at @xmath46 .
it is clear from eq .
( [ sigmatot ] ) that the extinction cross section can take negative values when the light and the particle rotate in the same direction ( e.g. , for @xmath131 and @xmath132 ) , provided the particle rotates faster than the light frequency ( @xmath133 ) .
this is illustrated in fig .
3 of the main paper , which shows that a negative cross section is possible due to either absorption or radiative processes , depending on the magnitude of @xmath9 relative to @xmath134 .
a negative extinction indicates that more light is leaving the system than it impinges on it .
that is , light amplification takes place due to the interaction with the rotating particle , and the extra light is coherent with the incident one .
these are necessary conditions for making a laser .
an electromagnetic amplifier could be constructed based on a medium filled with rotating particles .
a light beam with the right characteristics would be amplified at every encounter with one of the particles . as a result , the beam could be exponentially amplified as it propagates along this medium .
particle rotation can be sustained by external means ( e.g. , either mechanically for radio waves or via optical pumping at higher frequencies ) , so that the system acts as a transducer that converts the externally supplied energy into coherent radiation . in a more sophisticated arrangement
, we envision an electromagnetic cavity containing such an amplifying medium , in which the radiation is making several passes before leaving the cavity ( e.g. , in a parallel mirror configuration ) . the seed in this system can be provided either by external illumination or by intrinsic spontaneous emission within the system , which has been recently studied in detail @xcite .
this lasing and light amplification effects can be potentially achievable within various windows of the electromagnetic spectrum .
for example , one can rely on mechanically or magnetically induced motion up to radio frequencies @xmath135 mhz .
light - driven rotation of nanoparticles levitated in vacuum could be used for higher frequencies in the ghz regime before the particles break apart .
molecular rotators could also be employed up to thz frequencies , in a process resembling the rotational raman effect .
these regimes can be accompanied by intrinsic resonance frequencies of the rotating bodies relying on lc circuits at low frequencies or physically equivalent excitations such as mie resonances , plasmons , and optical phonons .
finally , dust clouds in cosmic environments provide a plausible scenario for electromagnetic amplification because they are formed by microparticles that are exposed to the effect of neighboring light sources ( e.g. , stars and galaxies ) , which are circularly polarized in some cases .
this illumination acts over enormous periods of time ( millions of years ) , thus potentially inducing large rotation in the particles in the range of ghz before they are destroyed by centrifugal explosion .
therefore , amplification can be expected up to ghz frequencies , compatible with the low - end of the cosmic microwave background . | the rotational dynamics of particles subject to external illumination is found to produce light amplification and inelastic scattering at high rotation velocities .
light emission at frequencies shifted with respect to the incident light by twice the rotation frequency dominates over elastic scattering within a wide range of light and rotation frequencies .
remarkably , net amplification of the incident light is produced in this classical linear system via stimulated emission .
large optically - induced acceleration rates are predicted in vacuum accompanied by moderate heating of the particle , thus supporting the possibility of observing these effects under extreme rotation conditions . |
the use of images of the earth in different wavelengths from infrared cameras on board of geostationary satellites is a powerful tool for the identification of new potential sites for astronomical observations in optical , infrared and submillimeter wavelengths ( erasmus & sarazin 2002 ; cavazzani et al . 2012 ; cavazzani & zitelli 2013 ) .
sites in remote areas with complex relief and scarse or null meteorological records can be identified by previous computation of the clear sky fraction ( csf ) .
as the cloud covering is a function of the height of the atmosphere , images from satellite are useful due to the detectors on board measure the emissivity of the atmosphere at different layers .
several studies for the determination of the csf are found in the literature .
soden & bretherton ( 1993 ) conducted an analysis of upper tropospheric humidity and clouds using goes observations , finding a correlation between humidity and cloud cover at high altitudes .
erasmus & sarazin ( 2002 ) were among the first to sucesfully use long - term satellite data for csf computations for a wide range of altitudes .
erasmus & van rooyen ( 2006 ) and della valle et al . ( 2010 ) reproduced with high precision the csf records of ground data at la palma with a model based on satellite data .
cavazzani et al .
( 2011,2012 ) found percentages of clear nights at five astronomical observatories around the world using long - term goes data .
hidayat et al .
( 2012 ) used 15 years of goes infrared images in 6.7 and 10.7@xmath2 m bands to compute the percentage of clear sky over a wide region in indonesia .
recently , marn et al . ( 2015 ) computed the csf at high altitude sites by using a new clearance algorithm based on global meteorological models .
csf values were then used for precipitable water vapour estimations at two astronomical sites in the north of chile . even though most of the methods are based on the comparison of radiances , fluxes or temperatures at each pixel of the satellite image with threshold values previously defined
there is no standard methodology for csf computations .
cavazzani & zitelli ( 2013 ) used monthly thresholds defined as the maximum radiance at clear nights along the month .
thresholds may also be defined by using an atmospheric model @xcite . in this study
, we introduce a novel methodology involving the use of monthly thresholds of temperature at different altitudes from the surface up to the upper troposphere .
our thresholds definitions were based on the altitude of the terrain , long - term records of weather stations and radiosonde data archives .
the thresholds definitions are based on monthly mean records of both , air temperature at the site and vertical temperature profile between 8 and 10 km of altitude which make more precise the predictions for clear covering than those involving atmospheric models for thresholds computations .
our methodology was applied in northern south america over the tropical region covering colombia , western venezuela and northern ecuador .
even though the zone is far from being interesting for infrared and submilimeter astronomy due to the requirement of excellent weather conditions , many universities and institutes from those countries are planning to develope optical observatories .
the countries are located within the intertropical convergence zone ( itcz ) showing complex topography and with a climate governed by a bimodal annual cycle which results of the double passage of the itcz for the region ( poveda et al .
the andes affect the climate by blocking zonalflows , influencing regional wind patterns and precipitation and hosting diverse micro - climates .
the main goal of this study is to seek prospective locations with middle quality csf values over the andes of those countries .
the paper is organized as follows . in section 2
we describe the satellite data used in this study .
threshold temperatures calculations are presented in sections 3 and 4 .
the methodology and validation are discussed in sections 5 and 6 while in sections 7 and 8 we present the results and the selected sites .
the geostationary operational enviromental satellites goes 12 and 13 are equiped with instrumentation capable of capture images in several infrared bands . for our purposes
we concentrate our attention on b3 and b4 bands centered at 6.7 and 10.7@xmath2 m , respectively .
images in b4 are strongly affected by emission and absorption of water vapour found at middle altitudes ( @xmath34 km ) whereas layers of the upper trophosphere between 8 and 10 km are sensed by images in the b3 band .
other bands as those centered in 13.3@xmath2 m provide information about fog and very low atmosphere characteristics .
since the data extends for decades these satellites are commonly used for long - term studies .
temporal domaine for this study begins on the 1th of january 2008 and ends on the 31st of december 2012 with a few missing data lower than 7% in the worst cases . goes missing data has different causes : technical troubles in goes-12 ( from december 15th 2008 to january 6th 2009 ) ; when goes-13 replaced goes-12 ( from april 14th to 28th 2010 ) ; and due to anomalies in the operational system ( from september 23th to october 18th 2012 ) .
we assumed the clear sky fraction as null for the days with goes missing data .
goes 12 - 13 images are freely available through an electronic library of noaa in format gvar ( goes variable format ) .
we selected a region of the image to download covering a big part of colombia , the western part of venezuela and the north of ecuador .
for each night we used available observed time at the three different hours 00:15:15 , 06:45:13 and 09:45:14 utc in order to cover the entire night .
gvar is an instrumental radiance given in counts .
we used the library of _ idl _ routines provided by the national climatic data center ( ncdc ) for converting the counts to radiance at each pixel .
usual calibration of the images was done using the scaling bias and gain values available from ncdc website in order to get the radiance in @xmath4 $ ] .
brightness temperatures @xmath5 at the instants @xmath6 of the night are computed from calibrated radiance values by assuming black body emission .
conversion to actual temperatures is done using the relation @xmath7 where @xmath8 and @xmath9 are conversion coefficients dependent on each band ( weinreb et al .
1997 , rossow et al .
1985 ) . reported values available from ncdc website , for goes 12 satellite are @xmath101.01 and @xmath11k for the b3 band and @xmath101.00 and @xmath12k for the b4 band .
the differences between the actual and brightness values increase with decreasing temperature but usually are lower than 4% .
for instance , a brightness temperature of @xmath13k corresponds to temperatures of @xmath14k and @xmath15k for the b3 and b4 bands , respectively . for most of cases
those differences are negligible for the csf computations .
we indicate actual temperatures in b3 and b4 bands as @xmath16 and @xmath17 , respectively .
the spatial resolution of goes images is 4@xmath184 km in both bands and thus a single pixel subtends only a small field of view from the ground compared with the total field of view available for an observer at ground . as the validation of our method involves comparison with log - books of the _ observatorio nacional de llano del hato _ in venezuela ( hereafter , cid ) we degraded the original resolution in order to make more real comparisons .
other authors have used similar or even lower resolutions .
cavazzani and zitelli ( 2012 ) computed goes temperatures for each site over a @xmath19 sub - matrix which corresponds to a linear projection of about 100@xmath18100 km centered on the coordinates observatories at ground .
we used a delanuay interpolation of actual temperatures @xmath16 and @xmath17 on a grid of 51@xmath20 pixels obtaining a resolution of 28@xmath1818 km which we consider wide enough for validation with ground data .
detection of mid - altitude clouds at a height of 4000 meters above a pixel location of the image in b4 band and at a fixed instant during the night requires the previous evaluation of a threshold or reference temperature . in this study
we compute monthly thresholds of temperature based on long - term records of weather stations located inside the region covered by the goes b4 and b3 images .
+ we used long - term records of meteorological data obtained from the _ instituto de hidrologa , meteorologa y estudios ambientales _ , hereafter ideam .
the ideam data archive comprise periods from 7 to 40 years of monthly mean values of temperature and precipitation for a set of weather stations in colombia .
the complete archive contains 400 records at altitudes between 400 and 4700 meters above sea level covering 25% of the country .
the archive presents a high number of gaps , especially those of temperature .
we considered only long - term records of temperature with a maximum of one monthly absence obtaining the reduced number of 63 weather stations irregularly distributed in a region comprising latitudes between 3@xmath1 and 7@xmath1 north , and longitudes from -76@xmath1 up to -72@xmath1 west .
we will refer to this region as the control region defined by the spatial distribution of the ideam long - term historical records .
the longest records of these selected stations have more than 40 yr as is shown in the right panel of figure [ f1 ] .
locations are widely distributed in different altitudes with a high number of them close around bogot ( bog ) ( @xmath21=2600 meters ) as shown in the left panel of the same figure .
+ the average of monthly mean records at each station for periods spanning at least 7 years is expected to be normally distributed around a central value @xmath22 being very likely that the monthly mean values are within @xmath23 of the distribution , where @xmath24 is the standard deviation .
the coldest temperature at each station may be then defined as @xmath25 .
however , an additional correction term must be applied since the measurements do not represent values during the night .
we recall that the ideam archive consists only of mean temperatures thus we must correct them for nightime observations . the monthly threshold temperature at each station may therefore computed as : @xmath26 \label{e1}\ ] ] where @xmath27 is the trend of the minima records of temperature given by short - term records of weather stations indicated with a solid line in figure [ f2 ] and represents the correction for night - time observations .
we note that these thresholds correlate very well with the altitude of the station as illustrated in the same figure suggesting that monthly threshold temperatures may be completely determined from the altitude of the stations . from linear fits for each month
we established monthly relationships between threshold and altitude of the form : @xmath28 where @xmath29 and @xmath30 are constants for each month with values indicated in table [ t1 ] .
slight differences along the year are observed for these two parameters with slope or lapse rates variations of the order of @xmath31 @xmath1c and negligible variations in the intercepts .
+ thresholds in the control region at locations where no weather stations are available require interpolation of the altitudes .
we used the geostatistical method called kriging , which is a spatial interpolation method that in its simplest form allows to compute altitudes at locations with no data , based on the spatial autocorrelation .
this means that two close stations have similar values of altitude whilst two stations farther apart do not . briefly ,
the height at the location @xmath32 refered as @xmath33 is given by a weighted combination of the ideam all records of altitude : @xmath34 where @xmath35 is the vector of kriging weights @xmath36 and @xmath21 is the vector of available samples of altitudes .
our sample is defined by the altitudes of all available records of weather stations .
the system of equations ( [ dos ] ) is equivalent to the matrix form @xmath37 where @xmath38 represents the matrix of covariances between all sampled values and @xmath39 the vector of covariances between sampled points and the point in which the value is assessed .
the vector of covariance is obtained from the semivariogram @xmath40 defined as : @xmath41 \label{tres}\ ] ] where @xmath42 represents the unit distance or lag from which semivariance is determined .
the two semivariograms indicated in figure [ f3 ] were computed for lags up to 200 km along two mutual perpendicular directions .
both semivariograms present similar spatial variances and are well described by a spherical model ( dashed line ) with a nugget of @xmath43 , sill of @xmath44 and range of @xmath45 .
the spherical model was found as the most suitable for modeling the covariances ( [ tres ] ) and for solving the weights in the system of equations ( [ dos ] ) by using gsl - gnu numerical libraries for matrix inversion .
a comparison of the interpolated heights with real altitudes from all - term ideam weather stations is shown in the figure [ f4 ] .
cross validation of the kriging interpolator was done by hiding each ideam data point during the computation of the weights and then comparing the prediction with the real value .
differences with an ideal kriging interpolation ( solid line ) are within 3@xmath46 which we consider enough good for our purposes . based on kriging results for altitude interpolation in the control region , we computed the monthly thresholds using the values appearing in the table [ t1 ] .
results are very similar for all the months with small differences due to the bimodal annual cycle .
threshold temperatures for january with resolution of 28@xmath1818 km are shown in the panel b ) of figure [ f4 ] .
the coldest region ( 0 - 2)@xmath1c corresponds to the _ nevado del huila _ the highest snow peak in the central branch of the andes . thresholds between ( 2 - 7)@xmath1c are observed in the eastern andes comprising _ villa de leyva _ ( vdl ) , _ pisba _ ( pis ) and _ macaravita _ ( mac ) .
warm regions with thresholds above 20@xmath1c characterize the valley of _ magdalena _ river between the central and eastern branches of the andes and the southestern regions on zero altitude zones . for comparison purposes ,
we show in panel a ) the thresholds obtained using delaunay triangulation of the 63 ideam long - term monthly records .
we note that the spatial domaine is constrained in this case and defined by the triangulation of the 63 irregularly distributed long - term ideam weather stations .
differences between both interpolations are significant mostly in regions surrounded by strong topographic gradients in which a better estimation of the threshold temperatures using kriging is expected .
this is particularly noticed in the poor ideam sampled region centered on pis where a maximum difference of @xmath47c is observed .
other regions with low number of weather stations are the south of _ bucaramanga _ ( buc ) and the south - east of _ medelln _ ( med ) . on these regions
the combination of the lack of weather stations and strong gradients in altitude make the delanuay interpolation unrealistic . mean differences of @xmath48c over the eastern andes are atribuited mainly to strong gradients as the region is well sampled by the ideam data archive .
thresholds obtained from altitude are more precise than those obtained by interpolation of weather station records and are used as a reference for cloud detection in the b4 band .
uncertainties in the threshold temperature determination are computed using the sigma error of the kriging interpolation .
the largest errors up to 6@xmath1c are observed over the sea level locations and into the valleys while errors over the andes are always below 1.5@xmath1c as shown in panel c ) of figure [ f4 ] .
for the control region the thresholds are computed with an uncertainty of @xmath49c . with the aim of computing monthly threshold temperatures over a wider region including the _ observatorio nacional de llano del hato _ where log - books of cloud cover fraction are available
, we apply monthly threshold - altitude relations over what we denoted as extended region covering latitudes between 0 and 13n and longitudes between 70 and 78w .
a set of @xmath50 altitudes over this region were obtained from _ google - maps _ and interpolated using the methodology discussed above .
monthly threshold temperatures between @xmath51 and @xmath52c are observed across caribbean coastal regions and in the valleys of the rivers magdalena and cauca in panels of figure [ f5 ] . in the proximity to the andes ,
thresholds get colder specially in the highest mountains indicated with purple in the figure such as the four highests locations over the colombian and venezuelan andes highlands with threshold temperatures below @xmath53c .
uncertainties in threshold temperatures over the extended region are assumed to be equal to those obtained in the control region .
interpolated threshold temperatures covering the surface of the extended region are then used in our algorithm for cloud detection in the b4 band .
for a fixed instant , if the temperature obtained from the goes image at 10.7 @xmath2 m is colder than @xmath54 then the presence of a cloud above the location is inferred .
clouds along the upper troposphere are detected following a similar schema , defining first a threshold temperature for high altitudes where the weighting functions of b3 shows the maximum value .
we denote the altitude of the atmosphere measured from the ground with @xmath55 , different to the altitude of a weather station @xmath21 .
we recall that the weighting functions depending on both , the wavelength and @xmath55 , determine the layer from which the emitted radiation originates . for the b3 band ,
the median height values of the weighting functions are in the interval @xmath56 km of altitude above the ground with a maximum at 9 km ( cavazzani et al.2011 ) .
threshold temperatures at upper troposphere may be obtained using different approaches .
hydayat et al .
( 2012 ) considered the reference atmospheric model of kneizys et al .
( 1996 ) which provides a vertical tropospheric profile of temperature for latitudes within @xmath57 and adopted a threshold equal to @xmath58c at the observatory location based on a @xmath59 analysis with data of sky visual inspection .
our approach is based on the use of three thresholds each one at heights of @xmath60 , @xmath61 and @xmath62 km obtained from radiosonde upper - air measurements .
we make a strong emphasis on the reliability of the temperature at upper troposphere .
particularly , we are interested in finding trustworthy threshold values between 8 and 10 km .
data were obtained twice a day from the freely available database of the university of wyoming which provides information of temperature , humidity , pressure and mixing ratio as a function of the altitude for two stations located inside the extended region , _
80222 skbo _
( 74@xmath190``5w , 4@xmath1420''n , 2546 m ) in central andes and _ 78988 tncc _
( 68@xmath15736``w , 12@xmath11160 '' , 54 m ) in the caribbean coast .
upper - air temperatures of both stations exhibit similar patterns as a function of the altitude .
around 60 radiosonde records were used per month adding data from both stations .
altitude resolution is variable with typical values ranging from 200 up to 600 m. we compute monthly mean temperatures by considering measurements made twice daily ( 00 and 12 utc ) during the years 2008 - 12 at each @xmath55 using linear interpolation along a uniform vertical grid of 21 points between @xmath60 and @xmath62 km . in order to follow the same approach than the one used for ground thresholds , we define the threshold temperature at the altitude @xmath55 as @xmath63 where @xmath64 is the monthly mean
value and @xmath65 is the standard deviation .
monthly thresholds as a function of the altitude are displayed in figure [ f6 ] .
monthly standard deviations change with altitude and time showing yearly variations up to @xmath49c in the worst of cases . in the tropics ,
the temperature anomalies at high altitudes is enhanced by the presence of moderate size clouds usually associated with shallow convection responsible for anomalies up to @xmath66c at altitudes from 5 up to 14 km ( folkins 2008 ) .
this tropospheric warming has been largely observed and we suggest that is the main source of the sigma variations . the further assumption that these thresholds describe the upper troposphere of the whole extended region in many aspects is a strong hypothesis , however the area of study is affected by atmospheric phenomena including the westerly low level _
choc _ jet coming from the pacific coasts of colombia ( poveda & mesa 2000 ) and the caribbean trade winds which leads to similar daily patterns of high altitude clouds .
middle and right panels in figure [ f7 ] show a typical evolution during the night of 8th of april 2009 .
we see two images in b3 taken at the beginning and at the end of the night .
high altitude clouds shown in green color in the left panel coming from the pacific ocean are blocked by the andes system leaving clear zones in the andes of venezuela and over the amazon zones .
we note that the fact of extrapolating thresholds over all the extended region , the possible changes in the instrumentation and the differences between night and daytime observations may affect the observed temperatures and therefore introduce considerable noise in the computed thresholds . however , the monthly thresholds obtained with these two stations are in agreement with the predictions of the tropical atmospheric model of kneizys ( 1996 ) indicated with the solid line in figure [ f6 ] .
we implemented an algorithm for percentage of clear sky calculation during the night through a comparison pixel to pixel of goes 12/13 interpolated temperatures with monthly threshold values obtained from long - term records of weather stations and radiosonde databases .
we constructed four monthly threshold images , one for band b4 labeled as @xmath54 and three for b3 labeled as @xmath67 each one at heights @xmath68 and @xmath62 km .
daily comparisons were done in both bands at 00:15:15 , 06:45:13 and 09:45:14 utc with similar approach to those described in hidayat et al .
( 2012 ) and erasmus & sarazin ( 2002 ) .
first we used daily radiances in the b4 band to compute brightness and actual temperatures .
if the satellite derived temperature is colder than the monthly threshold at b4 the presence of middle high clouds above the pixel image at the instant @xmath69 is inferred : @xmath70 where @xmath54 is a monthly constant depending only on the altitude of the terrain .
if a pixel is classified as covered in b4 it will be also covered in the b3 band . on the contrary , if the pixel is clear in b4 we still have to verify the presence of high altitude cirrus clouds by doing a further and similar analysis using the b3 band .
the pixel is classified as clear or covered in b3 at the altitude @xmath55 in the following way : @xmath71 where @xmath64 denotes the monthly mean temperature of the upper troposphere at height @xmath55 and @xmath72 the standard deviation at that height .
we compute three values of clear sky fraction at the instant @xmath69 and at each altitude @xmath55 using inequalities ( 4 ) and ( 5 ) by counting the number of clears or covered pixels .
the total clear sky fraction at altitude @xmath55 is then obtained from the average of the three instants @xmath69 along the night .
one value of csf is computed in this way at each altitude .
the nights in the extended region are classified as spectroscopic or photometric at each altitude @xmath55 based on the values of total clear sky fraction .
we define the quality of the night , at the height @xmath55 , as spectroscopic if the total clear sky fraction is larger than @xmath73 . as we computed the total csf with three values , spectroscopic nights are those with @xmath74 of the night free of clouds at least .
sites that fulfill this requirement are more suitable for spectroscopic observations than for photometric ones as long as spectra can be acquired quickly in comparison with photometeric data .
we note that this definition is based only on the amount of clear sky and does not involve any sky quality indicator .
spectroscopic night simulations ( hereafter , sns ) provide monthly averages of clear sky fraction at three different altitudes for each pixel of the image satisfying the condition that csf is larger than 33% .
nights with photometric quality on the other hand , are those with total sky fraction larger than 67% corresponding to @xmath75 of clear sky during the night at least . if this condition is fulfilled in a particular site , we argue than at the altitude @xmath55 will be possible to conduct photometric observations .
photometric night simulations ( hereafter , pns ) provide monthly maps of clear sky fraction at three different altitudes for each pixel of the image with the condition that csf is larger than 67% .
based on total csf values obtained with sns and pns outputs , monthly averages at altitude @xmath55 were computed .
cloud cover along the upper - troposphere is sensitive to the altitude , being larger at low altitudes mostly at @xmath76 km , indicating that high cirrus does not affect considerably the net cloud fraction .
we used the monthly csf averages at @xmath76 , @xmath61 and @xmath770 km to compute a weigthed mean for the three altitudes at each instant during the night .
weights were computed through a comparison of the amount of clear sky fraction for photometric nights coming from our model with those obtained from daily number of observing hours reported by log - books of cid or site 2 in venezuela .
site 2 is placed at 3600 meters height and enclosed by higher mountains as _ pico bolvar _ ( 4978 m ) at less than 40 km away .
as the resolution is 28 km in longitude and 18 km in latitude , the threshold temperature at site 2 is highly affected by the altitude of the closest higher mountains giving colder thresholds than expected .
using pns instead of sns makes more precise validation because pns output is closer to the condition of total clear night .
figure [ f8 ] corresponds to the clear hours above the reference site 2 or cid as reported in the log - book . despite the bias present in the log - books and the missing data for the years 2009 - 10 ,
due mainly to bad weather , we did not find considerable differences between observed and expected monthly clear sky fraction values when different weights were used .
therefore , for the purposes of this study , we consider sufficient to use equal weights for the three altitudes . the validation of our code is limited due to factors such as the lack of completeness of daily log - books , tecnhical issues , high humidity , strong winds and uncertainties of our model .
we used all the information available in the log books as the general description of the night done by the astronomer to compute the amount of daily clear sky fraction for each month of the years 2008 - 12 at site 2 .
comparison with our monthly results using pns is displayed in figure [ f9 ] .
the solid line represents no differences between predicted and measured monthly clear sky fraction values .
there is a large spread during the considered years , especially for clear sky fraction values between 20 - 60% , but in general terms they are in reasonable agreement .
monthly percentages of clear sky for the years 2008 - 12 over the extended region were obtained using the method described above .
the main common feature is that monthly averages display maxima at the same locations along the five years .
maximum monthly percentages of clear sky are observed during two periods in the years 2008 - 12 , one from december to february and another from june to july according with the bimodal annual cycle of precipitation ( poveda et al .
the year 2012 stands out as the best year displaying the highest percentages during both periods .
we take the year 2012 as representative of the monthly distributions of clear sky . as illustrated in figure [ f10 ] our monthly averages of csf for sns nights exhibit differences along the year with the highest csf values always located over a wide region in the central venezuelan andes with a maximum csf@xmath7880% at site 1 during both periods . for site 2 or cid corresponding to _
observatorio nacional de llano del hato _ we get a csf mean value of 58.5% , lower in comparison with site 1 during both periods . we have found two regions in colombia with similar night qualities to those observed at site 2 . the northern part of the _ sierra nevada de santamarta _ close to _ nabusimake _ ( nab ) and the central _ serrana del perij _ toward the east of _ codazzi _ ( cod ) .
percentages at these two regions are in the range of 80 - 90% during the first period , 50 - 70% during the second and below 30% from march to may and between august and october .
sites over this region contain the best places in colombia found in this study .
somewhat low quality places are found over the north - eastern andes of colombia , over mountain systems along buc , mac and pis with clear sky fraction between 30 - 50% during both periods of 2012 .
southern andes regions in the _ nudo de los pastos _ , especially those in northern ecuador display csf values below 30% from november to may and @xmath370% from june to october .
very low percentages of clear sky fraction below 20% characterize central andes of colombia over the regions comprising _ popayn _ ( pop ) and bog with the exception of the region centered at the high site _ nevado del huila _
( 75@xmath15512``w , 2@xmath15136''n ) at 3700 meters with a mean of 35% between december - february and june - july and 33% for the other months .
regarding monthly averages of csf for pns nights , the same regions were found but with low percentages in comparison as shown in figure [ f11 ] .
the absence of spectroscopic nights in central colombia from may to october confirms that november - february and june - july are marked by the largest csf percentages while march - may and august - november are characterized by lower csf values .
we also observe such behaviour for other years .
the existence of this bimodal behaviour with maxima during the dry season and minima during the rainy season is atributed , as we metioned already , to the double passage of the intertropical convergence zone for the extended region .
therefore we conducted yearly sns and pns simulations , bringing together the months december , january , february , june , july and august into one group that we call _ dry season _ and march , april , may , september , october and november into another group denoted as _
rainy season_. sns percentages for both seasons show maxima at the same locations than those observed for the year 2012 with seasonal differences of @xmath320% with the exception of low latitude locations close to the equator where slightly smaller differences are observed .
percentages for pns simulations result 10% lower in comparison to sns nights but with identical csf distribution over the extended region .
annual clear sky fraction averages obtained from our sns simulations during rainy seasons for the years 2009 and 2012 are larger than for the other years .
it seems to correlate with the changes in the oceanic nio index ( oni ) reported by the climate prediction center noaa .
changes from positive to negative anomalies in april 2010 associated to a _ la nia _ event increase the number and intensity of tropical storms in central colombian andes thus decreasing the clear sky percentages .
the _ nudo de los pastos _ and the _ sierra nevada de santamarta _ are not considerably affected by this _ el nio southern oscillation _ ( enso ) .
for selection purposes we give preference to the simulations for spectroscopic nights during the dry seasons . we have selected twelve sites based on five year averages during the dry seasons that can be suitable for astronomical observations by selecting locations with csf@xmath7830% .
factors such as accesibility , security , service , local humidity , atomospheric _ seeing _ were not considered and we need additional studies to confirm these locations as potential for astronomical sites .
these twelve selected sites are distributed mostly along the following four zones : the central andes of venezuela comprising the states of _ mrida _ and _ trujillo _ , the north east of the _ sierra nevada de santamarta _ around sites nab and cod , the north eastern andes of colombia in the proximity of _ el espino _ ( esp ) and a wide region on the southern andes in _ nudo de los pastos _ containing two locations in ecuador .
we show in figure [ f16 ] the contours of clear sky fraction for these selected zones during the dry season with a spatial pixel resolution of 28@xmath1818 km .
even though the selection criteria involves only spectroscopic nights during the dry seasons , we give a total number of clear nights per year at selected sites by adding the csf five year averages for each season and then multiplying by a factor of 1.827 .
tables [ tbl-2 ] and [ tbl-3 ] contain properties of the selected sites indicating the identifier , the nearest populated location , geographic coordinates of the site , altitude , yearly five year averages for the dry and rainy seasons , five year average of clear sky fraction , standard deviation and the total number of clear nights per year .
the best places are situated over an extense area of the central andes of venezuela comprising latitudes from 8@xmath148n to 9@xmath111n and longitudes between 70@xmath117w and 70@xmath142w .
the region contains a number of potential sites at high altitudes , in particular site 1 ( 70@xmath12848``w , 9@xmath1560''n ) where we obtain the maximum csf in the extended region .
site 1 has an altitude of 3480 meters and it is located @xmath335 km north - east from _ timotes _ ( tim ) very close to the protected area _
parque nacional general cruz carrillo_. five year average of clear sky fraction for spectroscopic nights during the dry seasons for this site is ( [email protected])% corresponding to 122.3 clear nights per year .
the percentage during the rainy seasons is ( [email protected])% leading to 97.7 clear nights per year .
subsequently we report a total number of ( 220@xmath030 ) spectroscopic nights per year at site 1 . regarding our reference site 2 or
cid , five year average of spectroscopic nights during dry seasons is ( [email protected])% while for rainy seasons it is ( [email protected])% giving a mean of 72.4% corresponding to ( 132@xmath025 ) total clear nights per year .
five year average for photometric nights at this site gives ( [email protected])% for dry seasons and
( [email protected])% for rainy seasons giving a total of 43% which is in agreement with the log - books of the observatory as shown in figure [ f9 ] even though seasonal differences were not considered for the validation of the algorithm as discussed in section 6 .
+ this is a large system of mountains in the north of colombia isolated from the andes and located just 52 km from the caribbean coast .
it rises up to 5700 meters above sea level over a protected area of 17000 km@xmath79 containing the national park _ sierra nevada de santamarta_. five year average of csf is ( 42@xmath010)% over a broad region of 40 km to the north of nab , specifically at site 3 ( 73@xmath13112``w , 10@xmath15512''n ) placed at an altitude of 3600 meters with 76.8 clear nights per year during the dry season and 49.3 nights per year during the rainy season .
the total spectroscopic clear nights at site 3 is ( 126@xmath034 ) in contrast with ( 72@xmath016 ) clear nights per year with photometric quality .
the steep gradient in elevation toward the south - east direction of nab make the percentage of csf lower along the plain regions on the valley of river _
cesar _ which separates the _ sierra nevada de santamarta _ from the _ serrana del perij _ , an extension of the eastern andean branch on the boundary between colombia and venezuela .
slightly lower percentages than those observed at site 3 are obtained at site 4 less than 10 km from the location cod .
the site has ( 111@xmath027 ) clear nights per year with spectroscopic quality and ( 66@xmath018 ) photometric nights .
this broad territory hosts a high number of _ pramos _ , dry forests , valleys and high snow peaks most ot them protected areas such as national parks _ el cocuy _ ( coc ) and _ pisba _ ( pis ) .
the selected locations are within the wide region between the _ sierra nevada del cocuy _ and the north part of the canyon of the river _
chicamocha_. the highlands of the sierra reach an elevation of more than 5000 meters including several snow peaks .
site 6 placed @xmath350 km west from these high snow peaks , is the center of a region with five year average for the dry seasons of ( 35@xmath04)% .
clear sky percentage toward the north west along the canyon reaches the maximum value of ( [email protected])% at site 5 located in _
cuchilla las torrecitas _
, next to _ cordillera de hoya grande _ in _ norte de santander _ at an altitude of 4060 meters . for rainy seasons
we obtain ( [email protected])% at this site giving a total of ( 107@xmath016 ) clear nights per year .
site 7 located 20 km north - west of pis with an altitude of 3600 meters belongs to the protected area _
parque nacional pramo de pisba_. five year average is ( [email protected])% whereas for rainy seasons site 7 has 25% giving a total of ( 99@xmath016 ) clear nights per year .
another selected site in this region is site 8 , the closest selected location to the snow peaks , with ( 90@xmath012 ) clear nights per year using only four years since the 2012 csf average is below our detection threshold .
percentages of clear sky fraction for photometric nights for sites 7 and 8 are indicated in table [ tbl-3 ] . lower quality sites with csf@xmath8015%
are clearly observed over the central branch of the andes comprising locations _ murillo _ ( mur ) and _ medellin _ ( med ) .
brieva ( 1985 ) , using cloud cover data from visual inspection of the sky reports between 130 and 250 useful nights per year at site _ otn _ ( 4@xmath14633``n , 75@xmath12453''w ) , located 25 km in the south west direction from mur .
we obtain ( 89@xmath011 ) clear nights per year at site mur ( 4@xmath15919``,75@xmath11012 '' ) .
this region consists of tropical humid forests where the andes splits into the western and central branches .
sites 9 and 11 placed in north of ecuador stand out as the best places in the region with ( 105@xmath0 36 ) and ( 101@xmath0 42 ) clear nights per year , respectively . with ( [email protected])% below our selection limit we include the site 12 _ laguna de la cocha _ ( lco ) located at the colombian side of the _ nudo de los pastos_. minimum seasonal differences are found at the sites in this region .
comparatively with the other sites , the percentage of clear sky during dry and rainy seasons is very similar in sites within the _ nudo de los pastos _ confirming that the sky cover does not change significantly from one season to another at locations near the equator .
we note from tables [ tbl-2 ] and [ tbl-3 ] that the minimum differences between csf values for dry and rainy seasons for all twelve sites are observed especially for the sites 9 , 11 and 12 with differences of 4 , 2 and 1% .
five year average for the selected sites exhibit an extended minima in 2010 - 11 during rainy seasons as shown in figure [ f17 ] .
we found a possible correlation to _
la nia _ episodes since positive values for the southern oscillation index ( soi ) are observed for these two years .
interior plots in figure [ f17 ] show soi values obtained from noaa website for 2008 - 12 .
the symbols for the months go along with those for the sites .
positive values mostly for 2008 , 2010 and 2011 are associated to intensification of storms over the andes and then to low csf values as observed .
in this work we describe a method to compute the clear sky fraction from goes data based on comparisons with long - term records of temperature from weather stations and radiosondes , over a wide region of northern south - america comprising sites in colombia , venezuela and ecuador .
we conducted a temporal analysis based on three daily values along five years doing a careful comparison of goes12/13 temperatures in bands b3 and b4 with thresholds at different altitudes .
this approach enabled us to carefully determine the percentage of clear sky simultaneosly over the whole region .
similar perentages were obtained for all locations along the five years confirming the reliability of our methodology .
threshold monthly temperatures as a function of the altitude of the terrain were obtained for a small region of control through kriging interpolation of long - term records of temperature at weather stations .
monthly linear correlations between thresholds and altitude were found and extrapolated to an extended region including _ observatorio nacional de llano del hato_. threshold monthly temperatures at 8 , 9 and 10 km along the upper atmosphere were also computed based on daily radiosonde data of two stations located at sea level and at the central andes .
the method differs from other cloud fraction algorithms that uses long - term records of temperature to obtain the csf .
our approach gives more precise csf values as long as we have used reliable reference temperatures at the surface and at the high troposphere .
furthermore , our model can be easily extended to other tropical regions of astronomical interest .
nights are classified as spectroscopic or photometric based on the three csf values during the night .
the total number of spectroscopic and photometric nights per year at a fixed site are computed separately from five year averages for dry and rainy seasons .
the best location found in this study is site 1 ( 70@xmath12848``w , 9@xmath1560''n ) in the central venezuelan andes with ( 220@xmath042 ) spectroscopic clear nights per year and ( 157@xmath024 ) photometric clear nights per year .
site 1 is placed at an altitude of 3480 meters and 35 km north - east from _ timotes _ within a protected area of venezuela .
two lower quality locations were found in _
sierra nevada de santamarta _ , site 3 ( 73@xmath13112``w , 10@xmath15512''n ) and site 4 ( 73@xmath1224``w , 10@xmath1824''n ) with ( 126@xmath034 ) and ( 111@xmath027 ) spectroscopic clear nights per year , respectively .
site 3 stands out as the best site identified in colombia .
the eastern andes of colombia host two regions with more than 90 spectroscopic clear nights per year .
the region centered at site 5 ( 72@xmath14312``w , 7@xmath1112''n ) in _ norte de santander _ with ( 107@xmath023 ) clear nights per year and the wide zone centered at _ el espino _ comprising _ macaravita _ and _ el cocuy _ in _ boyac _ with ( 94@xmath013 ) clear nights per year .
the nearest locations to the equator in the _ nudo de los pastos _ region are site 9 ( 77@xmath15024``w , 0@xmath14648''n ) and site 11 ( 77@xmath15024``w , 0@xmath11536''n ) with ( 105@xmath036 ) and ( 101@xmath042 ) clear nights per year , respectively , presenting minimum seasonal differences of number of clear nights as expected .
finally , csf values below 25% observed during the rainy seasons of the years 2010 and 2011 seems to be correlated to positive values of the southern oscillation index .
the selected sites obtained in this study may be considered as a first - step to decide the location of an astronomical observatory .
however , site - survey is absolutely required before start any astronomical facility .
results of an observational campaign with the aim to characterize the sites identified in this study will be reported in a forthcoming paper .
we would like to further point out that the value of our approach goes beyond the possibility of being useful for identification of potential sites in the region .
our approach allows us to quantify the contribution of the different layers of the atmosphere to the csf above the site , in particular , the high troposphere and the surface .
although we have had to make a number of simplifying assumptions , the results are self - consistently and giv this reasearch was supported by colciencias and universidad nacional de colombia through the project no .
we thank _ instituto de hidrologa , meteorologa y estudios ambientales de colombia _ ( ideam ) for providing meteorological data set .
this publication makes use of data products from the oan - venezuela , which is operated by cida for the _ ministerio para el poder popular para la educacin universitaria , ciencia y tecnologa _ of venezuela .
we thank the invaluable assistance of the observers and night assistants at the oan - venezuela and the support of the computer and technical department of cida in the creation of the database used in this contribution as validation of our method .
brieva , e. , rmxaa , 1985 , 10 , 397 cavazzani , s. , ortolani , s. and zitelli , v. , 2012 , , 419 , 3081 cavazzani , s. and zitelli , v. , 2013 , , 429 , 1849 cavazzani , s. , ortolani , s. , zitelli , v. and maruccia , y. , 2011 , , 411 , 1271 della valle , a. , maruccia , y . ,
ortolani , s. and zitelli , v. , 2010 , , 401 , 1904 - 1916 erasmus , d.a . and sarazin , m. , 2002 , asp conf .
ser . , 266 , ed . j. vernin , z. benkhaldoun , and c. muoz - tuon , san francisco , p.310 erasmus , d.a . and van rooyen , r. , 2006 , ground - based and airborne telescopes . ed .
stepp , l. m. , proceedings of the spie , volume 6267 folkins , i. , fueglistaler , s. , lesins , g. and mitovski , t. , 2008 , j. atmos .
, 65 , 1019 - 1034 hidayat , t. , mahasena , p. , dermawan , b. , hadi , p. and premadi , p.w . , 2012 , , 427 , 1903 kneizys , f.x . ,
abreu , l.w . , anderson , g. p. , 1996
, the modtran 2/3 report and lowtran 7 model , atmos .
measurement marn , j. c. , pozo , d. and cur , m. , 2015 , a&a , 573 , a41 poveda , g. & lvarez , d. , 2011 , clim .
, 36 , 2233 poveda , g. & mesa , o. , 2000 , geophysical research letters , 27,11,1675 rossow , w.r . , kinsella , e. , wolf , a. & gardner , l. , 1985 international satellite cloud climatology project ( isccp ) , wmo / td - no .
58 ( 1985 ) soden , b. j. and bretherton , f. p. , 1993 , j.geophys.res . , 98 , d9 , 669 - 688 weinreb , m. p. , jamieson , m. , fulton , n. , 1997 , applied optics , 36 , 6895 1 & tim & timotes & 70@xmath12848 `` & 9@xmath1560 '' & 3480 & 65.4 & 52.1 & 68.8 & 56.8 & 55.5 & 42.7 & 61.1 & 43.4 & 84.0 & 72.5 & [email protected] & [email protected] & 220@xmath042 2 & cid & cida & 70@xmath15212 `` & 8@xmath14724 '' & 3600 & 51.4 & 27.0 & 45.0 & 26.6 & 38.7 & 20.2 & 40.8 & 12.3 & 58.5 & 41.6 & [email protected] & [email protected] & 132@xmath0343 & nab & nabusimake & 73@xmath13112 ``
& 10@xmath15512 '' & 3300 & 43.4 & 29.8 & 40.4 & 36.2 & 33.0 & 17.8 & 35.3 & 17.9 & 58.1 & 33.3 & [email protected] & [email protected] & 126@xmath034 4 & cod & codazzi & 73@xmath1224 `` & 10@xmath1824 '' & 2620 & 42.6 & 26.7 & 38.0 & 31.8 & 33.5 & 17.0 & 28.6 & 13.1 & 46.7 & 26.2 & [email protected] & [email protected] & 111@xmath0275 & chi & chitag&72@xmath14312 `` & 7@xmath1112 '' & 4060 & 39.5 & 20.8 & 36.6 & 25.7 & 31.7 & 16.0 & 33.1 & 14.0 & 45.9 & 30.2 & [email protected] & [email protected] & 107@xmath0236 & esp & el espino & 72@xmath1240 `` & 6@xmath1300 '' & 3620 & 33.6 & 19.9 & 37.5 & 20.3 & 29.5 & 14.6 & 34.4 & 14.0 & 39.9 & 13.5 & [email protected] & [email protected] & 94@xmath013 7 & pis & p. pisba & 72@xmath13336 `` & 5@xmath15848 '' & 3600 & 37.4 & 19.6 & 35.4 & 25.4 & 32.0 & 15.5 & 31.7 & 12.2 & 37.7 & 25.5 & 34.8@xmath81 & [email protected] & 99@xmath016 8 & coc & el cocuy & 72@xmath1448 `` & 6@xmath1300 '' & 3600 & 31.8 & 17.2 & 35.2 & 21.7 & 29.8 & 12.9 & 35.5 & 14.0 & ... & ... & [email protected]&[email protected] & 90@xmath0129 & tuf & p. tuffino & 77@xmath15024 `` & 0@xmath14648 '' & 3400 & 26.6 & 12.2 & 31.8 & 39.5 & 27.9 & 15.7 & 28.0 & 22.0 & 39.6 & 44.5 & [email protected]&[email protected] & 105@xmath03610 & hui & n. huila & 75@xmath15512 `` & 2@xmath15136 '' & 3700 & 27.6 & 12.4 & 28.2 & 24.8 & 34.0 & 10.5 & 24.9 & 17.8 & 35.4 & 33.2 & [email protected]&[email protected] & 91@xmath02511 & iba & ibarra & 77@xmath15024 `` & 0@xmath11536 '' & 3660 & 20.3 & 10.8 & 33.4 & 42.5 & 22.1 & 14.5 & 26.9 & 23.8 & 41.0 & 40.7 & [email protected]&[email protected] & 101@xmath04212 & lco & l. cocha & 77@xmath1224 `` & 1@xmath1224 '' & 3400 & 18.3 & 9.3 & 23.3 & 32.01 & 28.2 & 13.1 & 24.5 & 17.3 & 27.0 & 42.1 & [email protected]&[email protected] & 86@xmath032 1 & tim & timotes & 70@xmath12848 `` & 9@xmath1560 '' & 3480 & 54.3 & 36.7 & 55.5 & 35.2 & 44.2 & 29.1 & 44.8 & 28.5 & 57.9 & 44.8 & [email protected]&[email protected] & 157@xmath024 2 & cid & cida & 70@xmath15212 `` & 8@xmath14724 '' & 3600 & 29.6 & 14.1 & 29.3 & 13.1 & 25.8 & 13.1 & 25.7 & 3.1 & 34.0 & 27.3 & [email protected]&[email protected] & 79@xmath0223 & nab & nabusimake & 73@xmath13112 `` & 10@xmath15512 '' & 3300 & 31.8 & 17.9 & 24.2 & 16.9 & 24.3 & 8.7 & 24.2 & 7.1 & 30.5 & 11.4 & [email protected]&[email protected] & 72@xmath0164 & cod & codazzi & 73@xmath1224 `` & 10@xmath1824 '' & 2620 & 29.5 & 15.8 & 26.7 & 13.8 & 22.3 & 9.3 & 16.5 & 4.9 & 28.7 & 12.3 & [email protected] & [email protected] & 66@xmath018 5 & chi & chitag & 72@xmath14312 `` & 7@xmath1112 '' & 4060 & 27.6 & 10.8 & 23.2 & 11.4 & 20.7 & 8.2 & 20.7 & 6.7 & 29.3 & 10.9 & 24.3@xmath04 & [email protected] & 62@xmath0116 & esp & el espino & 72@xmath1240 `` & 6@xmath1300 '' & 3620 & 22.0 & 7.6 & 23.4 & 8.1 & 23.2 & 8.8 & 24.0 & 8.1 & 26.2 & 4.8 & [email protected]&[email protected] & 57@xmath067 & pis & p. pisba & 72@xmath13336 `` & 5@xmath15848 '' & 3600 & 25.4 & 10.3 & 23.2 & 12.2 & 23.2 & 6.3 & 21.3 & 5.6 & 26.8 & 6.4 & [email protected] & [email protected] & 59@xmath09 8 & coc & el cocuy & 72@xmath1448 `` & 6@xmath1300 '' & 3600 & 21.6 & 6.4 & 23.1 & 9.9 & 21.7 & 7.2 & 22.1 & 7.3 & ... & ... & [email protected]&[email protected] .1 & 54@xmath04 9 & tuf & p. tuffino & 77@xmath15024 `` & 0@xmath14648 '' & 3400 & 6.8 & 6.5 & 18.6 & 17.5 & 15.6 & 8.5 & 15.2 & 9.6 & 25.1 & 28.6 & [email protected]&[email protected]&56@xmath02910 & hui & n. huila & 75@xmath15512 `` & 2@xmath15136 '' & 3700 & 12.3 & 5.1 & 10.7 & 10.6 & 19.9 & 6.0 & 13.3 & 8.7 & 13.5 & 8.8 & [email protected]&[email protected] & 40@xmath011 11 & iba & ibarra & 77@xmath15024 `` & 0@xmath11536 '' & 3660 & 7.9 & 6.0 & 21.1 & 17.8 & 10.3 & 8.0 & 15.5 & 14.6 & 22.7 & 29.0 & [email protected] & [email protected] & 56@xmath02912 & lco & l. cocha & 77@xmath1224 `` & 1@xmath1224 '' & 3400 & 5.3 & 2.7 & 11.0 & 12.1 & 14.3 & 4.0 & 11.3 & 7.3 & 13.3 & 22.8 & [email protected]&[email protected] & 38@xmath021 | in this study we describe an innovative method to determine potential sites for optical and infrared astronomical observations in the andes region of northern south america .
the method computes the clear sky fraction ( csf ) from geostationary observational environmental satellite ( goes ) data for the years 2008 - 12 through a comparison with temperatures obtained from long - term records of weather stations and atmospheric temperature profiles from radiosonde .
criteria for sky clearance were established for two infrared goes channels in order to determine potential sites in the andes region of northern south - america .
the method was validated using the reported observed hours at _ observatorio nacional de llano del hato _ in venezuela .
separate csf percentages were computed for dry and rainy seasons for both , photometric and spectroscopic night qualities .
twelve sites with five year averages of csf for spectroscopic nights larger than 30% during the dry seasons were found to be suitable for astronomical observations .
the best site with ( 220@xmath042 ) spectroscopic clear nights per year is located in the andes of venezuela ( 70@xmath12848``w , 9@xmath1560''n ) at an altitude of 3480 meters .
lower quality regions were found in _
sierra nevada de santamarta _ and _ serrana del perij _ with ( 126@xmath034 ) and ( 111@xmath027 ) clear nights per year , respectively . sites over the andes are identified in _
norte de santander _ with ( 107@xmath023 ) and in the north - east part of _ boyac _ with a mean of ( 94@xmath013 ) clear nights per year .
two sites at low latitude located in ecuador with more than 100 clear nights per year and with similar seasonal csf percentages were also identified .
five year evolution suggest a possible correlation between the lowest percentages observed during the rainy seasons of 2010 and 2011 with positive values of the southern oscillation index . |
in hole doped high-@xmath1 cuprates , the hole density @xmath3 in cuo@xmath0 planes has been determined by various methods : indirect chemical methods like solid solutions @xcite , bond valence sums determined from structural bond lengths @xcite , or the fermi surface topology @xcite .
moreover , the thermoelectric power is a universal function of @xmath3 @xcite , and the phase diagram for hole doped cuprates is well described by @xmath1=@xmath10 $ ] @xcite , which is applicable for the estimation of @xmath3 when suitable structural data are not available .
these methods are , however , inapplicable when we deal with multilayered compounds because they are composed of more than two inequivalent cuo@xmath0 planes in a unit cell ( see fig .
[ fig : n3nmr](a ) as an example ) . in multilayered cuprates ,
doped hole carriers are inequivalent between outer cuo@xmath0 planes ( ops ) and inner cuo@xmath0 planes ( ips ) due to the imbalance of the madelung potential on each cuo@xmath0 plane ; the above - mentioned methods would evaluate not _ each hole density _ inherent to cuo@xmath0 planes but a _ total hole density_. precise determination of hole density on respective cuo@xmath0 planes is indispensable for gaining deep insight into the phenomena observed in multilayered cuprates @xcite .
it is known that the spin part of the knight shift @xmath2(300 k ) at room temperature increases with @xmath3 in hole doped cuprates @xcite .
the @xmath2(300 k ) for op and ip are separately determined by cu - nmr . therefore , each value of @xmath3 for op and for ip can be estimated if the relationship between @xmath2(300 k ) and @xmath3 is established . meanwhile , a linear equation @xmath3=0.502@xmath2(300 k)+0.0462 has been reported @xcite , where @xmath3 was derived from the nqr frequencies of cu and o in cuo@xmath0 planes @xcite .
the values of @xmath3 estimated by the linear equation , however , seem relatively large . for example
, an optimal doping level has been evaluated to be @xmath3 @xmath11 0.22 - 0.24 in five - layered hg - based compounds @xcite , the value of which is larger than a widely - accepted optimal doping level in hole - doped cuprates , @xmath3 @xmath11 0.16 .
this inconsistency is probably , in part , due to the calculation that connects nqr frequencies to @xmath3 @xcite .
in addition to that , it is nontrivial whether the @xmath3-dependence of @xmath2(300 k ) is described with a single equation which holds among different materials .
@xmath2 is proportional to spin susceptibility @xmath12 and hyperfine coupling constants @xmath13 as @xmath2=@xmath14 , where the supertransferred magnetic field @xmath15 in @xmath13=@xmath16 + 4@xmath15 depends on the materials @xcite . in this paper
, we report that the planar cuo@xmath0 hole density @xmath3 in high-@xmath1 cuprates is consistently evaluated from the spin part of the @xmath17cu - knight shift @xmath2(300 k ) at room temperature .
we show the @xmath3 dependences of @xmath2(300k ) and of @xmath15 in bi - layered ba@xmath0cacu@xmath0o@xmath7(f , o)@xmath0 ( 0212f ) , together with other single- and bi - layered materials with tetragonal symmetry .
nmr data on three - layered ba@xmath0ca@xmath0cu@xmath4o@xmath5(f , o)@xmath0 ( 0223f ) , hgba@xmath0ca@xmath0cu@xmath4o@xmath6 ( hg1223 ) @xcite , and hgba@xmath0ca@xmath7cu@xmath8o@xmath9 ( hg1245 ) @xcite suggest that the @xmath3-dependencies of @xmath2(300 k ) and of @xmath15 hold in multilayered compounds as well .
these results show that the present method , based on the knight shift , enables us to separately estimate @xmath3 for each cuo@xmath0 plane in multilayered high-@xmath1 cuprates .
= 3.5 mm [ cols="^,^,^,^,^,^",options="header " , ] [ t : ggg2 ] in multilayered cuprates , doped hole carriers reside on op and ip with different doping levels , and the cuo@xmath0 layers show different physical properties due to the charge distribution even in the same sample .
therefore , it is required to separately estimate @xmath3 for op and ip in order to study the electronic states of multilayered cuprates .
it is invalid to apply the estimation methods used in single- and bi - layered compounds , for example , the relationship @xmath1=@xmath18[1 - 82.6(@xmath3 - 0.16)@xmath19 @xcite , the thermoelectric power @xcite , and the bond valence sums @xcite .
those methods are applicable to evaluate total hole density , but not to evaluate each hole density at cuo@xmath0 planes . on the other hand
, cu - nmr measures the respective values of @xmath20(300 k ) for op and ip . if the @xmath20(300 k)-@xmath3 relationship in fig .
[ fig : bterm](a ) holds even in multilayered cuprates , it allows us to separately estimate @xmath3 for op and ip from experimental values of @xmath20(300 k ) .
table [ t : ggg2 ] lists @xmath20(300 k ) , @xmath21 , and @xmath15 for multilayered cuprates : three - layered 0223f , hgba@xmath0ca@xmath0cu@xmath4o@xmath6 ( hg1223 ) @xcite , and five - layered hgba@xmath0ca@xmath0cu@xmath4o@xmath9 ( hg1245 ) @xcite . here , @xmath21 is the hole density tentatively estimated by using the @xmath20(300 k)-@xmath3 relationship in fig .
[ fig : bterm](a ) .
we plot @xmath15 of those compounds as crosses ( @xmath22 ) in fig .
[ fig : bterm](b ) .
the data plots fit into the other data , suggesting that the @xmath20(300 k)-@xmath3 relationship is also valid in multilayered cuprates .
we may consider the case that the @xmath20(300 k)-@xmath3 relationship is modified in multilayered cuprates . according to eq .
( [ eq : ks ] ) , the possible source of the modification is the @xmath3 dependence of @xmath15 or of @xmath12 or both .
experimental @xmath15 values in multi - layered compounds , however , fall on the same universal curve as shown in fig .
[ fig : bterm](b ) ; some unexpected modifications on the side of @xmath12 can indeed be considered as unlikely .
therefore , we conclude that the @xmath20(300 k)-@xmath3 relationship in fig . [
fig : bterm](a ) is valid for cuo@xmath0 planes in multilayered compounds , regardless of op and ip .
moreover , when we apply the @xmath20(300 k)-@xmath3 relationship to five - layered compounds , the fact that the maximum @xmath1 takes place at @xmath3 @xmath11 0.16 @xcite supports the validity of the application to multilayered compounds .
several attempts have been used to determine @xmath3 of hole - doped cuprates so far the relationship between @xmath3 and @xmath20(300 k ) in this work is a promising approach to estimate @xmath3 , which is _ applicable to multilayered compounds_. to our knowledge , it is the only method to separately determine @xmath3 for op and ip in multilayered compounds with a reliable accuracy .
we have shown that the planar cuo@xmath0 hole densities @xmath3 in high-@xmath1 cuprates are consistently determined with the cu - nmr knight shift .
it has been demonstrated that the spin part of the knight shift @xmath20(300 k ) at room temperature is material - independent in cuo@xmath0 planes with tetragonal symmetry , and that @xmath20(300 k ) monotonically increases with @xmath3 from underdoped to overdoped regions .
these observations suggests that @xmath20(300 k ) is a reliable method for determining planar cuo@xmath0 hole densities .
the experimental values of @xmath20(300 k ) and of hyperfine magnetic fields for three - layered ba@xmath0ca@xmath0cu@xmath4o@xmath5(f , o)@xmath0 and other multilayered compounds support the application of the @xmath3-@xmath2(300 k ) relationship to multilayered compounds .
we remark that , to our knowledge , the relationship is the only method to separately determine @xmath3 for op and ip in multilayered compounds with a reliable accuracy .
the authors are grateful to m. mori for his helpful discussions .
this work was supported by grant - in - aid for specially promoted research ( 20001004 ) and by the global coe program ( core research and engineering of advanced materials - interdisciplinary education center for materials science ) from the ministry of education , culture , sports , science and technology ( mext ) , japan .
m. w. shafer , t. penney , and b. l. olson , phys .
b * 36 * , 4047 ( 1987 ) . k. kishino , j .- i .
shimoyama , t. hasegawa , k. kitazawa , and k. fueki , jpn .
* 26 * , l1228 ( 1987 )
. y. tokura , j. b. torrance , t. c. huang , and a. i. nazzal , phys .
b * 38 * , 7156 ( 1988 ) .
j. b. torrance , y. tokura , a. i. nazzal , a. bezinge , t. c. huang , and s. s. p. parkin , phys .
lett . * 61 * , 1127 ( 1988 ) .
s. d. obertelli , j. r. cooper , and j. l. tallon , phys .
b * 46 * , 14928 ( 1992 )
. j. l. tallon , c. bernhard , h. shaked , r. l. hitterman , and j. d. jorgensen , phys .
b * 51 * , r12911 ( 1995 ) .
m. r. presland , j. l. tallon , r. g. buckley , r. s. liu , and n. e. flower , physica c * 176 * , 95 ( 1991 ) .
y. tokunaga , k. ishida , y. kitaoka , k. asayama , k. tokiwa , a. iyo , and h. ihara , phys .
b * 61 * , 9707 ( 2000 ) .
h. kotegawa , y. tokunaga , k. ishida , g .- q .
zheng , y. kitaoka , h. kito , a. iyo , k. tokiwa , t. watanabe , and h. ihara , phys .
b * 64 * , 064515 ( 2001 ) .
h. mukuda , y. yamaguchi , s. shimizu , y. kitaoka , p. m. shirage , and a. iyo , j. phys .
. jpn . * 77 * , 124706 ( 2008 ) .
s. shimizu , h. mukuda , y. kitaoka , h. kito , y. kodama , p. m. shirage , and a. iyo , j. phys .
. jpn . * 78 * , 064075 ( 2009 ) .
s. shimizu , s .-
tabata , h. mukuda , y. kitaoka , p. m. shirage , h. kito , and a. iyo , arxiv:1102.5282 , j. phys .
( to be published , vol .
k. fujiwara , y. kitaoka , k. asayama , y. shimakawa , t. manako , and y. kubo , j. phys .
jon . * 59 * , 3459 ( 1990 ) .
r. e. walstedt , w. w. warren , jr .
, r. f. bell , r. j. cava , g. p. espinosa , l. f. schneemeyer , and j. v. waszczak , phys .
b * 41 * , 9574 ( 1990 ) .
s. ohsugi , y. kitaoka , k. ishida , g .- q .
zheng , and k. asayama , j. phys .
63 * , 700 ( 1994 ) .
k. magishi , y. kitaoka , g .- q .
zheng , k. asayama , t. kondo , y. shimakawa , t. manako , and y. kubo , phys .
b * 54 * , 10131 ( 1996 )
. k. ishida , k. yoshida , t. mito , y. tokunaga , y. kitaoka , k. asayama , a. nakayama , j. shimoyama , and k. kishio , phys .
b * 58 * , r5960 ( 1998 ) .
j. g. storey , j. l. tallon , and g. v. m. williams , phys . rev .
b * 76 * , 174522 ( 2007 ) .
y. tokunaga , h. kotegawa , k. ishida , g .- q .
zheng , y. kitaoka , k. tokiwa , a. iyo , and h. ihara , j. low temp .
phys . * 117 * , 314 ( 1999 ) . g .- q .
zheng , y. kitaoka , k. ishida , and k. asayama , j. phys .
. jpn . * 64 * , 2524 ( 1995 ) .
j. haase , o. p. sushkov , p. horsch , and g. v. m. williams , phys .
b * 69 * , 094504 ( 2004 ) .
f. mila and t. m. rice , phys .
b * 40 * , 11382 ( 1989 ) .
k. magishi , y. kitaoka , g .- q .
zheng , k. asayama , k. tokiwa , a. iyo , and h. ihara , j. phys .
. jpn . * 64 * , 4561 ( 1995 ) .
h. kotegawa , y. tokunaga , y. araki , g .- q .
zheng , y. kitaoka , k. tokiwa , k. ito , t. watanabe , a. iyo , y. tanaka , and h. ihara , phys .
b * 69 * , 014501 ( 2004 ) .
a. iyo , y. tanaka , m. tokumoto , and h. ihara , physica c * 366 * , 43 ( 2001 ) .
a. iyo , m. hirai , k. tokiwa , t. watanabe , and y. tanaka , physica c * 392 * , 140 ( 2003 ) .
p. m. shirage , d. d. shivagan , y. tanaka , y. kodama , h. kito , and a. iyo , appl .
. lett . * 92 * , 222501 ( 2008 ) .
s. shimizu , h. mukuda , y. kitaoka , a. iyo , y. tanaka , y. kodama , k. tokiwa , and t. watanabe , phys .
* 98 * , 257002 ( 2007 ) .
s. shimizu , t. sakaguchi , h. mukuda , y. kitaoka , p. m. shirage , y. kodama , and a. iyo , phys .
b * 79 * , 064505 ( 2009 ) .
h. yasuoka , t. imai , and t. shimizu , in _ strong correlation and superconductivity _ , edited by h. fukuyama , s. maekawa , and a. p. malozemoff ( springer , berlin , 1989 ) vol .
89 , p. 254 .
r. e. walstedt , in _ the nmr probe of high-@xmath1 materials _
( springer , berlin , 2008 ) .
y. kitaoka , k. fujiwara , k. ishida , k. asayama , y. shimakawa , t. manako , and y. kubo , physica c * 179 * , 107 ( 1991 ) .
julien , m. horvati , p. carretta , c. berthier , y. berthier , p. sgransan , s. m. loureiro , and j .- j .
capponi , physica c * 268 * , 197 ( 1996 ) .
s. e. barrett , d. j. durand , c. h. pennington , c. p. slichter , t. a. friedmann , j. p. rice , and d. m. ginsberg , phys .
b * 41 * , 6283 ( 1990 ) .
m. takigawa , a. p. reyes , p. c. hammel , j. d. thompson , r. h. heffner , z. fisk , and k. c. ott , phys .
b. * 43 * , 247 ( 1991 ) .
h. zimmermann , m. mali , m. bankay , and d. brinkmann , physica c * 185 - 189 * , 1145 ( 1991 ) .
y. kitaoka , k. fujiwara , k. ishida , k. asayama , y. shimakawa , t. manako , and y. kubo , physica c * 176 * , 107 ( 1991 ) .
k. fujiwara , y. kitaoka , k. ishida , k. asayama , y. shimakawa , t. manako , and y. kubo , physica c * 184 * , 207 ( 1991 ) . c. h. pennington , d. j. durand , c. p. slichter , j. p. rice , e. d. bukowski , and d. m. ginsberg , phys .
b * 39 * , 2902 ( 1989 ) . | we report that planar cuo@xmath0 hole densities in high-@xmath1 cuprates are consistently determined by the cu - nmr knight shift . in single- and bi - layered cuprates , it is demonstrated that the spin part of the knight shift @xmath2(300 k ) at room temperature monotonically increases with the hole density @xmath3 from underdoped to overdoped regions , suggesting that the relationship of @xmath2(300 k ) vs. @xmath3 is a reliable measure to determine @xmath3 .
the validity of this @xmath2(300 k)-@xmath3 relationship is confirmed by the investigation of the @xmath3-dependencies of hyperfine magnetic fields and of spin susceptibility for single- and bi - layered cuprates with tetragonal symmetry .
moreover , the analyses are compared with the nmr data on three - layered ba@xmath0ca@xmath0cu@xmath4o@xmath5(f , o)@xmath0 , hgba@xmath0ca@xmath0cu@xmath4o@xmath6 , and five - layered hgba@xmath0ca@xmath7cu@xmath8o@xmath9 , which suggests the general applicability of the @xmath2(300 k)-@xmath3 relationship to multilayered compounds with more than three cuo@xmath0 planes .
we remark that the measurement of @xmath2(300 k ) enables us to separately estimate @xmath3 for each cuo@xmath0 plane in multilayered compounds , where doped hole carriers are inequivalent between outer cuo@xmath0 planes and inner cuo@xmath0 planes . |
the past four years have seen dramatic advances in our understanding of low - mass x - ray binaries ( lmxbs ) thanks to precision - timing observations made with the _ rossi x - ray timing explorer _ satellite ( _ rxte _ ; bradt , rothschild & swank 1993 ) .
quasi - periodic oscillations ( qpos ) in the 2001200 hz frequency range have been detected in the persistent emission of many lmxbs .
these qpos often occur simultaneously separated by a frequency difference that remains roughly constant or decreases as the frequencies increase ; these frequencies may reflect the relativistic motion of material at the inner edge of the accretion disk in these systems .
for a recent review see van der klis ( 2000 ) .
separately , nearly - coherent oscillations ( ncos ) with @xmath3 ( @xmath4 ) values of @xmath5300 have been detected during many type i x - ray bursts from six lmxbs , and there is strong evidence that these frequencies reflect the underlying neutron star spin period for these sources ( strohmayer 1996 ; for a review see strohmayer 1999 ) . in particular , ( 1 ) the high coherence of the signals is a challenge to alternative models : @xmath6 in a single burst from ks 1731@xmath7260 ( smith , morgan , & bradt 1997 ) , and @xmath3 values of @xmath84000 have been demonstrated , by modeling of the frequency evolution during the burst , for bursts from and 4u 1702@xmath7429 ( strohmayer & markwardt 1999 ) and ( muno 2000 ) ; ( 2 ) the long - term stability ( to better than one part in 1000 ) of the nco period in at least three sources has been demonstrated over a time scale of @xmath81 year ( strohmayer 1998b ; muno 2000 ) ; ( 3 ) the strength of the oscillations varies throughout the burst as expected for a spin modulation : the oscillations are generally strongest on the leading edge and tail of the bursts and undetectable at the burst peak , when a large portion of the ns surface is in conflagration ( strohmayer , swank & zhang 1998a ) ; and ( 4 ) pulse phase - resolved spectroscopy of a burst from 4u 1636@xmath7536 indicates that the flux modulation during the tail of the burst is accompanied by a modulation of the blackbody temperature of the spectrum , again as expected ( strohmayer 1998a ) .
the rapid burster ( , or rb hereafter ; lewin 1976 ) is a recurrent transient lmxb , located in the galactic plane at a distance of approximately 8.6 kpc ( frogel 1995 ) in the highly reddened globular cluster liller 1 ( liller 1977 ) .
it is the only lmxb known to produce both type i and type ii x - ray bursts ( hoffman , marshall & lewin 1978 ) . whereas the type i bursts result from the explosive thermonuclear burning of accreted material on the surface of a neutron star , the type ii bursts result from spasmodic accretion
the release of gravitational potential energy presumably because of a recurrent accretion instability . for a review , see lewin , van paradijs , & taam ( 1993 ) .
guerriero ( 1998 ) undertook a search for pulsations from the type i bursts , type ii bursts , and persistent emission of the rapid burster , as observed through the first four outbursts of the source seen with _ rxte_. no significant detections were made , with upper limits for persistent signals of @xmath93 per cent rms variation in the range 202048 hz .
given the unique nature of the rapid burster , we undertook to perform a more sensitive search .
we used the knowledge that has been gleaned from other sources about the nco phenomenon , focusing our search on the leading edge of the type i bursts , and on the higher energy photons only ( see below , sect .
[ sec : obs ] ) .
preliminary results of this work were reported by fox & lewin ( 1999a ) .
as of january 2000 , seven outbursts of the rapid burster have been observed with _ rxte _ ( lewin 1996a , b ; guerriero , lewin & kommers 1997 ; guerriero 1998 ; fox & lewin 1998 ; fox 1998 ; fox & lewin 1999b ; the sixth outburst began in march 1999 and was not reported ) .
_ rxte _ all - sky monitor ( asm ; levine 1996 ) data for these outbursts are available in `` quicklook '' and fully - processed form from the web sites of the asm team and the high - energy astrophysics science archive research center , respectively .
alerted by the asm , we are able to commence target of opportunity observations with the proportional counter array ( pca ; jahoda 1996 ) within one day of the start of an outburst , and to follow it throughout its month - long evolution ( guerriero 1999 ) . outburst profiles , as observed with the _ rxte _ asm , typically exhibit a sharp 13 d rise to a maximum flux of @xmath9400 mcrab ( 212 kev ; the crab produces 75 c s@xmath10 in the asm cameras ) , as time - averaged over the 90-second asm dwells .
the x - ray flux then declines quasi - exponentially over the next month with a time constant of @xmath98 d ( guerriero 1999 ) .
note , however , that the sixth outburst in march 1999 deviated significantly from this pattern ( fox 1999 ) .
the timing analysis presented here relies exclusively on the pca data .
our observations used individually described , event - encoded data with a time resolution of 122 @xmath11s and 64 energy channels . for the usual spectral parameters of the rb , one pca c s@xmath10 @xmath12 ergs @xmath13 s@xmath10 , 220 kev ( guerriero 1999 ) .
when the outburst reaches a maximum ( pe and type i bursts ) , count rates are @xmath95000 pca c s@xmath10 corresponding to an isotropic luminosity of @xmath14 ergs s@xmath10 ( 220 kev ) at a distance of 8.6 kpc .
the brightest type i burst peak luminosities , after pe subtraction , are similarly @xmath8@xmath15 ergs s@xmath10 ( see table 1 ) .
l@ llrrrrr & & & & & & & + & & & & & & + + * 1 & ( a ) & 1996 - 11 - 06 & 20:00:49 & 3256 & 4960 & 1.4 & * 2138 + * 2@xmath16&(a ) & 1996 - 11 - 09 & 01:06:20 & 1056 & 2664 & 1.3 & * 1496 + 3@xmath16&(c?)&1996 - 11 - 11 & 21:52:45 & 816 & 656 & 0.2 & 1856 + + 4 & ( b ) & 1997 - 06 - 26 & 05:03:43 & 5360 & 2072 & 2.4 & 240 + 5 & ( b ) & 1997 - 06 - 26 & 05:13:23 & 5288 & 2272 & 5.6 & 123 + 6@xmath16&(b ) & 1997 - 06 - 26 & 05:22:42 & 2112 & 1000 & 4.6 & 70 + 7 & ( b ) & 1997 - 06 - 26 & 08:21:13 & 5696 & 2368 & 2.8 & 251 + 8 & ( b ) & 1997 - 06 - 26 & 08:28:59 & 5784 & 1632 & 7.2 & 50 + 9@xmath16&(b ) & 1997 - 06 - 26 & 08:40:36 & 2368 & 1144 & 2.6 & 145 + 10@xmath16&(b ) & 1997 - 06 - 26 & 08:48:51 & 2200 & 1120 & 3.6 & 105 + * 11 & ( b ) & 1997 - 06 - 27 & 18:13:15 & 5152 & 3448 & 2.2 & 642 + * 12@xmath17&(b ) & 1997 - 06 - 27 & 18:30:34 & 5264 & 2768 & 1.8 & 536 + * 13@xmath16&(a ) & 1997 - 06 - 29 & 07:03:31 & 1496 & 2016 & 1.5 & 783 + * 14 & ( a ) & 1997 - 06 - 29 & 18:07:23 & 4016 & 4000 & 1.4 & 1409
+ * 15@xmath16&(a ) & 1997 - 06 - 29 & 18:35:24 & 1656 & 1560 & 0.6 & 1226 + * 16@xmath16&(a ) & 1997 - 07 - 07 & 13:29:06 & 664 & 2088 & 1.1 & 1391 + * 17@xmath18&(a ) & 1997 - 07 - 10 & 14:03:41 & 528 & 2208 & 1.2 & * 1473 + * 18 & ( a ) & 1997 - 07 - 13 & 12:01:19 & 1136 & 1728 & 1.5 & 706 + + * 19 & ( a ) & 1998 - 01 - 30 & 18:31:04 & 4680 & 3424 & 2.0 & 716 + * 20@xmath17&(a ) & 1998 - 01 - 30 & 18:54:18 & 4784 & 3424 & 1.6 & 867 + * 21 & ( a ) & 1998 - 01 - 30 & 20:02:22 & 4344 & 3584 & 1.5 & 1088 + * 22 & ( a ) & 1998 - 01 - 30 & 20:28:12 & 4328 & 3864 & 1.7 & 1082 + 23@xmath19&(a ) & 1998 - 01 - 30 & 20:54:18 & 4232 &
4200 & 1.5 & n / a + * 24 & ( a ) & 1998 - 01 - 31 & 23:25:27 & 3672 & 3664 & 1.6 & 1169 + * 25 & ( a ) & 1998 - 01 - 31 & 23:53:09 & 3712 & 3736 & 1.3 & * 1494 + * 26 & ( a ) & 1998 - 02 - 02 & 17:00:47 & 2480 & 4720 & 1.3 & * 2349 + * 27 & ( a ) & 1998 - 02 - 02 & 18:46:18 & 2552 & 4912 & 1.2 & * 2703 + * 28 & ( a ) & 1998 - 02 - 04 & 20:21:48 & 2288 & 4960 & 1.7 & * 1954 + * 29 & ( a ) & 1998 - 02 - 07 & 19:00:34 & 1776 & 5328 & 1.5 & * 2746 + * 30 & ( a ) & 1998 - 02 - 07 & 20:04:54 & 1744 & 4920 & 1.3 &
* 2800 + * 31@xmath20&(c ) & 1998 - 02 - 10 & 22:06:44 & 1424 & 4208 & 1.2 & * 2587 + * 32 & ( c ) & 1998 - 02 - 16 & 14:04:17 & 1256 & 3632 & 0.5 & * 5292 + 33@xmath21&(c ) & 1998 - 02 - 16 & 15:13:01 & 1168 & 1608 & 2.3 & n / a + * 34@xmath16&(c ) & 1998 - 02 - 19 & 14:29:55 & 192 & 2896 & 0.3 & * 8103 + + * 35@xmath22&(a ? ) & 1998 - 08 - 19 & 05:07:19 & 13664 & 5232 & 4.7 & 306 + * 36@xmath23 & ( a ? ) & 1998 - 08 - 20 & 13:21:38 & 10848 & 3256 & 1.0 & 765 + * 37@xmath24 & ( d ) & 1998 - 08 - 22 & 08:24:11 & 3328 & 3448 & 1.3 & 1313 + * 38@xmath24 & ( d ) & 1998 - 08 - 22 & 09:01:59 & 3272 & 3384 & 1.6 & 1107 + * 39@xmath25&(d ) & 1998 - 08 - 22 & 10:15:45 & 1184 & 1384 & 1.6 & 453 + * 40 & ( c ) & 1998 - 09 - 01 & 02:13:16 & 1152 & 4480 & 0.6 & * 5603 + * 42x@xmath24 & ( c ? ) & 1998 - 09 - 07 & 03:35:12 & 608 & 2288 & 3.3 & 541 + + 41@xmath20&(a ) & 1999 - 03 - 12 & 21:50:02 & 3440 & 2776 & 2.7 & 467 + 42@xmath20&(a ) & 1999 - 03 - 12 & 22:12:53 & 3344 & 2928 & 1.5 & 935 + 43@xmath26&(a ) & 1999 - 03 - 16 & 02:32:05 & 1616 & 1488 & 1.2 & 593 + 44@xmath26&(a ) & 1999 - 03 - 16 & 03:00:04 & 1576 & 1448 & 1.4 & 501 + 45@xmath20&(a ) & 1999 - 03 - 16 & 03:55:32 & 3152 & 2600 & 0.9 & 1336 + 46@xmath20&(a ) & 1999 - 03 - 16 & 04:23:23 & 3104 & 2968 & 1.0 & 1411 + 47@xmath20&(a ) & 1999 - 03 - 19 & 19:57:22 & 1056 & 1448 & 1.4 & 586 + * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * l@ llrrrrr & & & & & & & + & & & & & & + + 48@xmath27 & ( b ) & 1999 - 09 - 30 & 20:35:00 & 2632 & 1848 & 1.6 & 466 + 49@xmath27 & ( b ) & 1999 - 09 - 30 & 20:55:54 & 2512 & 1880 & 2.7 & 299 + 50@xmath27 & ( b ) & 1999 - 09 - 30 & 21:17:23 & 2464 & 2048 & 2.2 & 420 + 51@xmath27 & ( b ) & 1999 - 09 - 30 & 22:03:53 & 2488 & 2136 & 1.6 & 614 + 52@xmath27 & ( a ) & 1999 - 09 - 30 & 22:26:54 & 2368 & 2408 & 2.0 & 607 + 53@xmath26 & ( a ) & 1999 - 10 - 02 & 09:28:25 & 1520 & 1728 & 1.8 & 518 + 54@xmath26 & ( a ) & 1999 - 10 - 02 & 09:54:57 & 1528 & 1896 & 1.8 & 572 + 55@xmath26 & ( a ) & 1999 - 10 - 02 & 10:49:25 & 1528 & 1744 & 1.6 & 565 + 56@xmath26 & ( a ) & 1999 - 10 - 02 & 11:16:28 & 1520 & 1760 & 1.6 & 596 + 57@xmath28&(a)&1999 - 10 - 02 & 11:44:19 & 544 & 736 & 1.3 & 335 + 58@xmath29&(a ) & 1999 - 10 - 05 & 12:11:53 & 1752 & 3112 & 1.6 & 1273 + 59@xmath27&(a ) & 1999 - 10 - 05 & 12:49:39 & 1768 & 3160 & 1.7 & 1193 + 60 & ( a ) & 1999 - 10 - 08 & 12:40:07 & 2736 & 4672 & 1.3 &
* 2335 + 61@xmath20&(a ) & 1999 - 10 - 12 & 12:29:52 & 1664 & 4024 & 1.0 & * 2926 + 62@xmath30&(a ) & 1999 - 10 - 12 & 14:28:52 & 632 & 1744 & 1.6 & 825 + 63 & ( a ) & 1999 - 10 - 16 & 04:13:44 & 1344 & 4688 & 1.0 & * 3487 + 64@xmath31&(c ) & 1999 - 10 - 19 & 06:12:23 & 176 & 1872 & 0.2 & * 8612 + * * * * dates and times are in ut .
profile classes for the bursts correspond to the examples shown in fig .
[ fig : profiles ] . persistent emission + background ( pe ) and burst peak count rates ( peak ; for 1/8-second bins ) are in pca cts / sec ( 260 kev )
. burst rise times ( in seconds ) are characteristic times for exponential fits ; merit figures are calculated as merit = ( peak@xmath32/pe)/rise .
boldface burst numbers indicate the 31 bursts of the original search ; boldface merit figures indicate the 17 bursts of the final search ; see text for details .
@xmath16observed during offset pointing ( see text ) ; @xmath17burst tail truncated by slew or data gap ; @xmath19not recorded in event mode data ; @xmath21only partially recorded in event mode data ; @xmath23occurred during a type ii burst ; @xmath24identification as type i burst is uncertain ; @xmath204 of 5 pcus operational ; @xmath273 of 5 pcus operational ; @xmath262 of 5 pcus operational type
i x - ray bursts are distinguished from the type ii bursts of the rb in that they show a pronounced cooling over the course of the burst ; type ii bursts , on the other hand , maintain a roughly constant temperature through the burst .
near the maximum of the rb outbursts , when the persistent emission is at its strongest , we have occasionally observed bursts with a slow rise , slow decay and relatively low peak flux classified as b - profile bursts below that we have difficulty identifying definitively as type i or type ii . for the most part
we have identified them as type i on the basis of spectral analyses ( guerriero 1999 ) ; however , this identification is not certain in all cases . in figure [
fig : profiles ] we present the profiles of four type i bursts from the rapid burster that indicate the variety of burst profiles that we observe .
table 1 identifies all of the bursts by profile class ; note , however , that these identifications are merely gross characterizations , and in particular , the distinction between profile classes ( a ) and ( b ) is probably not well - defined at this point .
we note that an occasional type i burst from is visible at @xmath950 per cent collimation efficiency during direct pointings at the rb , but that these bursts exhibit spectra ( fitted blackbody temperatures ( @xmath92 kev ) , peak count rates ( @xmath910,000 c s@xmath10 ) , and light curves ( @xmath910 sec decays ) that readily distinguish them from the rb bursts .
our analysis evolved historically and must be described as such .
it began in 1998 after _ rxte _ had observed five outbursts from the rapid burster .
we compiled a list of all rb type i bursts that were observed with _
rxte _ during these outbursts ( table 1 , data through september 1998 ) ; note that no type i bursts were observed during the first outburst ( april and may 1996 ) .
we selected one - second time segments from the start of each burst : an automated procedure found the first eighth - second bin to represent a 4.5@xmath33 fluctuation above the background level , and these start times were then inspected visually and adjusted by a half - second or less by hand . to maximize our sensitivity to pulsations modulating the burst flux alone ( and not the background or pe ) , we rank - ordered the bursts in terms of peak@xmath32/pe ( table 1 ) , and selected the top 31 bursts for our sample , as these seemed the most attrative targets ; the selected bursts are indicated in bold in table 1 . to increase signal - to - noise for a relatively hard pulsation we selected photons in the 5.516 kev bandpass , roughly the upper half of the energy range for the rapid burster spectrum .
we binned the light curves from these one - second intervals to @xmath34 sec time resolution , corrected times to the solar system barycentre , took the fourier transform , and averaged the 31 resulting power density spectra producing a single power density spectrum ( pds ) of 1 hz frequency resolution and nyquist frequency 2048 hz .
the frequency of the ncos often drifts by roughly 2 hz during the bursts ( strohmayer 1996 ) , so we do not necessarily anticipate a one - bin excess in the pds . rather than search for the highest individual powers , then , we began by inspecting the running two - bin average of adjacent powers from the pds .
then , in a check for broader signals , we rebinned the pds by successive factors of two and , at each stage , performed a two - bin running - average search on the averaged powers .
since the statistical distribution of poisson noise powers is known ( for raw powers it is the @xmath35 distribution with two degrees of freedom ; leahy 1983 ) , we were able to compare the search results from different stages of this process and select those which are most significant in a global sense .
in fact , averaging 31 pdss already puts the powers in a near - gaussian regime , and we adopted this approximation in ranking prospective triggers of different bin sizes .
we have performed monte carlo calculations to determine the absolute significance of any particular excess over the entire power - spectral search .
we did this because , first , the running - average search procedure ( with its multiple rebinnings and reexaminations of the pds ) results in a distribution of tested powers that does not have an analytical form ; and second , because the burst rise injects strong low - frequency power and places sidebands on noise spikes , increasing the noise powers for an integral - power search such as ours .
our monte carlo ( mc ) procedure used the actual burst data , scrambling the data on short time scales to destroy real high - frequency power while preserving the burst rise and low - frequency noise responsible for the amplification of noise spikes in the pds .
the 120 time bins from each 30 msec interval of a burst were randomly interchanged prior to performing the fft .
we then averaged the 31 `` fake '' pdss from the individual bursts and performed our full search protocol , recording the highest three signals that were found .
for the purposes of this paper , we performed an mc run of 100,000 trials .
there are two aspects of this procedure which , if ignored , will systematically affect the resulting statistics . first , real noise is present in our fake pdss at low frequencies , @xmath0100 hz , because of the nature of our scrambling procedure .
we therefore ignore false triggers in the frequency range @xmath160 hz .
second , the mc procedure destroys real broad - band power at high frequencies , which ( if not accounted for ) will systematically enhance powers in the real pds relative to powers in the fake ( mc ) ones .
we have therefore characterized the broad - band excess power that is present in the real data .
subtracting off the deadtime - corrected pca poisson level ( morgan , remillard & greiner 1997 ) from our real ( averaged ) pds , and logarithmically rebinning , we find that we may fit the residuals with a power law of exponent @xmath70.7 and total normalization 1.0 per cent rms over 12048 hz .
we corrected our real pds powers by dividing by this background level before comparing our powers to the mc results . the average power density spectrum of the first one second of the 31 type i bursts ( see table 1 ) from the rb is presented in fig . [ fig : discovery ] .
the two highest peaks , centred at 306.5 hz and 155 hz , respectively , extend noticeably above the noise level , and are placed in a near - harmonic relationship .
specifically , our monte carlo results indicate that a peak with the strength of the 306.5 hz peak ( or stronger ) is present in 1.8 per cent of the simulated power density spectra , and that a peak with the strength of the 155 hz peak , or stronger , is present in 15.3 per cent of the simulated pdss .
note that although the single - bin 155 hz power is the strongest in the pds , the 306.5 hz candidate signal is more significant because of the strong power in two adjacent bins .
the chances of any pds having two peaks , both stronger than the strength of our 155 hz peak , is 1.2 per cent .
the near - harmonic frequencies of the two peaks may be evidence for a common physical origin . however , the 1 per cent difference in frequencies is uncomfortably large to result from the frequency drifts that have been observed in type i bursts from other sources ( strohmayer & markwardt 1999 ; muno 2000 ) though a connection can certainly not be excluded .
the burst frequencies in other sources drift by less than 1 per cent , total , and exhibit the most pronounced evolution during the tails of the bursts . moreover , the signal we observe here , if real , results from the contribution of several bursts , which would tend to diminish the observable effects of the frequency drift . in the event that the two signals are related
then their combined significance is still dependent upon _ post factum _
assumptions ; we give an illustrative example . if we allow for a @xmath363 per cent shift in the frequency of a harmonic or subharmonic relative to the main signal ( conservative in that the actual shift is 1.1 per cent ) then we would consider two frequency windows , 153158 hz and 594632 hz centred on one - half and double the main 306.5 hz signal frequency , respectively as candidate windows for harmonic detection .
the chance of observing a signal as strong as our 155 hz signal ( or stronger ) within this range is roughly 0.38 per cent ( 15.3 per cent full - search probability multiplied by the size of this restricted range and divided by the size of the full 602048 hz window ) . multiplying the odds for the main signal ( 54:1 ) by the odds for this additional signal ( 259:1 ) would give a combined probability for the dual detection of @xmath37 , corresponding to a 4.0@xmath33 level of confidence .
after we had finished the analysis of the type i bursts of the first five outbursts as described above , and before we were ready to submit our results for publication , two additional outbursts of the rb occurred in march and september 1999 , respectively .
the few bursts from outburst 6 were of generally low quality , but during outburst 7 better bursts were collected and so we performed a full analysis on those 23 bursts ( see table 1 ) .
we found no evidence for modulation near 306 hz , or at any other frequency .
we can set 95 per cent confidence rms upper limits of 8.1 per cent , 7.7 per cent , and 5.0 per cent , respectively , on the average 306 + 307 hz power during the one second before the rise , the one second rise , and the one second after that .
we then analyzed all 61 type i bursts with fast - timing data ( covering the first seven _ rxte _ outbursts ) .
as with our initial search , we found the strongest signal in the 306 + 307 hz bin .
the rms variation in the average pds is [email protected] per cent during the first one second of the bursts ( 5.516 kev energy band ) .
monte carlo simulations indicate , however , that this signal is not significant : 19 per cent of the simulated datasets have signals of equal or greater significance .
if the signal is not real , then it sets our quoted 95 per cent confidence upper limit for a narrow pulsation : 8.8 per cent rms on average .
95 per cent confidence upper limits on the average 306 + 307 hz rms modulation in the one second before and the one second after the first second , for the full burst set , are 5.7 per cent and 2.2 per cent , respectively .
cumming & bildsten ( 2000 ) have shown that the vertical thermal time during mixed hydrogen / helium burning is longer than the wrapping time from the expansion - induced differential rotation , and that the opposite is true for helium - rich bursts : the heat from these bursts can get out faster and , in addition , the shearing frequency is lower .
this makes it easier for a periodic signal from some underlying asymmetry to escape from a helium - rich burst .
muno ( 2000 ) have found that the `` fast '' bursts from bursts with high peak fluxes , short decay times , and radius expansion , indicative of helium burning occur in the same part of the x - ray colour - colour diagram , and all exhibit burst pulsations .
the `` slow '' bursts , likely the result of mixed hydrogen / helium burning with lower peak fluxes , longer decay times , and no radius expansion do not show pulsations , consistent with this picture .
lewin ( 1987 ) showed that the bursts from 1636@xmath753 with radius expansion had substantially shorter rise times than those that did not exhibit radius expansion .
inspired by the above , we restricted our attention to a set of 17 fast - rise bursts from all seven outbursts , and we recomputed an average pds , and performed our signal search again .
these bursts were selected on the basis of their `` merit '' function ( see table 1 ) , defined as ( peak@xmath32/pe)/rise to take into account the expected strength of any burst signal , as well as the distinction between fast- and slow - rise bursts .
note however that we have no particular evidence that these bursts ( or any bursts from the rb ) resulted from helium - rich as opposed to hydrogen - rich burning ; this question will be addressed in future work .
we ignore burst 3 , a burst with low signal - to - noise that has high `` merit '' only because of its unusually short rise time .
we find that the signal is strengthened significantly , with an average power in this burst sample of 4.13 for the two - bin 306.5 hz average ( mean of 34 powers ) .
a signal of this strength or stronger is found in only 0.07 per cent of the monte carlo simulations of this dataset , corresponding to a 3.4-sigma level of confidence .
the strength of the signal corresponds to an average 8.0 per cent rms modulation of the burst flux during these intervals .
similar searches of the burst flux preceding and following the selected one second do not show significant signal power ; the 95 per cent confidence upper limit on the excess average 306 + 307 hz power during these intervals is 0.76 and 0.47 , respectively .
these correspond to average upper limits of 10 per cent and 2 per cent rms , respectively , for modulation of the burst flux during these intervals . likewise ,
a search for power in the soft band ( 2.05.5 kev ) photons from the first second of each burst does not show significant power : the 95 per cent confidence upper limit on the excess average 306 + 307 hz power during these intervals is 0.69 , corresponding to an average rms upper limit of 4.3 per cent for modulation of the soft burst flux .
we performed a separate systematic search for pulsations near 307 hz during the tails of all type i bursts ( except for burst 3 , which is too weak for useful limits to be set ) .
the search parameters here were chosen to maximize signal for the relatively weaker pulsations during the burst tails : we select higher - energy photons only ( 5.516 kev ) , divide the first 16 s of each burst into four 4 s segments , fft and search for the highest power in the 304310 hz range .
we found no significant power , and our mean per - burst 95 per cent confidence upper limits on the strength of any coherent power ( 0.25 hz bins ) in the tails of the bursts are 5 per cent14 per cent rms , relative to the 5.516 kev burst flux , depending on the strength of the burst .
given the weaker signal strengths that are generally observed during burst tails , and the large search space we necessarily considered , we do not consider this non - detection surprising . since our only finding , subsequent to the evidence uncovered in our initial search , was a strengthening of that evidence within an overlapping set of bursts , this result is subject to a posteriori effects and is of uncertain absolute significance .
we have performed a search for burst pulsations from the type i bursts of the rapid burster .
the search was targeted to maximize the chances for detection of a signal like those found in other type i burst sources , where the maximum modulation amplitude occurs at the leading edge of a burst and the pulsations are spectrally hard compared to the burst and persistent emission .
we find evidence at the @xmath298 per cent confidence level for a modulation near 307 hz which , if confirmed , will indicate a spin period for the rapid burster of 3.25 or 6.5 milliseconds , depending on whether the main burst signal is the fundamental or first harmonic of the spin frequency .
we thank deepto chakrabarty , tod strohmayer and fred lamb for helpful discussions .
w.l . gratefully acknowledges nasa support .
was partially supported by nasa via grant nag 5 - 8658 and by the national science foundation under grant no .
phy 94 - 07194 .
l.b . is a cottrell scholar of the research corporation .
strohmayer t.e . ,
swank j.h . , zhang , w. , 1998a , in proceedings of the symposium `` the active x - ray sky : results from bepposax and rossi - xte '' , eds .
l. scarsi , h. bradt , p. giommi , & f. fiore , rome : nuclear physics b proceedings supplements , astro - ph/9801219 | we have searched the rising portion of type i x - ray bursts observed from the rapid burster with the _ rossi x - ray timing explorer _ for the presence of periodicities .
the 95 per cent confidence upper limit on the average root - mean - square variation of near coherent pulsations with a width of @xmath01 hz ( in 602048 hz ) during the first second of the bursts is @xmath18.8 per cent .
we find a possible detection ( @xmath298 per cent significance ) at 306.5 hz .
[ firstpage ] x - rays : bursts x - rays : stars stars : individual : rapid burster |
the determination of the mass of the pulsars psr j1614 - 2230 with @xmath6 @xcite and psr j0348 - 0432 with @xmath7 @xcite has brought the necessity to re - examine many aspects of the physics of neutron stars ( nss ) . in particular
, these observations revived the discussions about whether compact stars are purely hadronic , may have quark - matter cores in their interior , or may be pure strange quark stars @xcite .
in fact , several authors have looked along the years for features that may allow to distinguish unequivocally these stars through e.g. the analysis of the mass - radius relationship and the cooling behaviour since both aspects strongly depend on the microscopic composition . in the case of the mass - radius relationship
a discrimination is not easy because the stellar radius is difficult to determine observationally .
moreover , at present , most observed compact stars have masses in a range where many models for hadronic , hybrid and strange stars overlap in the @xmath8 diagram .
cooling studies rely on the fact that the neutrino emission , the specific heat , the thermal conductivity , and other relevant quantities strongly depend on the microscopic composition
. however , while different models lead to different thermal evolution , there are many degrees of freedom in the problem and a univocal interpretation of observed data is difficult . on the other hand ,
it is now clear that the increase in sensitivity of gravitational wave ( gw ) detectors such as advanced ligo and advanced virgo will bring within the next few years the science of gravitational radiation to a mode of regular astrophysical observation @xcite .
since nss can be conspicuous emitters of gravitational radiation , gws of nss will provide in the near future an important piece of information about several aspects of ns physics .
in particular , transient phenomena involving the excitation of oscillation modes have long been considered as an important tool for the exploration of ns s interiors because several oscillation modes may emit gravitational waves @xcite .
a lot of work has been carried out in the last three decades in order to describe the non - radial oscillatory properties of nss ; however , these studies employed equations of state that in most cases render maximum stellar masses below @xmath9 .
recent observations have shifted the maximum stellar mass above @xmath9 and therefore it is worth re - examining the oscillation spectra because the change in the allowed equations of state may bring new ways to distinguish hadronic , hybrid and strange stars . in this work ,
we study the @xmath10 , @xmath11 and @xmath4 modes of hadronic , hybrid and pure self - bound strange quark stars with maximum masses above 2 @xmath12 within the relativistic cowling approximation .
the paper is organized as follows : in sec .
ii we describe the eos used for the description of hadronic and quark matter . in sec .
iii we present the equations that govern non - radial fluid oscillations of compact stars . in sec .
iv we present our results , and in sec . v a summary and our conclusions .
_ hadronic matter . _
the relativistic mean - field model is widely used to describe hadronic matter in compact stars . in this paper
we adopt the following standard lagrangian for matter composed by nucleons and electrons , @xmath13\psi_{b } + \frac{1}{2}({\partial_\mu \sigma \partial^\mu \sigma - m_{\sigma}^2 \sigma^2 } ) \nonumber \\ & - & \frac{1}{4 } \omega_{\mu \nu}\omega^{\mu \nu}+ \frac{1}{2 } m_{\omega}^2 \omega_\mu \omega^\mu - \frac{1}{4 } \vec \rho_{\mu \nu}.\vec \rho^{\mu \nu } \nonumber \\ & + & \frac{1}{2 } m_\rho^2 \vec \rho_{\mu}. \vec \rho^{\mu } -\frac{1}{3}bm_{n}(g_{\sigma}\sigma)^{3}-\frac{1}{4}c(g_{\sigma}\sigma)^{4 } \nonumber \\ & + & \sum_{l } \bar{\psi}_{l } [ i \gamma_{\mu } \partial^{\mu } - m_{l } ] \psi_{l}. \label{eq1}\end{aligned}\ ] ] leptons @xmath14 are treated as non - interacting and baryons @xmath15 are coupled to the scalar meson @xmath16 , the isoscalar - vector meson @xmath17 and the isovector - vector meson @xmath18 . for more details about the eos obtained from the above lagrangian the reader is referred to e.g. @xcite and references therein .
the five constants in the model are fitted to the bulk properties of nuclear matter @xcite . in this work
we use the parametrizations gm1 @xcite and nl3 @xcite whose coupling constants are shown in table [ table1 ] . at low densities we use the baym , pethick and sutherland ( bps ) model @xcite . .coupling constants for the parametrizations gm1 and nl3 of the hadronic eos .
@xmath19 is the maximum mass of a pure hadronic star for matter composed by nucleons and electrons . [
cols="<,^,^,^",options="header " , ] in this paper we have investigated non - radial fluid oscillations of hadronic , hybrid and strange quark stars with maximum masses above the mass of the recently observed pulsars psr j1614 - 2230 and psr j0348 - 0432 with @xmath0 . for the hadronic equation of state we employed two different parametrizations of a relativistic mean - field model with nucleons and electrons . for
quark matter we have included the effect of strong interactions and color superconductivity within the mit bag model .
the equations of non - radial oscillations were integrated within the cowling approximation in order to determine the frequency of the @xmath10 , @xmath1 and @xmath4-modes .
we find that the fundamental mode is sensitive to the internal composition , but due to the uncertainties in the equations of state , there is an overlapping of the curves corresponding to hadronic , hybrid and strange quark stars for stellar masses larger that @xmath20 . as a consequence
it would be difficult to distinguish hybrid and hadronic stars through the @xmath10-mode frequency , even if the mass or the surface @xmath21 of the object is determined concomitantly with @xmath22 .
however , there are features that in some cases may allow a differentiation between strange stars and hadronic / hybrid stars . for example
, strange stars can not emit gravitational waves with frequency below @xmath23 khz for any value of the mass .
also , sources with a mass in the range @xmath24 emitting a signal in the range @xmath25 khz would be strange stars .
the frequency of the @xmath1 mode is much more affected by the internal composition of the star . for hadronic and hybrid stars , we find that @xmath26 is in the range @xmath27 khz for objects with masses in the range @xmath28 , but for strange quark stars it is always significantly larger than @xmath29 khz .
thus , a compact object emitting a signal above @xmath29 khz could be identified as a strange star even if its mass or gravitational redshift are unknown .
high frequency @xmath4-modes are only present in hybrid stars and fall in the range @xmath5 khz .
thus , they are clearly distinguishable from the fundamental mode , and of low - frequency @xmath4-modes associated with chemical inhomogeneities in the outer layers or thermal profiles .
our results are summarized in table [ table2 ] and show that based on the simultaneous analysis of the frequency of the @xmath10 , @xmath1 and @xmath4-modes it would be possible to discriminate between hadronic , hybrid and strange quark stars .
the spectrum of a pulsating compact star is very rich and therefore the possibility of having other kind of modes in the same frequency range of the above studied fluid modes should be studied in more detail in the light of modern equations of state leading to stellar models compatible with the recent observations of psr j1614 - 2230 and psr j0348 - 0432 .
if other modes are present in the same range , the criterion presented above may be less effective .
additionally , rotation is known to change the frequency range of the modes , but these modifications are not expected to alter qualitatively the scenario presented in this paper .
in particular , it looks quite robust the conclusion that compact objects emitting a signal above 6 khz should be interpreted as strange quark stars and those emitting a signal in the range @xmath30 khz should be interpreted as hybrid stars .
c. vsquez flores and g. lugones acknowledge the financial support received from fapesp - brazil . | one of the key aspects in the comprehension of neutron star interiors is the identification of observables that may impose constraints on the equation of state . at present
, limits are obtained mainly through the study of the mass - radius relationship , the maximum rotational frequency and the cooling behaviour . however , since gravitational wave observatories such as advanced ligo and advanced virgo will open a new window of observation of neutron stars in the very near future , it is crucial identifying observables that may emerge from the analysis of the gravitational wave emission of neutron stars . to this end , we investigate non - radial oscillations of hadronic , hybrid and pure self - bound strange quark stars with maximum masses above the mass of the recently observed massive pulsars psr j1614 - 2230 and psr j0348 - 0432 with @xmath0 . for the hadronic equation of state
we employ different parametrizations of a relativistic mean - field model and for quark matter we use the mit bag model including the effect of strong interactions and color superconductivity .
we find that the first pressure mode for strange quark stars has a very different shape than for hadronic and hybrid stars . for strange quarks
stars the frequency of the @xmath1 mode is larger than @xmath2 khz and diverge at small stellar masses , but for hadronic and hybrid stars it is in the range @xmath3 khz .
this allows an observational identification of strange stars even if extra information such as the mass , the radius or the gravitational redshift of the object is unavailable or uncertain . also , we find as in previous works that the frequency of the @xmath4-mode associated with the quark - hadron discontinuity in a hybrid star is in the range @xmath5 khz for all masses .
thus , compact objects emitting gravitational waves above 6 khz should be interpreted as strange quark stars and those emitting a signal within @xmath5 khz should be interpreted as hybrid stars .
stars : neutron , stars : oscillations , dense matter , gravitational waves |
let @xmath1 be a bounded smooth domain in @xmath2 , and fix @xmath3 . we suppose that the boundary @xmath4 consists of two nonempty open subsets , that is , @xmath5 .
we are concerned with the non - stationary incompressible navier - stokes equations in @xmath1 : u + ( u)u -u + p = f & in@xmath6 , [ a.3 ] + u = 0 & in@xmath6 , with the initial condition @xmath7 here , @xmath8 , @xmath9 , @xmath10 , and @xmath11 denote a viscosity constant , velocity field , pressure , and external force respectively ; @xmath12 means the time derivative @xmath13 . as for the boundary condition
, we impose the adhesive b.c . on @xmath14
: @xmath15 on the other hand , we consider one of the following nonlinear b.c . on @xmath16
: @xmath17 which is called the _ slip boundary condition of friction type _
( sbcf ) , and @xmath18 which is called the _ leak boundary condition of friction type _ ( lbcf ) . here
, @xmath19 is an outer unit normal vector defined on @xmath20 , and we write @xmath21 and @xmath22 .
the stress tensor @xmath23 is given by @xmath24 , @xmath25 being kronecker delta .
we define the stress vector @xmath26 as @xmath27 , and write @xmath28 and @xmath29 .
one can easily see that @xmath30 may depend on @xmath10 , whereas @xmath31 does not .
the function @xmath32 , given on @xmath16 and assumed to be strictly positive , is called a _ modulus of friction_. its physical meaning is the threshold of the tangential ( resp .
normal ) stress .
in fact , if @xmath33 ( resp .
@xmath34 ) then ( [ a.1 ] ) ( resp .
( [ a.2 ] ) ) implies @xmath35 ( resp .
@xmath36 ) , namely , no slip ( resp .
leak ) occurs ; otherwise non - trivial slip ( resp .
leak ) can take place .
we notice that if we make @xmath37 formally , ( [ a.1 ] ) and ( [ a.2 ] ) reduce to the usual slip and leak b.c . respectively . in summary ,
sbcf and lbcf are non - linearized slip and leak b.c
. obtained from introduction of some friction law on the stress .
it should be also noted that the second and third conditions of ( [ a.1 ] ) ( resp .
( [ a.2 ] ) ) are equivalently rewritten , with the notation of subdifferential , as @xmath38 though we will not pursue this matter further , one can refer to @xcite for the navier - stokes equations with general subdifferential b.c .
see also @xcite , which considers the motion of a bingham fluid under b.c . with nonlocal friction against slip .
sbcf and lbcf are first introduced in @xcite for the stationary stokes and navier - stokes equations , where existence and uniqueness of weak solutions are established .
generalized sbcf is considered in @xcite .
the @xmath39-@xmath40 regularity for the stokes equations is proved in @xcite . in terms of numerical analysis ,
@xcite deal with finite element methods for sbcf or lbcf .
applications of sbcf and lbcf to realistic problems , together with numerical simulations , are found in @xcite . for non - stationary cases ,
@xcite study the time - dependent stokes equations without external forces under sbcf and lbcf , using a nonlinear semigroup theory .
the solvability of nonlinear problems are discussed in @xcite for sbcf , and in @xcite for a variant of lbcf .
they use the stokes operator associated with the linear slip or leak b.c . , and do not take into account a compatibility condition at @xmath0 .
the purpose of this paper is to prove existence and uniqueness of a strong solution for ( [ a.3])([a.5 ] ) with ( [ a.1 ] ) or ( [ a.2 ] ) .
we employ the class of solutions of ladyzhenskaya type ( see @xcite ) , searching @xmath41 such that @xmath42 there are several reasons we focus on this strong solution .
first , from a viewpoint of numerical analysis , we would like to construct solutions in a class where uniqueness and regularity are assured also for 3d case .
second , we desire an @xmath43-estimate with respect to time for @xmath10 , which may not be obtained for weak solutions of leray - hopf type ( cf .
* proposition iii.1.1 ) ) .
third , in lbcf , it is not straightforward to deduce a weak solution because of ( [ a.4 ] ) below .
similar difficulty already comes up in the linear leak b.c .
( see @xcite ) the rest of this paper is organized as follows .
basic symbols , notation , and function spaces are given in section 2 . in section 3
, we investigate the problem with sbcf .
the weak formulation is given by a variational inequality , to which we prove uniqueness of solutions . to show existence
, we consider a regularized problem , approximate it by galerkin s method , and derive a priori estimates which allow us to pass on the limit to deduce the desired strong solution . using the compatibility condition that @xmath44 must satisfy sbcf , we can adapt @xmath44 to the regularized problem , which makes an essential point in the estimate .
section 4 is devoted to a study of the problem with lbcf .
there are two major differences from sbcf .
first , as was pointed out in the stationary case ( * ? ? ?
* remark 3.2 ) , we can not obtain the uniqueness of an additive constant for @xmath10 if no leak occurs , namely , @xmath45 on @xmath16 .
second , under lbcf , the quantity @xmath46 need not vanish because @xmath47 can be non - zero .
this fact affects our a priori estimates badly , and we can extract a solution only when the initial leak @xmath48 is small enough . incidentally , if we use the so - called bernoulli pressure @xmath49 instead of standard @xmath10 , the mathematical difficulty arising from ( [ a.4 ] ) are resolved ; nevertheless the leak b.c . involving the bernoulli pressure is known to cause an unphysical effect in numerical simulations ( see @xcite ) .
thereby we employ the usual formulation .
finally , in section 5 we conclude this paper with some remarks on higher regularity .
throughout the present paper , the domain @xmath1 is supposed to be as smooth as required . for the precise regularity of @xmath1 which is sufficient to deduce our main theorems , see remarks [ rem c.10 ] and [ rem d.10 ] .
we shall denote by @xmath50 various generic positive constants depending only on @xmath1 , unless otherwise stated .
when we need to specify dependence on a particular parameter , we write as @xmath51 , and so on .
we use the lebesgue space @xmath52 , and the sobolev space @xmath53 for a nonnegative integer @xmath54 , where @xmath55 means @xmath56 .
@xmath57 is also defined for a non - integer @xmath58 ( e.g. ( * ? ? ?
* definition 1.2 ) ) .
we put @xmath59 . for spaces of vector - valued functions , we write @xmath60 , and so on . the lebesgue and sobolev spaces on the boundary @xmath20 , @xmath14 , or @xmath16 , are also used .
@xmath61 means @xmath62 , and we put @xmath63 , where @xmath64 denotes the surface measure . for a positive function @xmath32 on @xmath16 , the weighted lebesgue spaces @xmath65 and @xmath66 are defined by the norms @xmath67 respectively .
the dual space of @xmath65 is @xmath66 ( see ( * ? ? ?
* lemma 2.1 ) ) . the usual trace operator @xmath68 is defined from @xmath69 onto @xmath70 .
the restrictions @xmath71 , @xmath72 of @xmath73 , are also considered , and we simply write @xmath74 to indicate them when there is no fear of confusion .
in particular , @xmath75 and @xmath76 means @xmath77 and @xmath78 respectively , for @xmath79 . note that @xmath80 and @xmath81 because @xmath19 is smooth on @xmath20 .
the inner product of @xmath82 is simplified as @xmath83 , while other inner products and norms are written with clear subscripts , e.g. , @xmath84 or @xmath85 . for a banach space @xmath86
, we denote its dual space by @xmath87 and the dual product between @xmath87 and @xmath86 by @xmath88 . moreover , we employ the standard notation of bochner spaces such as @xmath89 , @xmath90 . for function spaces corresponding to a velocity and pressure , we introduce closed subspaces of @xmath91 or @xmath56 as follows : @xmath92 to indicate a divergence - free space , we set @xmath93 .
we use the notation @xmath94 , @xmath95 , @xmath96 , and @xmath97 .
let us define bilinear forms @xmath98 , @xmath99 , and a trilinear form @xmath100 by @xmath101 the bilinear forms @xmath102 are continuous , and from korn s inequality ( ( * ? ? ? *
lemma 6.2 ) ) there exists a constant @xmath103 such that @xmath104 concerning the trilinear term @xmath100 , we obtain the following two lemmas .
\(i ) when @xmath105 , for all @xmath106 it holds that @xmath107 \(ii ) when @xmath105 or @xmath108 , for all @xmath106 it holds that @xmath109 in particular , we see from ( [ b.6 ] ) that @xmath110 by the sobolev embedding @xmath111 ( resp .
@xmath112 ) which is valid for @xmath105 ( resp .
@xmath113 ) , combined with an interpolation inequality between @xmath56 and @xmath69 , we have @xmath114 therefore , since @xmath115 by hlder s inequality , we conclude ( [ b.5 ] ) ( resp .
( [ b.6 ] ) ) .
[ lem b.2 ] ( i ) for all @xmath116 and @xmath117 , @xmath118 .
\(ii ) for all @xmath119 and @xmath117 , @xmath120 , and @xmath121 where @xmath122 is a constant depending only on @xmath1 . by integration by parts
, we have @xmath123 from which the conclusion of ( i ) and the first assertion of ( ii ) follow . combining the hlder inequality @xmath124 , the sobolev embedding @xmath125 @xmath126 , and the continuity of the trace operator @xmath127 , we derive ( [ b.10 ] ) . whether @xmath122 is small or not , especially when compared to @xmath128 in ( [ b.3 ] ) , is a very crucial point in our a priori estimates for lbcf ( see proposition [ prop d.1 ] ) .
this is why we distinguish @xmath122 from other constants @xmath50 and do not combine @xmath122 with them . as ( i )
above shows , this problem does not happen when we consider sbcf .
furthermore , we introduce nonlinear functionals @xmath129 and @xmath130 by @xmath131 where @xmath132 is a modulus of friction mentioned in section [ sec a ] .
they are obviously nonnegative and positively homogeneous .
in addition , they are lipschitz continuous when @xmath133 for a.e . @xmath134 .
the followings , which are readily obtainable consequences of standard trace and ( solenoidal ) extension theorems ( ( * ? ? ?
* theorems i.1.5 - 6 , lemma i.2.2 ) , see also ( * ? ? ?
* section 5.3 ) ) , are frequently used in subsequent arguments .
[ lem b.1 ] ( i ) for @xmath135 , it holds that @xmath136 .
\(ii ) for @xmath137 satisfying @xmath138 , there exists @xmath139 such that @xmath140 on @xmath16 and @xmath141 .
[ lem b.10 ] ( i ) for @xmath142 , it holds that @xmath143 .
\(ii ) for @xmath144 ( resp .
@xmath145 ) , there exists @xmath142 ( resp .
@xmath146 ) such that @xmath147 on @xmath16 and @xmath148 .
the definition of @xmath149 given in section [ sec a ] becomes ambiguous when @xmath41 has only lower regularity , say @xmath150 .
thus we propose a redefinition of it , based on the following green formula : @xmath151 [ def b.1 ] let @xmath152 , @xmath153 , @xmath154 , @xmath155 .
if ( [ a.3 ] ) holds in the distribution sense for a.e .
@xmath156 , that is , @xmath157 then we define @xmath158 by @xmath159 where @xmath160 is given by @xmath161 .
the above @xmath162 is well - defined by virtue of the trace and extension theorem .
it coincides with the previous definition when @xmath41 is sufficiently smooth .
in addition , by lemmas [ lem b.1 ] and [ lem b.10 ] , @xmath163 and @xmath164 are characterized by @xmath165 and @xmath166 respectively . by lemma [ lem b.1](ii ) , @xmath167 actually does not depend on @xmath10 .
throughout this section , we assume @xmath168 , @xmath169 , and @xmath170 . further regularity assumptions on these data will be given before theorem [ thm c.1 ] .
in addition , the barrier term @xmath129 is simply written as @xmath171 .
a primal weak formulation of ( [ a.3])([a.5 ] ) with ( [ a.1 ] ) is as follows : pde - sbcf for a.e .
@xmath156 , find @xmath172 such that @xmath173 , @xmath174 , @xmath167 is well - defined in the sense of definition [ def b.1 ] , @xmath175 a.e . on @xmath16 , and @xmath176 a.e .
on @xmath16 .
more precisely ,
@xmath175 " implies that @xmath177 actually belongs to @xmath178 with @xmath179 .
in particular , @xmath180 . throughout this section
, we refer to problem pde - sbcf just as problem pde .
similar abbreviation will be made for other problems .
one can easily find that a classical solution of ( [ a.3])([a.5 ] ) with ( [ a.1 ] ) solves problem pde , and that a sufficiently smooth solution of problem pde is a classical solution .
as the next theorem shows , problem pde is equivalent to the following variational inequality problem .
vi@xmath181-sbcf for a.e .
@xmath156 , find @xmath182 such that @xmath173 , @xmath174 , and @xmath183 [ thm c.2 ] problems @xmath184 and @xmath185 are equivalent .
the precise meaning of equivalent " is that if @xmath41 solves problem pde , @xmath9 solves problem vi@xmath181 ; if @xmath9 solves problem vi@xmath181 , there exists unique @xmath10 such that @xmath41 solves problem pde .
hereafter we will frequently use the terminology equivalent " in a similar sense .
let @xmath186 be a solution of problem pde .
then it follows that @xmath187 using this equation together with @xmath175 and @xmath176 , we have @xmath188 for all @xmath139 .
hence @xmath9 is a solution of problem vi@xmath181 .
next , let @xmath9 be a solution of problem vi@xmath181 .
taking @xmath189 as a test function in ( [ c.4 ] ) , with arbitrary @xmath190 , we find that @xmath191 by a standard theory ( see ( * ? ? ?
* propositions i.1.1 and i.1.2 ) ) , there exists unique @xmath192 such that ( [ b.50 ] ) holds .
therefore , @xmath177 is well - defined , and thus @xmath193 combining this equation with ( [ c.4 ] ) , we obtain @xmath194 and as a result of triangle inequality , @xmath195 for @xmath139 . in view of lemma [ lem b.1](ii )
, this implies that for @xmath137 @xmath196 by a density argument , we can extend @xmath167 to an element of @xmath197 such that @xmath198 since @xmath199 , we conclude @xmath175
. then @xmath176 follows from ( [ c.5 ] ) with @xmath200 .
hence @xmath41 is a solution of problem pde .
we are now in a position to state our main theorem .
we assume : 1
. @xmath201 .
[ s1 ] 2 .
@xmath202 with @xmath203 .
[ s2 ] 3 .
@xmath204 , and sbcf is satisfied at @xmath0 , namely , [ s3 ] @xmath205 note that @xmath206 can be defined in a usual sense .
[ thm c.1 ] under @xmath207@xmath208 , when @xmath105 there exists a unique solution @xmath9 of problem @xmath185 such that @xmath209 when @xmath108 , the same conclusion holds on some smaller time interval @xmath210 .
we call the solution in the above theorem a _ strong solution _ of problem vi@xmath181 .
first we prove the uniqueness of a strong solution
. the existence will be proved in section [ sec 3.4 ] after some additional preparations .
[ prop c.4 ] if @xmath211 and @xmath212 are strong solutions of problem @xmath185 , then @xmath213 .
taking @xmath214 and @xmath215 in ( [ c.4 ] ) for @xmath211 and that for @xmath212 respectively , and adding the resulting two inequalities , for a.e .
@xmath216 we obtain @xmath217 we deduce from ( [ b.6 ] ) , together with young s inequality , that @xmath218 @xmath219 combining ( [ b.3 ] ) and these estimates with ( [ c.12 ] ) , we have @xmath220 by gronwall s inequality , we conclude @xmath221 since @xmath222 .
( note that @xmath223 remains finite because @xmath224 thus @xmath225 . in the case of sbcf here , the last term of ( [ c.12 ] )
vanishes , according to lemma [ lem b.2](i ) .
we did not use that fact because we would like to make our proof of uniqueness remain unchanged when we deal with lbcf .
concerning the associated pressure , we find : [ prop c.5 ] under the assumptions of theorem @xmath226 , let @xmath9 be the strong solution of problem @xmath185 , and @xmath10 be the associated pressure obtained in the proof of theorem @xmath227 . then @xmath228 . for a.e .
@xmath216 , the well - known inf - sup condition ( see ( * ? ? ?
* i.(5.14 ) ) ) , together with ( [ c.2 ] ) , ( [ b.7 ] ) , and @xmath175 a.e . on @xmath16 ,
yields @xmath229 since rhs is bounded uniformly in @xmath230 , @xmath10 is in @xmath231 . to prove the solvability of problem vi@xmath181
, we consider a regularized problem vi@xmath232 , which is shown to be equivalent to a variational equation problem , denoted by problem ve@xmath232 . before stating those problems in detail , for fixed @xmath233 we introduce @xmath234 where @xmath235 is a regularization of @xmath236 having the following properties : 1 .
@xmath237 is a nonnegative convex function . 2 . for all @xmath238
, it holds that @xmath239 3 .
if @xmath240 denotes @xmath241 , for all @xmath238 it holds that @xmath242 4 .
let @xmath243 denote the hessian of @xmath235 , namely , @xmath244 for @xmath245 .
then @xmath243 is semi - positive definite , that is , @xmath246 where @xmath247 means the transpose of @xmath248 .
this is a consequence of the convexity of @xmath235 .
such @xmath235 does exist ; for example , let @xmath235 be given by @xmath249 if @xmath250 , @xmath251 if @xmath252 .
then some elementary computation shows that @xmath235 enjoys all of ( a)(d ) above .
one could use the moreau - yoshida approximation of @xmath236 as @xmath235 , which is considered in @xcite , but it is only in @xmath253 , not in @xmath254 .
since @xmath235 is differentiable , the functional @xmath255 is gteaux differentiable , with its derivative @xmath256 computed by @xmath257 for @xmath258 .
we are ready to state the regularized problems mentioned above .
vi@xmath232-sbcf for a.e .
@xmath156 , find @xmath259 such that @xmath260 , @xmath261 and @xmath262 ve@xmath232-sbcf for a.e .
@xmath156 , find @xmath259 such that @xmath260 , @xmath261 and @xmath263 here , @xmath264 is a perturbation of the original initial velocity @xmath44 .
the way one obtains @xmath264 from @xmath44 is described later . by an elementary observation ( e.g. ( * ? ? ?
* section 3.3 ) or ( * ? ? ?
* lemma 3.3 ) ) , we see that : [ prop c.1 ] problems @xmath265 and @xmath266 are equivalent
. now we focus on the construction of a perturbed initial velocity @xmath264 .
since @xmath267 satisfies sbcf by ( s[s3 ] ) , it follows from the green formula @xmath268 , for @xmath139 , that @xmath269 & \hspace{9 cm } ( \forall v\in v_{n,\sigma } ) .
\label{c.34}\end{aligned}\ ] ] here we consider the regularized problem : find @xmath270 such that @xmath271 & \hspace{9.5 cm } ( \forall v\in v_{n,\sigma } ) , \label{c.25}\end{aligned}\ ] ] which is equivalent to ( cf . proposition [ prop c.1 ] ) @xmath272 by a standard theory of elliptic variational inequalities @xcite , ( [ c.25 ] ) admits a unique solution @xmath264 , which is the perturbation of @xmath44 in question . with this setting ,
we find : [ lem c.1 ] ( i ) when @xmath273 , @xmath274 strongly in @xmath91 .
\(ii ) @xmath275 and @xmath276 \(i ) taking @xmath277 in ( [ c.25 ] ) and @xmath278 in ( [ c.34 ] ) , adding the resulting two inequalities , applying korn s inequality , and using ( [ c.28 ] ) , we conclude @xmath279 \(ii ) since @xmath280 by ( s[s2 ] ) , we can directly apply the regularity result ( * ? ? ?
* lemma 5.2 ) to the elliptic variational inequality ( [ c.25 ] ) , and obtain ( [ c.27 ] ) .
though our @xmath235 and @xmath240 are different from those of @xcite , it makes no difference in the proof of that lemma .
[ rem c.10 ] ( i ) as a result of ( i ) above , for sufficiently small @xmath233 we have @xmath281 \(ii ) concerning the regularity of the domain , @xcite assumes that @xmath14 and @xmath16 are class of @xmath282 and @xmath283 respectively , which is sufficient for our theory as well . in @xcite , dealing with the stationary problem , the author stated that @xmath284 was enough to derive @xmath285 and @xmath286 . however , it turned out that his proof presented there worked only for @xmath287 ; see the errata by the same author .
this is why we have assumed @xmath203 in ( s[s2 ] ) , not @xmath288 . due to proposition [ prop c.1 ] ,
we concentrate on solving problem ve@xmath232 . in doing so ,
we construct approximate solutions by galerkin s method .
since @xmath289 is separable , there exist members @xmath290 , linear independent to each other , such that @xmath291 dense in @xmath91 . here
@xmath292 is fixed , and thus we may assume @xmath293 .
ve@xmath294-sbcf find @xmath295)\ , ( k=1 , . .
, m)$ ] such that @xmath296 defined by @xmath297 satisfies @xmath298 and @xmath299 & \hspace{9 cm } ( k=1 , . . . ,
\label{c.17 } \end{aligned}\ ] ] since @xmath300 , the system of ordinal differential equations ( [ c.17 ] ) admits unique solutions @xmath301)\ , ( k=1 , . . . ,
m)$ ] for some @xmath302 .
the a priori estimate below shows @xmath303 can be taken as @xmath304 , so that we write @xmath304 instead of @xmath303 from the beginning .
[ prop c.2 ] let ( s[s1])(s[s3 ] ) be valid and @xmath292 be small enough so that @xmath305 holds .
\(i ) when @xmath105 , @xmath306 and @xmath307 are bounded independently of @xmath308 and @xmath292 .
\(ii ) when @xmath108 , the same conclusion holds for some smaller interval @xmath309 , which can be taken independently of @xmath308 and @xmath292 .
due to space limitations , we simply write @xmath310 , @xmath311 , @xmath312 , . .
instead of @xmath313 , @xmath314 , @xmath315 , . .
and so on .
\(i ) multiplying ( [ c.17 ] ) by @xmath316 , and adding the resulting equations for @xmath317 , we obtain @xmath318 where we have used lemma [ lem b.2](i ) .
it follows from ( [ b.3 ] ) and ( [ c.18 ] ) that @xmath319 which gives @xmath320 consequently , for @xmath321 , @xmath322 since @xmath323 by assumption , we find that @xmath324 and @xmath325 are bounded by @xmath326 independently of @xmath308 and @xmath292 .
next , we differentiate @xmath327 with respect to @xmath230 , which is possible because @xmath316 s are in @xmath328)$ ] , to deduce @xmath329 multiplying this by @xmath330 , and adding the resulting equations , we obtain @xmath331 where we have again used lemma [ lem b.2](i ) . here , @xmath332 & & \hspace{-3cm}\left| \int_{\gamma_1 } g ' \alpha_\epsilon(u_{m\tau})\cdot u_{m\tau } '
ds \right| & \le \|g'\|_{l^2 } \|u_{m\tau}'\|_{l^2(\gamma_1)^d } \qquad(\text{by } ( \ref{c.18 } ) ) \notag \\ & & \hspace{-3cm}&\le c\|g'\|_{l^2 } \|u_m'\|_{h^1 } \qquad(\text{by lemma \ref{lem b.1}(i ) } ) \notag \\ & & \hspace{-3cm}&\le \frac\alpha6 \|u_m'\|_{h^1}^2 + c\|g'\|_{l^2}^2 , \notag \\[1 mm ] & & \hspace{-3cm}\int_{\gamma_1 } g\;^t u_{m\tau } ' \beta_\epsilon(u_{m\tau } ) u_{m\tau}'\ , ds & \ge 0 , \qquad(\text{by $ g>0 $ and ( \ref{c.19 } ) } ) \notag \\[1 mm ] & & \hspace{-3cm}|(f ' , u_m')| & \le \|f'\|_{l^2 } \|u_m'\|_{h^1 } \le \frac\alpha6 \|u_m'\|_{h^1}^2 + c\|f'\|_{l^2}^2 .
\notag \end{aligned}\ ] ] collecting these estimates , it follows from ( [ c.20 ] ) that for @xmath333 @xmath334 if the second term of lhs is neglected , gronwall s inequality leads to @xmath335 provided that @xmath336 is bounded independently of @xmath308 and @xmath292 , estimate ( [ c.21 ] ) gives the boundedness of @xmath337 because we know that of @xmath325 due to ( [ c.35 ] ) .
then , by ( [ c.23 ] ) and ( [ c.35 ] ) we have @xmath338 which implies @xmath339 is bounded . finally , integrating ( [ c.22 ] )
, we see that @xmath340 is also bounded .
to show the boundedness of @xmath336 , we multiply ( [ c.17 ] ) by @xmath330 , add the resulting equations , and make @xmath0 , arriving at @xmath341 from the construction of @xmath264 , especially ( [ c.26 ] ) , we have @xmath342 furthermore , by schwarz s inequality , sobolev s inequality and ( [ c.27 ] ) , @xmath343 combining these estimates with ( [ c.24 ] ) , we obtain @xmath344 which proves the boundedness of @xmath336 .
this completes the proof of ( i ) .
\(ii ) the discussion before ( [ c.29 ] ) and the observation for @xmath345 are the same as ( i ) .
what changes from the case @xmath105 is that when @xmath108 , instead of ( [ c.29 ] ) , we only have ( by ( [ b.6 ] ) and young s inequality ) @xmath346 for a constant @xmath347 which can be arbitrarily small .
we choose @xmath348 satisfying @xmath349 , and from ( [ c.51 ] ) we obtain @xmath350 .
let @xmath351 , which may depend on @xmath352 at this stage , be the maximum value of @xmath230 such that @xmath353 .
if @xmath354 for all @xmath333 , we set @xmath355 .
since @xmath356 and @xmath357 is continuous with respect to @xmath230 , such @xmath358 does exist , and furthermore if @xmath359 then @xmath360 .
therefore , in place of ( [ c.22 ] ) we obtain @xmath361 which leads to the boundedness of @xmath362 and @xmath363 , together with @xmath364 .
finally , let us prove that @xmath358 is bounded from below independently of @xmath308 and @xmath292 .
in fact , if @xmath359 then we see that @xmath365 since we already know @xmath362 is bounded , we obtain the lower bound for @xmath358 .
this completes the proof of proposition [ prop c.2 ] .
\(i ) a naive computation gives , by ( [ c.18 ] ) , @xmath366 but @xmath367 can not be bounded by @xmath368 in general .
therefore , the perturbation of @xmath44 , which is based on the compatibility condition in ( s[s3 ] ) , is essential in deriving ( [ c.60 ] ) .
\(ii ) if @xmath108 and @xmath11 , @xmath32 , @xmath44 are sufficiently small , we can prove @xmath369 for all @xmath333 , and consequently the existence of a global solution . as a final step for our proof of the existence
, we discuss passing to the limits @xmath370 and @xmath273 .
the proof below is valid for both @xmath113 , except that when @xmath108 we have to replace @xmath304 with @xmath358 given in proposition [ prop c.2 ] .
[ prop c.3 ] ( i ) under the assumptions of proposition @xmath371 , there exists a solution @xmath372 of problem @xmath265 such that all of @xmath373 , @xmath374 , and @xmath375 are bounded independently of @xmath292 .
\(ii ) there exists a strong solution of problem @xmath185 .
\(i ) as a consequence of proposition [ prop c.2 ] , there exists some @xmath372 and a subsequence of @xmath376 , denoted again by @xmath376 , such that @xmath377 , @xmath378 , and when @xmath370 @xmath379 we also find that all of @xmath373 , @xmath374 , and @xmath375 are bounded independently of @xmath292 .
let us prove @xmath372 solves problem vi@xmath232 . by proposition [ prop c.1 ]
, it suffices to show that @xmath372 is a solution of problem ve@xmath232 . multiplying ( [ c.17 ] ) by an arbitrary @xmath380 and integrating over @xmath381
, we obtain @xmath382 it follows from ( * ? ? ?
* theorem iii.2.1 ) that the embedding @xmath383 is compact , so that @xmath384 strongly in @xmath385
. moreover , since the trace operator @xmath386 is compact , @xmath387 strongly in @xmath388 . in particular , @xmath387 a.e . on @xmath389 , and thus the continuity of @xmath390 yields @xmath391 a.e . on @xmath389 .
making @xmath370 in ( [ c.30 ] ) , together with lebesgue s convergence theorem , we see that @xmath392 since @xmath393 , the above equation is valid for all test functions @xmath139 .
hence ( [ c.14 ] ) holds for a.e .
@xmath394 , which implies that @xmath372 is a solution of problem ve@xmath232 .
\(ii ) as a result of ( i ) , there exists some @xmath9 and a sequence @xmath395 @xmath396 , to which we drop the subscript @xmath397 for simplicity , such that @xmath398 , @xmath399 , and when @xmath273 @xmath400 as before , one sees that @xmath401 strongly in @xmath385 and @xmath402 strongly in @xmath403 .
in addition , @xmath404 weakly in @xmath405 , and thus it follows that @xmath406 . following the technique of @xcite , we let @xmath407 be arbitrary . for a.e .
@xmath394 , we take @xmath408 in ( [ c.15 ] ) and integrate the resulting equation over @xmath381 to deduce @xmath409 & \hspace{5 cm } + j_\epsilon(\tilde v_\tau ) - j_\epsilon(u_{\epsilon\tau } ) - ( f , \tilde v - u_\epsilon ) \big\}dt \ge 0 .
\label{c.31 } \end{aligned}\ ] ] in view of ( [ c.28 ] ) , together with triangle inequality and lipschitz continuity of @xmath171 , we have @xmath410 and @xmath411 when @xmath273 .
therefore , taking the lower limit @xmath412 in ( [ c.31 ] ) gives @xmath413 a technique using the lebesgue differentiation theorem allows us to conclude that @xmath9 satisfies ( [ c.4 ] ) at a.e .
@xmath414 ( for more detail , see @xcite ) .
concerning the initial condition , since the trace operator @xmath415 is continuous , lemma [ lem c.1](i ) leads to @xmath416 .
hence @xmath9 is a strong solution of problem vi@xmath181 .
propositions [ prop c.4 ] and [ prop c.3](ii ) complete the proof of theorem [ thm c.1 ] .
throughout this section , we assume @xmath417 , @xmath418 , and @xmath170 . further regularity assumptions on these data will be given before theorem [ thm d.1 ] .
in addition , the barrier term @xmath130 is simply written as @xmath171 .
a primal weak formulation of ( [ a.3])([a.5 ] ) with ( [ a.2 ] ) is as follows : pde - lbcf for a.e .
@xmath156 , find @xmath419 such that @xmath173 , @xmath174 , @xmath420 is well - defined in the sense of definition [ def b.1 ] , @xmath421 a.e . on @xmath16 , and @xmath422 a.e .
on @xmath16 .
more precisely ,
@xmath421 " implies that @xmath423 actually belongs to @xmath66 with @xmath424 .
in particular , @xmath425 . throughout this section
, we refer to problem pde - lbcf just as problem pde .
similar abbreviation will be made for other problems .
next , as in sbcf , we propose : vi - lbcf for a.e .
@xmath156 , find @xmath419 , such that @xmath173 , @xmath174 and ( u , v - u ) + a_0(u , v - u ) + a_1(u , u , v - u ) + b(v - u , p ) + + j(v_n ) - j(u_n ) ( f , v - u ) & @xmath426 , [ d.3 ] + b(u , q ) = 0 & @xmath427 .
vi@xmath181-lbcf for a.e .
@xmath156 , find @xmath428 such that @xmath173 , @xmath174 and @xmath429 unlike the case of sbcf , problem vi@xmath181 is not exactly equivalent to problem pde , as is shown in the following theorem .
\(i ) if @xmath41 solves problem @xmath184 , then @xmath9 solves problem @xmath185 .
\(ii ) if @xmath9 solves problem @xmath185 , then there exists at least one @xmath10 such that @xmath41 solves problem @xmath184 .
if another @xmath430 satisfies the same condition , then for a.e .
@xmath134 there exists a unique @xmath431 such that @xmath432 \(iii ) in ( ii ) , if we assume furthermore @xmath433 , then @xmath434 .
namely , the associated pressure is uniquely determined .
\(i ) this can be proved by the same way as theorem [ thm c.2 ] .
\(ii ) for a.e . @xmath394 and @xmath190 , it follows from ( [ d.5 ] ) that @xmath435 , and thus there exists unique @xmath436 such that @xmath437 according to definition [ def b.1 ] , @xmath438 is well - defined , so that @xmath439 substituting this equation into ( [ d.5 ] ) , we obtain @xmath440 for all @xmath441 .
it follows from lemma [ lem b.10](ii ) that @xmath442 the hahn - banach theorem allows us to extend @xmath443 to a linear functional @xmath444 satisfying the same inequality as above for all @xmath445
. therefore , @xmath446 and @xmath421 .
in addition , @xmath447 follows .
since @xmath448 vanishes on @xmath449 , there exists a constant @xmath450 such that @xmath451 .
now , by setting @xmath452 , it follows that @xmath420 given above actually equals @xmath453 and that @xmath454 solves problem pde .
relation ( [ d.9 ] ) can be verified by a similar argument .
\(iii ) since @xmath455 , the assumption @xmath433 implies that there exist subsets @xmath456 of @xmath16 with positive @xmath457 dimensional lebesgue measure satisfying @xmath458 on @xmath459 and @xmath460 on @xmath461 . because @xmath421 and @xmath447 on @xmath16 , @xmath462 on @xmath459 and @xmath463 on @xmath461 .
hence @xmath450 in ( [ d.9 ] ) can not be other than zero . since @xmath421
, @xmath450 is no more than @xmath464 nor less than @xmath465 .
let us state our main theorems for the case of lbcf .
as in sbcf , some compatibility condition is necessary ; it is rather complicated because normal stress at @xmath0 involves a pressure at @xmath0 , which is not given as a data .
the precise description is as follows : we say that lbcf is satisfied at @xmath0 if @xmath466 and there exists @xmath467 such that @xmath468 we remark that a similar compatibility condition appears in nonlinear semigroup approaches ( see @xcite ) .
furthermore , in order to overcome a difficulty arising from ( [ a.4 ] ) , we need no - leak condition at @xmath0 , that is , @xmath469 on @xmath16 . in view of ( [ d.53 ] ) , this is automatically satisfied if @xmath470 on @xmath16 .
examining our proof of the a priori estimates carefully , one finds that this assumption can be weaken to the condition that @xmath48 is sufficiently small . including what we have discussed above
, we assume the followings : 1 .
[ l1 ] 2 .
@xmath202 with @xmath203 .
[ l2 ] 3 .
@xmath471 , and lbcf is satisfied at @xmath0 .
[ l3 ] 4 .
@xmath472 a.e .
on @xmath16 .
[ l4 ] [ thm d.1 ] under ( l[l1])(l[l4 ] ) above , there exists a unique solution @xmath9 of problem @xmath185 on some interval @xmath309 , with @xmath473 , such that @xmath474 the uniqueness can be proved by the same way as proposition [ prop c.4 ] .
we can also obtain @xmath475 by a similar manner to proposition [ prop c.5 ] , using the rather infamous inf - sup condition ( see ( * ? ? ?
* lemma 2.2 ) ) @xmath476 the rest of this section is devoted to the proof of the existence .
to state regularized problems , for fixed @xmath233 we introduce @xmath234 where @xmath235 is a function such that 1 .
@xmath477 is a nonnegative convex function . 2 . for all @xmath478 , @xmath479 .
if @xmath240 denotes @xmath480 , for all @xmath478 it holds that @xmath481 4 .
let @xmath482 .
then @xmath483 ( due to the convexity of @xmath235 ) .
such @xmath235 does exist ; for example , if we define @xmath249 if @xmath250 , @xmath251 if @xmath252 , then this @xmath235 enjoys all of ( a)(d ) above .
since @xmath235 is differentiable , the functional @xmath255 is gteaux differentiable , with its derivative @xmath484 computed by @xmath485 now let us state the regularized problems .
vi@xmath232-lbcf for a.e .
@xmath156 , find @xmath486 such that @xmath260 , @xmath261 and @xmath487 ve@xmath232-lbcf for a.e .
@xmath156 , find @xmath486 such that @xmath260 , @xmath261 and @xmath488 as in proposition [ prop c.1 ] , problems @xmath265 and @xmath266 are equivalent .
the construction of the perturbed initial velocity @xmath264 is similar to that of sbcf .
in fact , since lbcf holds at @xmath0 by ( l[l3 ] ) , the green formula leads to @xmath489 \ge & ( -\nu\delta u_0 + \nabla p_0 , v - u_0 ) \hspace{5 cm } ( \forall v\in v_{\tau,\sigma}).\end{aligned}\ ] ] we consider the regularized problem : find @xmath490 such that @xmath491 \ge & ( -\nu\delta u_0 + \nabla p_0 , v - u_0^\epsilon ) \hspace{5 cm } ( \forall v\in v_{\tau,\sigma } ) , \label{d.11}\end{aligned}\ ] ] which is equivalent to ( cf .
proposition [ prop c.1 ] ) @xmath492 the elliptic variational inequality ( [ d.11 ] ) admits a unique solution @xmath264 , which is the perturbation of @xmath44 in question . with this setting ,
we find : [ lem d.1 ] ( i ) when @xmath273 , @xmath274 strongly in @xmath91 .
in particular , it follows that @xmath493 in @xmath62 .
\(ii ) @xmath275 and @xmath494 \(i ) is proved by the same way as lemma [ lem c.1](i ) . since @xmath280 by ( l[l3 ] ) , ( ii ) is a direct consequence of ( * ? ? ?
* lemma 4.1 ) . by ( i ) and ( l[l4 ] ) , for sufficiently small @xmath233 we have @xmath495 where @xmath128 and @xmath122 are the constants in ( [ b.3 ] ) and ( [ b.10 ] ) respectively .
[ rem d.10 ] as in sbcf , if @xmath14 is @xmath282 and @xmath16 is @xmath283 , then we can apply lemma 4.1 of @xcite . on the other hand , @xmath288 ,
stated in @xcite , is actually insufficient to deduce the @xmath39-@xmath40 regularity ( see the errata of @xcite ) . to solve problem @xmath266 ,
let us construct approximate solutions by galerkin s method .
since @xmath496 is separable , there exist @xmath497 , linear independent to each other , such that @xmath498 dense in @xmath91 .
here we may assume @xmath293 .
ve@xmath294-lbcf find @xmath295)\ , ( k=1 , . .
, m)$ ] such that @xmath499 defined by @xmath297 satisfies @xmath298 and @xmath500 & \hspace{9 cm } ( k=1 , . . . ,
\label{d.10 } \end{aligned}\ ] ] since @xmath501 , there exist unique solutions @xmath301)\ , ( k=1 , . .
, m)$ ] for some @xmath303 , which may depend on @xmath308 and @xmath292 at this stage .
[ prop d.1 ] assume ( l[l1])(l[l4 ] ) , and let @xmath233 be sufficiently small so that @xmath502 holds .
then there exists some interval @xmath309 such that @xmath503 and @xmath504 are uniformly bounded with respect to @xmath308 and @xmath292 . here
, @xmath358 is independent of @xmath308 and @xmath292 . due to space limitations
, we sometimes simply write @xmath505 , instead of @xmath506 , when there is no fear of confusion .
first we consider the case @xmath105 . multiplying ( [ d.10 ] ) by @xmath316 for @xmath317 , adding them , using ( [ b.3 ] ) , ( [ b.10 ] ) and ( [ d.12 ] )
, we obtain @xmath507 since @xmath508 is continuous with respect to @xmath230 and ( [ d.20 ] ) holds , there exists a maximum value @xmath509 $ ] of @xmath230 such that @xmath510 if this inequality holds for all @xmath511 , we take @xmath512 . noting @xmath513 , we find from ( [ d.13 ] ) that @xmath514 hence @xmath515 is bounded independently of @xmath308 , @xmath292 .
next , differentiating ( [ d.10 ] ) , multiplying the resulting equation by @xmath330 , and adding them , we obtain @xmath516 here , we estimate each term in ( [ d.14 ] ) as follows : @xmath517 @xmath518 collecting these estimates , we derive from ( [ d.14 ] ) that for @xmath519 @xmath520 combining the technique used in proposition [ prop c.2 ] with ( [ d.17 ] ) and ( [ d.55 ] ) , we observe that @xmath521 , @xmath522 , and @xmath523 are bounded by @xmath524 .
it remains to show that @xmath525 is bounded from below independently of @xmath526 .
if @xmath527 and thus @xmath528 , we can extend @xmath357 beyond @xmath529 and repeat the above discussion until we reach either @xmath530 in the former case @xmath531 . in the latter case ,
we have @xmath532 hence @xmath525 is bounded from below , and we complete the proof for @xmath105 .
second let us consider the case @xmath108 .
what changes from @xmath105 is that ( [ d.15 ] ) is replaced with @xmath533 where @xmath534 can be arbitrarily small .
we choose @xmath534 satisfying @xmath535 , so that @xmath536 by virtue ( [ d.20 ] ) .
let @xmath537 be the maximum value of @xmath538 $ ] such that @xmath539 .
if this inequality holds for all @xmath538 $ ] , we set @xmath540 .
such @xmath537 does exist , and if @xmath541 then @xmath542 .
therefore , setting @xmath543 , instead of ( [ d.16 ] ) we get @xmath544 as a consequence , we see that @xmath545 , @xmath363 , @xmath546 are bounded by @xmath524 . now ,
if @xmath547 or @xmath541 then @xmath358 are bounded from below as follows : @xmath548 when @xmath549 and @xmath540 , we can extend @xmath357 beyond @xmath529 and repeat the above discussion .
this completes the proof of proposition [ prop d.1 ] .
the last step of the proof , namely , passing to the limits @xmath370 and @xmath273 can be carried out by the same way as proposition [ prop c.3 ] , with @xmath19 replaced by @xmath550 and vice versa .
this proves that a solution of problem vi@xmath181 exists , which , combined with the uniqueness result , completes the proof of theorem [ thm d.1 ] . at first glance
one may think theorem [ thm d.1 ] , where we get only a time - local solution in spite of a smallness assumption on @xmath44 even if @xmath105 , is too poor .
however , in view of the fact that we obtain only time - local solutions in 2d case under the linear leak b.c .
( see ( * ? ? ?
* theorem 6 ) or @xcite ) , such limitations can not be avoided to some extent . under additional smallness assumptions on the data @xmath551
, we can derive global existence results for both @xmath105 and @xmath108 .
by the discussion presented above , we have established the existence and uniqueness , while we did not get in touch with higher regularity such as @xmath552 this is because some regularity results for the elliptic cases are not available .
for instance , problem vi@xmath553-sbcf is rewritten as @xmath554 with @xmath555 for some @xmath556 . if we prove this elliptic variational inequality has a unique solution in @xmath557 when @xmath556 , then a technique similar to ( * ? ? ?
* theorems iii.3.6 and iii.3.8 ) allows us to deduce @xmath558 .
thereby , we need to extend the regularity theory of @xcite to cases @xmath559 .
the author would like to thank for professor norikazu saito for encouraging him through valuable discussions .
this work was supported by grant - in - aid for jsps fellows and crest , jst . | strong solutions of the non - stationary navier - stokes equations under non - linearized slip or leak boundary conditions are investigated .
we show that the problems are formulated by a variational inequality of parabolic type , to which uniqueness is established . using galerkin s method and deriving a priori estimates , we prove global and local existence for 2d and 3d slip problems respectively .
for leak problems , under no - leak assumption at @xmath0 we prove local existence in 2d and 3d cases .
compatibility conditions for initial states play a significant role in the estimates . |
in many practical networks , including sensor networks and vehicular networks , receivers demand a function of the messages generated by the sources that are distributed across the network rather than the generated messages . this situation is studied in the framework of network computing @xcite .
the classical network coding model of ahlswede , cai , li , and yeung @xcite can be viewed as a the special case of network computing in which the function to be computed at the receivers corresponds to a subset of the source messages and communication occurs over a network with noiseless links . in the same noiseless set up of @xcite , we consider the scenario in which a set of source nodes generate messages over a finite field and a single receiver node computes a linear function of these messages .
we ask whether this linear function can be computed by performing linear coding operations at the intermediate nodes . in multiple - receiver networks , if each receiver node demands a subset of the source messages ( which is an example of a linear function ) , then dougherty , freiling , and zeger @xcite showed that linear codes are not sufficient to recover the source messages .
similarly , if each receiver node demands the sum of the source messages , then ray and dei @xcite showed that linear codes are also not sufficient to recover the source messages .
in contrast , in single - receiver networks linear codes are sufficient for both the above problems and a simple cut - based condition can be used to test whether a linear solution exists .
our contribution is as follows .
we extend above results investigating if a similar cut - based condition guarantees the existence of a linear solution when the receiver node demands an arbitrary linear function of the source messages .
we identify two classes of functions , one for which the cut - based condition is sufficient for solvability and the other for which it is not .
these classes are complements of each other when the source messages are over the binary field . along the way
, we develop an algebraic framework to study linear codes and provide an algebraic condition to test whether a linear solution exists , similar to the one given by koetter and mdard @xcite for classical network coding .
the paper is organized as follows .
we formally introduce the network computation model in section [ sec : intro ] . in section
[ sec : main ] we develop the necessary algebraic tools to study linear codes and introduce the cut - based condition . in section [ sec : linearfunctions ] , we show the main results for the two classes of functions . section [ sec : conclusions ] concludes the paper , mentioning some open problems . in this paper , a _ network _
@xmath0 consists of a finite , directed acyclic multigraph @xmath1 , a set of _ source nodes _
@xmath2 , and a _ receiver _ @xmath3 .
such a network is denoted by @xmath4 .
we use the word graph " to mean a multigraph , and network " to mean a single - receiver network .
we assume that @xmath5 , and that the graph @xmath6 contains a directed path from every node in @xmath7 to the receiver @xmath8 .
for each node @xmath9 , let @xmath10 and @xmath11 denote the in - edges and out - edges of @xmath12 respectively .
we also assume ( without loss of generality ) that if a network node has no in - edges , then it is a source node .
we use @xmath13 to denote the number of sources @xmath14 in the network .
an _ alphabet _ @xmath15 is a nonzero finite field . for any positive integer @xmath16 , any vector @xmath17 , and any @xmath18 , let @xmath19 denote the @xmath18-th component of @xmath20 . for any index set @xmath21 with @xmath22 , let @xmath23 denote the vector @xmath24 .
network computing _ problem consists of a network @xmath0 , a source alphabet @xmath15 , and a _ target function _
@xmath25 where @xmath26 is the _ decoding alphabet_. a target function @xmath27 is _ linear _ if there exists a matrix @xmath28 over @xmath15 such that @xmath29 where ` @xmath30 ' denotes matrix transposition . for linear target functions the decoding alphabet is of the form @xmath31 , with @xmath32 . without loss of generality , we assume that @xmath28 is full rank ( over @xmath15 ) and has no zero columns .
for example , if @xmath28 is the @xmath33 identity matrix , then the receiver demands the complete set of source messages , and this corresponds to the classical network coding problem .
on the other hand , if @xmath28 is the row vector of @xmath34 s , then the receiver demands a sum ( over @xmath15 ) of the source values .
let @xmath35 be a positive integer . given a network @xmath0 with source set @xmath36 and alphabet @xmath15 , a _ message generator _
is a mapping @xmath37 for each source @xmath38 , @xmath39 is called a _ message vector _ and it can be viewed as an element of @xmath40 ( rather than as a vector ) . a _ linear network code _ in a network @xmath0 consists of the following : * every edge @xmath41 _ carries an element _ of @xmath40 and this element
is denoted by @xmath42 . for any node @xmath43 and any out - edge @xmath44
, the network code specifies an _ encoding function _
@xmath45 of the form : @xmath46 where @xmath47 for all @xmath48 . * the _ decoding function _
@xmath49 outputs a vector of length @xmath50 whose @xmath51-th component is of the form : @xmath52 where @xmath53 for all @xmath54 .
the arithmetic in and is performed over @xmath40 . in this paper , by a _ network code _ , we always mean a linear network code . in the literature ,
the class of network codes we define here is referred to as _ scalar linear codes .
_ these codes were introduced and studied in @xcite .
a more general class of linear codes over @xmath40 were defined and studied in @xcite .
depending on the context , we may view @xmath42 as a vector of length-@xmath35 over @xmath55 or as an element of @xmath40 . without explicit mention , we use the fact that the addition of @xmath56 as elements of a finite field coincides with their sum as elements of a vector space over @xmath55 .
furthermore , we also view @xmath55 as a subfield of @xmath40 without explicitly stating the inclusion map .
let @xmath57 denote the vectors carried by the in - edges of the receiver .
[ def : sol ] a linear network code over @xmath40 is called _ a linear solution for computing @xmath27 in @xmath0 _ ( or simply a _ linear solution _ if @xmath27 and @xmath0 are clear from the context ) if the decoding function @xmath49 is such that for every message generator @xmath58 , @xmath59 each source generates @xmath35 symbols over @xmath55 ( viewing @xmath40 as a vector space over @xmath55 ) and the decoder computes the target function @xmath27 for each set of source symbols .
a set of edges @xmath60 is said to _ separate _
sources @xmath61 from the receiver @xmath8 , if for each @xmath62 , every path from @xmath63 to @xmath8 contains at least one edge in @xmath64 .
a set @xmath65 is said to be a _ cut _ if it separates at least one source from the receiver .
let @xmath66 denote the set of all cuts in network @xmath0 . for any matrix @xmath67 ,
let @xmath68 denote its @xmath18-th column .
for an index set @xmath69 , let @xmath70 denote the @xmath71 submatrix of @xmath72 obtained by choosing the columns of @xmath72 indexed by @xmath73 .
if @xmath64 is a cut in a network @xmath0 , we define the set @xmath74 finally , for any network @xmath0 and matrix @xmath28 , we define @xmath75
linear solvability for the classical network coding problem was shown to be equivalent to the existence of a non - empty algebraic variety in @xcite . in the following , we present an analogous characterization for computing linear functions , providing an algebraic test to determine whether a linear solution for computing a linear function exists .
the reverse problem of constructing a multiple - receiver network coding ( respectively , network computing ) problem given an arbitrary set of polynomials , which is solvable if and only if the corresponding set of polynomials is simultaneously solvable is considered in reference @xcite ( respectively , @xcite ) .
we begin by giving some definitions and stating a technical lemma , followed by the main theorem below . for any edge @xmath76 ,
let @xmath77 and @xmath78 .
associated with a linear code over @xmath40 , we define the following three types of matrices : * for each source @xmath79 , define the @xmath80 matrix @xmath81 as follows : @xmath82 * similarly define the @xmath83 matrix @xmath84 as follows : @xmath85 * define the @xmath86 matrix @xmath87 as follows : @xmath88 since the graph @xmath6 associated with the network is acyclic , we can assume that the edges @xmath89 are ordered such that the matrix @xmath87 is strictly upper - triangular .
let @xmath90 denote the identity matrix of suitable dimension .
consider a network @xmath0 with alphabet @xmath55 and consider a linear code over @xmath40 with associated matrices @xmath91 and @xmath87 .
for every @xmath92 , define the @xmath93 matrix @xmath94 now let @xmath95 be a vector containing all the non - zero entries of the matrices @xmath96 , and let @xmath97 ( respectively , @xmath98 ) be a vector containing all the non - zero entries of the matrix @xmath84 ( respectively , @xmath87 ) . by abusing notation , depending on the context
we may view @xmath99 , @xmath100 , @xmath101 as elements of @xmath40 or as indeterminates .
thus , each of the matrices defined above may either be a matrix over @xmath40 or a matrix over the polynomial ring @xmath102}$ ] .
the context should make it clear which of these two notions is being referred to at any given point .
[ lemma : koetter ] the following two statements hold : 1 .
the matrix @xmath103 has a polynomial inverse with coefficients in @xmath104}$ ] , the ring of polynomials in the variables constituting @xmath98 .
the decoding function can be written as @xmath105 the first assertion is a restatement of ( * ? ? ?
* lemma 2 ) and the second assertion follows from ( * ? ? ?
* theorem 3 ) .
let @xmath106 be a polynomial ring .
the ideal generated by a subset @xmath107 and denoted by @xmath108 is the smallest ideal in @xmath106 containing @xmath109 .
let @xmath0 be a network with alphabet @xmath55 .
let @xmath110}$ ] and @xmath111 .
consider a linear network code for computing the linear function corresponding to @xmath28 in @xmath0 and the associated matrices @xmath112 over @xmath106 and define @xmath113 let @xmath114 denote the ideal generated by the elements of @xmath115 in the ring @xmath106 .
more formally , let @xmath116 the polynomials @xmath117 are referred to as the _ generating polynomials _ of the ideal @xmath114 .
we denote the grob@xmath118ner basis of an ideal generated by subset @xmath107 of a polynomial ring @xmath106 by @xmath119 .
the following theorem is a consequence of hilbert nullstellensatz ( see ( * ? ? ?
* lemma viii.7.2 ) and the remark after ( * ? ? ?
* proposition viii.7.4 ) ) .
[ th : gbasiscondition ] consider a network @xmath0 with alphabet @xmath55 and the linear target function @xmath27 corresponding to a matrix @xmath120 .
there exists an @xmath121 and a linear solution over @xmath40 for computing @xmath27 in @xmath0 if and only if @xmath122 . from lemma [ lemma : koetter ]
, the vector computed at the receiver can be written as @xmath123 on the other hand , to compute the linear function corresponding to @xmath28 , the decoding function must satisfy @xmath124}\end{aligned}\ ] ] it follows that the encoding coefficients in a linear solution must be such that @xmath125}\end{aligned}\ ] ] if we view the coding coefficients as variables , then it follows that a solution must simultaneously solve the generating polynomials of the corresponding ideal @xmath114 . by (
* lemma viii.7.2 ) , such a solution exists over the algebraic closure @xmath126 of @xmath55 if and only if @xmath127}$ ] .
furthermore , @xmath127}$ ] if and only if @xmath122 .
moreover , a solution exists over the algebraic closure @xmath126 of @xmath55 if and only if it exists over some extension field @xmath40 of @xmath55 and the proof is now complete .
it is clear that the set of linear functions that can be solved in a network depends on the network topology .
it is easily seen that a linear solution for computing a linear target function corresponding to @xmath111 exists only if the network @xmath0 is such that for every @xmath128 , the value of the cut @xmath129 is at least the rank of the submatrix @xmath130 ( recall that @xmath131 is the index set of the sources separated by the cut @xmath64 ) .
this observation is stated in the following lemma which is an immediate consequence of the cut - based bound in ( * ? ? ?
* theorem 2.1 ) .
[ lemma : ub ] for a network @xmath0 , a necessary condition for the existence of a linear solution for computing the target function corresponding to @xmath111 is @xmath132 we now consider two special cases .
first , consider the case in which the receiver demands all the source messages .
the corresponding @xmath28 is given by the @xmath33 identity matrix @xmath90 and the condition @xmath133 reduces to @xmath134 i.e. , the number of edges in the cut be at least equal to the number of sources separated by the cut .
second , consider the case in which the receiver demands the sum of the source messages .
the corresponding matrix @xmath28 is an @xmath135 row vector and the requirement that @xmath133 reduces to @xmath136 i.e. , all the sources have a directed path to the receiver . for both of the above cases , the cut condition in lemma [ lemma : ub ] is also sufficient for the existence of a solution .
this is shown in ( * ? ? ?
* theorem 3.1 and theorem 3.2 ) and is reported in the following lemma : [ th : extremematrix ] let @xmath137 . for a network @xmath0 with the linear target function
@xmath27 corresponding to a matrix @xmath120 , a linear solution exists if and only if @xmath133 .
the focus in the rest of the paper is to extend above results to the case @xmath138 by using the algebraic test of theorem [ th : gbasiscondition ] .
in the following , we first define an equivalence relation among matrices and then use it to identify a set of functions that are linearly solvable in every network satisfying the condition @xmath133 .
we then construct a linear function outside this set , and a corresponding network with @xmath133 , on which such a function can not be computed with linear codes .
finally , we use this example as a building block to identify a set of linear functions for which there exist networks satisfying the min - cut condition and on which these functions are not solvable .
notice that for a linear function with matrix @xmath111 , each column of @xmath72 corresponds to a single source node .
hence , for every @xmath33 permutation matrix @xmath139 , computing @xmath140 is equivalent to computing @xmath141 after appropriately renaming the source nodes .
furthermore , for every @xmath142 full rank matrix @xmath143 over @xmath55 , computing @xmath140 is equivalent to computing @xmath144 .
these observations motivate the following definition : let @xmath145 and @xmath146 .
we say @xmath147 if there exist an invertible matrix @xmath143 of size @xmath142 and a permutation matrix @xmath139 of size @xmath33 such that @xmath148 , and @xmath149 if such @xmath143 and @xmath139 do not exist .
since @xmath72 is assumed to be a full rank matrix , @xmath139 can be chosen such that the first @xmath50 columns of @xmath150 are linearly independent .
let @xmath151 denote the first @xmath50 columns of @xmath150 . by choosing @xmath152 , we have @xmath153 where @xmath154 is an @xmath155 matrix .
so for an arbitrary linear target function @xmath27 and an associated matrix @xmath28 , there exists an @xmath155 matrix @xmath154 such that @xmath156 . without loss of generality
, we assume that each column of @xmath72 associated with a target function is non - zero .
[ th : solvable ] consider a network @xmath0 with a linear target function corresponding to a matrix @xmath157 ( i.e. , @xmath158 ) . if @xmath159 where @xmath12 is a column vector of units , then a necessary and sufficient condition for the existence of a linear solution is @xmath133 .
let @xmath160 . the ` necessary ' part is clear from lemma [ lemma : ub ] .
we now focus on the ` sufficiency ' part .
notice that for each @xmath161 , the matrix @xmath162 ( computed as in ) is a row vector of length @xmath163 . stack these @xmath13 row vectors to form an @xmath164 matrix @xmath165 as follows , @xmath166 let @xmath167 denote the @xmath168 submatrix of @xmath165 obtained by deleting its @xmath18-th row .
claim @xmath34 _ : the matrix @xmath169 has a non - zero determinant over the ring @xmath110}$ ] .
claim @xmath170 _ : for each @xmath171 , we have @xmath172 .
+ by claim 1 and the sparse zeros lemma @xcite , @xcite , it follows that that there exists some @xmath121 such that the variables @xmath173 can be assigned values over @xmath40 so that the @xmath164 matrix @xmath174 is such that any of its @xmath175 submatrices @xmath176 obtained by deleting the @xmath18-th row in @xmath165 , is full rank over @xmath40 .
define two @xmath177 diagonal matrices @xmath178 and @xmath179 such that for @xmath180 @xmath181 now define the following matrices over @xmath40 : @xmath182 by by claim 2 it follows that @xmath183 exists .
if the matrices @xmath184 , and @xmath185 define a linear network code , then by lemma [ lemma : koetter ] , the vector received by @xmath8 can be written as , @xmath186 where , @xmath187 we have @xmath188 } \nonumber \\ & = \begin{pmatrix } a_1 ( i - f)^{-1 } b^{t } m_{(s)}^{-1 } \\ a_2 ( i - f)^{-1 } b^{t } m_{(s)}^{-1 } \\ \vdots \\ a_s ( i - f)^{-1 } b^{t } m_{(s)}^{-1 } \end{pmatrix } \ d^{-1 } u & { & [ \mbox{from $ \left ( ( m_{(s)}^{t})^{-1}\right)^{t } = m_{(s)}^{-1}$ } ] } \nonumber \\ & = \begin{pmatrix } i \\ a_s ( i - f)^{-1 } b^{t } m_{(s)}^{-1 } \end{pmatrix } \ d^{-1 } u & { & [ \mbox{from construction of $ m_{(s)}$ } ] } \label{eq : nulling } \\
\begin{pmatrix } \bar{a}_1 ( i - f)^{-1 } \bar{b}^{t } \\ \bar{a}_2 ( i - f)^{-1 } \bar{b}^{t } \\ \vdots \\ \bar{a}_s ( i - f)^{-1 } \bar{b}^{t } \end{pmatrix } & = \begin{pmatrix } u^{-1 } d \\ a_s ( i - f)^{-1 } b^{t } m_{(s)}^{-1 } \end{pmatrix } \
d^{-1 } u & { & [ \mbox{from \eqref{eq : redefinea } and \eqref{eq : nulling } } ] } \nonumber \\ & = \begin{pmatrix } u^{-1 } \\ { \mathbf}{1}^{t } \end{pmatrix } \ u & { & [ \mbox{from \eqref{eq : diagonalmatrices } } ] } \nonumber \\ & = \begin{pmatrix } i \\ { \mathbf}{1}^{t } u \end{pmatrix } \nonumber \\ & = \begin{pmatrix } i \\
u^{t } \end{pmatrix } \label{eq : solmstar } \\
\bar{m}^{t } & = \begin{pmatrix } i \
u \end{pmatrix}. & { & [ \mbox{from \eqref{eq : defmstar } and \eqref{eq : solmstar } } ] } \label{eq : solmstar1 } \ ] ] by substituting in , we conclude that the receiver computes the desired linear function by employing the network code defined by the encoding matrices @xmath189 , @xmath185 , and @xmath87 .
the proof of the theorem is now complete for the case when @xmath160 .
if @xmath190 , then there exists a full - rank matrix @xmath143 and a column vector @xmath191 of non - zero elements over @xmath55 such that @xmath192}\end{aligned}\ ] ] since a full - rank linear operator preserves linear - independence among vectors , for every such full - rank matrix @xmath143 , we have @xmath193 equation implies that @xmath194 . since @xmath195 ,
from the first part of the proof , there exist an @xmath121 and coding matrices @xmath96 , @xmath87 , and @xmath84 over @xmath40 such that the receiver can compute the linear target function corresponding to @xmath196 if and only if @xmath133 .
it immediately follows that by utilizing a code corresponding to the coding matrices @xmath96 , @xmath87 , and @xmath197 , the receiver can compute the target function corresponding to @xmath198 .
all that remains to be done is to provide proofs of claims @xmath34 and @xmath170 .
+ _ proof of claim 1 _ : if a cut @xmath64 is such that @xmath199 , then @xmath200 } \\ & = { { \left\lvert { { k}_{c } } \right\rvert}}. & { & [ \mbox{from $ t = ( i \ u ) $ } ] } \end{aligned}\ ] ] thus by ( * ? ? ?
* theorem 3.1 ) , there exists a routing solution to compute the identity function of the sources @xmath201 at the receiver .
let @xmath202 and let @xmath203 for some ( arbitrary ) @xmath51 . by lemma [ lemma : koetter ] , after fixing @xmath204 , the vector received by @xmath8 can be written as @xmath205 the existence of a routing solution for computing the identity function guarantees that there exist @xmath206 such that the matrix @xmath207 has a non - zero determinant over @xmath55 .
it follows that the determinant of @xmath207 is non - zero over @xmath208}$ ] .
since @xmath209 was arbitrary in the above argument , it follows that the determinant of each @xmath210 is non - zero over @xmath208}$ ] and the claim follows .
+ _ proof of claim 2 _ : we have @xmath211 where , @xmath212 follows from the definition of @xmath213 . by contraction ,
assume that there exists an @xmath214 such that @xmath215 .
it then follows that @xmath216 } \label{eq : dependance}\end{aligned}\ ] ] for some choice of @xmath217 and @xmath218 } \nonumber \\
\big ( a_s(i - f)^{-1}b^{t } - \sum_{j=1}^{s-2 } \left(a_s(i - f)^{-1}b^{t } m_{(s)}^{-1}\right)_{i_j } ( a_{i_j}(i - f)^{-1}b)^{t } \big ) & = 0 . & { & [ \mbox{from $ m_{(s)}^{-1 } $ is full rank } ] } \label{eq : depend}\end{aligned}\ ] ] equation implies a linear dependence among @xmath163 rows of the matrix @xmath165 .
this contradicts the fact that for each @xmath219 , @xmath167 is full rank .
thus @xmath220 for @xmath171 and the claim follows .
we provide the following communication - theoretic interpretation of our method of proof above .
we may view the computation problem as a mimo ( multiple input multiple output ) channel where the multiple input is given by the vector of symbols generated by the sources , the output is the vector decoded by the receiver , and the channel is given by the network topology and the network code .
our objective is to choose a channel to guarantee the desired output , by way of code design subject to the constraints imposed by network topology .
the channel gain from source @xmath221 to the receiver is given by the vector @xmath222 of length @xmath163 .
the first part of the proof utilizes the sparse zeros lemma to establish that there exists a choice of channels such that the channel between every set of @xmath163 sources and the receiver is invertible .
this is similar to the proof of the multicast theorem in @xcite .
in the second part of the proof , we recognize that the interference from different sources must also be `` aligned '' at the output for the receiver to be able to compute the desired function . accordingly
, we have modified the code construction to provide such alignment .
we now show the existence of a linear function that can not be computed on a network satisfying the min - cut condition .
this network will then be used as a building block to show an analogous result for a larger class of functions .
let @xmath223 denote the matrix @xmath224 and let @xmath225 denote the corresponding linear function .
it is possible to show with some algebra that @xmath226 for any column vector @xmath12 of units , so that the conclusion of theorem [ th : solvable ] does not hold . indeed , for the function @xmath225 the opposite conclusion is true , namely @xmath225 can not be computed over @xmath227 using linear codes .
this is shown by the following lemma .
[ lemma : gbasis1 ] let @xmath228 be the network shown in figure [ fig : gbasis1 ] with alphabet @xmath55 .
we have 1 .
there does not exist a linear solution for computing @xmath225 in @xmath227 . that @xmath229 is easily verified by considering the cut @xmath230 which attains the minimum .
we now proceed to show , using theorem [ th : gbasiscondition ] , that a linear solution does not exist .
we may assume , without loss of generality , that the node @xmath231 sends its message directly to nodes @xmath232 and @xmath233 ( i.e. , @xmath234 ) .
the matrices @xmath235 , and @xmath236 over @xmath106 can then be written as @xmath237 consequently , the ideal @xmath114 is given by @xmath238 we have @xmath239 thus , it follows that @xmath240 . by theorem
[ th : gbasiscondition ] , a linear solution does not exist for computing @xmath225 in @xmath228 .
we now identify a much larger class of linear functions for which there exist networks satisfying the min - cut condition but for which linear solutions do not exist .
let @xmath154 be an @xmath155 matrix with at least one zero element and @xmath156 .
for each @xmath28 in this equivalence class we show that there exist a network @xmath0 that does not have a solution for computing the linear target function corresponding to @xmath28 but satisfies the cut condition in lemma [ lemma : ub ] .
the main idea of the proof is to establish that a solution for computing such a function in network @xmath0 implies a solution for computing the function corresponding to @xmath241 in @xmath228 , and then to use lemma [ lemma : gbasis1 ] .
[ th : insolvable ] consider a linear target function @xmath27 corresponding to a matrix @xmath67 .
if @xmath242 such that at least one element of @xmath154 is zero , then there exists a network @xmath0 such that 1 .
there does not exist a linear solution for computing @xmath27 in @xmath0 .
let @xmath244 and let @xmath245 denote the corresponding linear target function .
it is enough to show that there exists a network @xmath246 such that @xmath247 but @xmath246 does not have a linear solution for computing @xmath245 .
this is because a network @xmath0 that does not have a solution for computing @xmath28 is easily obtained by renaming the sources in @xmath246 as follows : since @xmath156 , there exist @xmath143 and @xmath139 such that @xmath248 .
let @xmath249 denote the permutation function on the set @xmath250 defined by the permutation matrix @xmath251 .
obtain the network @xmath0 by relabeling source @xmath221 in @xmath246 as @xmath252 . to see that there does not exist a solution for computing @xmath27 in @xmath0 , assume to the contrary that a solution exists . by using the same network code in @xmath246 ,
the receiver computes @xmath253 thus the receiver in @xmath246 can compute @xmath254 , which is a contradiction .
now we construct the network @xmath246 as claimed .
since @xmath154 has at least once zero element , there exists a @xmath255 such that @xmath151 has a zero in @xmath256-th column .
define @xmath257 denote the elements of @xmath258 by @xmath259 let @xmath260 be an element of @xmath261 ( such a @xmath260 exists from the fact that the @xmath256-th column contains at least one zero ) and define @xmath262 and denote the elements of @xmath263 by @xmath264 since @xmath265 does not contain an all - zero column , @xmath266 .
now , let @xmath246 denote the network shown in figure [ fig : insolvability ] where , @xmath267 denotes a relay node .
it follows from the construction that @xmath268 which is equal to the transfer matrix @xmath223 defined in .
notice that in the special case when @xmath269 and @xmath270 , the network shown in figure [ fig : insolvability ] reduces to the network shown in figure [ fig : reducedsum2 ] which is equivalent to the network @xmath228 in figure [ fig : gbasis1 ] with target function @xmath225 .
since @xmath228 does not have a solution for computing @xmath225 by lemma [ lemma : gbasis1 ] , we conclude that @xmath228 can not have a solution either .
similarly , we now show that in the general case , if the network @xmath271 has a solution for computing @xmath245 , then such a solution induces a solution for computing @xmath225 in network @xmath228 , contradicting lemma [ lemma : gbasis1 ] .
let there exist an @xmath121 for which there is a linear solution for computing @xmath245 over @xmath246 using an alphabet over @xmath40 .
in any such solution , for each @xmath272 , the encoding function on the edge @xmath273 must be of the form @xmath274 for some @xmath275 . since @xmath273 is the only path from source @xmath276 to the receiver , it is obvious that @xmath277 .
we define the map @xmath58 as follows .
let @xmath278 be arbitrary elements of @xmath40 and let @xmath279 note that @xmath58 has been chosen such that for any choice of @xmath280 , and @xmath281 , every edge @xmath282 carries the zero vector .
furthermore , for the above choice of @xmath58 , the target function associated with @xmath265 reduces to @xmath283 substituting @xmath284 and @xmath285 in , it follows that the receiver can compute @xmath286 from the vectors received on edges @xmath287 and @xmath288 .
consequently , it follows that there exist a linear solution over @xmath40 for computing the linear target function associated with the transfer matrix @xmath289 in the network shown in figure [ fig : reducedsum2 ] .
it is easy to see that the existence of such a code implies a scalar linear solution for computing @xmath225 in @xmath228 .
this establishes the desired contradiction . finally , we show that @xmath243 .
let @xmath128 be a cut such that @xmath290 ( i.e , @xmath64 separates sources from only the top and middle rows in the network @xmath246 ) .
we have the following two cases : 1 . if @xmath291 , then it is easy to see that @xmath292 .
similarly , if @xmath293 and @xmath294 , then again @xmath292 .
consequently , we have @xmath295 } \nonumber \\ & \ge 1 .
& { & [ \mbox{from $ { { \left\lvert c \right\rvert } } \ge { { \left\lvert k_c \right\rvert}}$ } ] } \label{eq : mincutfornp1}\end{aligned}\ ] ] 2 . if @xmath293 and @xmath296 , then from figure [ fig : reducedsum2 ] , @xmath297 and @xmath298 .
moreover , the index set @xmath258 was constructed such that @xmath299 consequently , we have @xmath300 } \nonumber \\ & \le { { \left\lvert k \right\rvert } } + 1 & { & [ \mbox{from \eqref{eq : linearrelation } } ] } \nonumber \\ & = { { \left\lvert c \right\rvert}}. \label{eq : mincutfornp2}\end{aligned}\ ] ] from and , we conclude that if @xmath290 , then @xmath301 for an arbitrary cut @xmath128 , let @xmath302 denote the number of sources in @xmath263 that are separated from the receiver by @xmath64 ( i.e , @xmath303 ) .
we have @xmath304 since each source in @xmath263 is directly connected to the receiver , @xmath305 is equal to the number of edges in @xmath64 separating the sources in @xmath306 from the receiver .
consequently , from , it follows that @xmath307 substituting in , we conclude that for all @xmath128 @xmath132 since the edge @xmath308 disconnects the source @xmath309 from the receiver , @xmath310 is immediate and the proof of the theorem is now complete .
we now consider the case in which the source alphabet is over the binary field . in this case
, we have that the two function classes identified by theorems [ th : solvable ] and [ th : insolvable ] are complements of each other , namely either @xmath311 or @xmath156 with @xmath154 containing at least one zero element .
[ lemma : binaryfield ] let @xmath138 and let @xmath145 . if @xmath312 , then there exists an @xmath313 matrix @xmath154 such that @xmath154 has at least one zero element and @xmath156 . since @xmath28 is assumed to have a full row rank , @xmath314 for some @xmath313 matrix @xmath315 over @xmath316 .
if @xmath317 has @xmath318 s , then we are done .
assume to the contrary that @xmath317 is a matrix of non - zero elements .
we only need to consider the case when @xmath319 ( since @xmath312 ) . for @xmath320 ,
let @xmath321 denote the @xmath18-th column vector of the @xmath142 identity matrix .
define @xmath322 and let @xmath139 be a permutation matrix that interchanges the @xmath50-th and @xmath323-th columns and leaves the remaining columns unchanged .
it is now easy to verify that @xmath324 where @xmath154 is an @xmath155 matrix with at least one @xmath318 element : for @xmath325 @xmath326 thus , @xmath327 and by transitivity we conclude that @xmath156 which proves the claim .
we wish to mention the following open problems arising from this work .
* is there a graph - theoretic condition that allows to determine whether a given network is solvable with reference to a given linear function ? we have provided an algebraic test in terms of the grob@xmath118ner basis of a corresponding ideal , but we wish to know whether there is there an algorithmically more efficient test . *
we showed that @xmath243 is not sufficient to guarantee solvability for a certain class of linear functions .
a possible direction of future research is to ask whether there is a constant @xmath328 such that @xmath329 guarantees solvability . alternatively , for every constant @xmath328
, does there exist a network @xmath0 and a matrix @xmath28 such that @xmath329 and @xmath0 does not have a linear solution for computing the linear target function associated with @xmath28 ?
[ lemma : equivalencelinearfunctions ] let @xmath67 .
if @xmath330 is a column vector of non - zero elements and @xmath190 , then there exists a full rank matrix @xmath143 and a column vector @xmath191 of non - zero elements over @xmath55 such that @xmath331 .
let @xmath143 denote the matrix obtained by collecting the first @xmath332 columns of @xmath28 .
we will first show that the matrix @xmath143 is full - rank .
after factoring out @xmath143 , we then prove that the last column must have non - zero entries .
since @xmath333 , there exists a full - rank matrix @xmath334 and a permutation matrix @xmath335 such that @xmath336 from , the columns of @xmath143 are constituted by the columns of @xmath334 in which case @xmath143 is full - rank , or columns of @xmath143 contains @xmath337 columns of @xmath334 and @xmath338 .
we will now show that the vector @xmath338 can not be written as a linear combination of any set of @xmath339 column vectors of @xmath334 .
assume to the contrary that there exist @xmath340 for @xmath341 such that @xmath342 where @xmath343 denotes the @xmath51-th column of @xmath334 .
let @xmath344 denote the vector such that @xmath345 , and @xmath346 .
we have @xmath347 } \nonumber \\ \bar{q } ( u - a ) & = 0 & { & [ \mbox{from \eqref{eq : dependence}}]}. \label{eq : contradiction}\end{aligned}\ ] ] contradicts the fact that @xmath334 is full - rank .
hence @xmath348 s satisfying do not exist and consequently , @xmath143 is a full - rank matrix .
we now have @xmath349 where @xmath350 and hence @xmath351 .
furthermore , @xmath352 and @xmath353 implies that @xmath354 .
thus , there exists a full - rank matrix @xmath154 and a permutation matrix @xmath139 such that @xmath355 let @xmath321 denote the @xmath18-th column of @xmath90 .
it follows from that either @xmath212 @xmath356 and @xmath154 itself is an @xmath175 permutation matrix , or @xmath357 for some @xmath358 , @xmath51-th column of @xmath154 is @xmath12 , and the remaining columns must constitute the @xmath339 columns @xmath359 of @xmath90 for some @xmath256
. if @xmath212 is true , then @xmath360 and the elements of @xmath191 are non - zero since @xmath361 is another permutation matrix . if @xmath357 is true , then @xmath362 and it must be that @xmath363 ( if @xmath364 , then @xmath365 which contradicts @xmath362 )
. let @xmath366 .
we must have @xmath367 if we denote the number of non - zero entries in a vector @xmath12 by @xmath368 , then we have @xmath369 } \nonumber \\ & = ( s-1 ) - { { \left\lvert d \right\rvert } } \nonumber \\ & \ge 1 & { & [ \mbox{from
$ { { \left\lvert d \right\rvert } } \le s-2 $ } ] } \label{eq : sandwich}\end{aligned}\ ] ] from , it follows that @xmath370 and consequently that every element of @xmath191 is non - zero .
the proof of the lemma is now complete .
10 r. ahlswede , n. cai , s .- y .
r. li , and r. w. yeung , `` network information flow '' , _ ieee transactions on information theory _ , vol .
it-46 , no .
4 , pp . 12041216 , july 2000 .
r. koetter and m. mdard , `` an algebraic approach to network coding , '' _ ieee / acm transactions on networking _ , vol .
5 , pp . 782795 , oct . 2003 .
h. kowshik and p. r. kumar , `` zero error function computation in sensor networks '' , _ proceedings of the ieee conference on decision and control _ , 2009
. b. k. rai , and b. k. dey , `` sum - networks : system of polynomial equations , unachievability of coding capacity , reversibility , insufficiency of linear network coding , '' http://arxiv.org/abs/0906.0695 , 2009 .
a. ramamoorthy , `` communicating the sum of sources over a network , '' _ proceedings of the ieee international symposium on information theory _ ,
toronto , canada , 2008 .
n. ma , p. ishwar , and p. gupta , `` information - theoretic bounds for multiround function computation in collocated networks , '' _ proceedings of the ieee international symposium on information theory _
, pp . 23062310 , 2009 .
b. nazer and m. gastpar , `` computing over multiple - access channels , '' _ ieee transactions on information theory _ , vol .
53 , pp . 34983516 , oct .
r. dougherty , c. freiling , and k. zeger , `` insufficiency of linear coding in network information flow , '' _ ieee transactions on information theory _ ,
51 , no . 8 , pp . 2745 - 2759 , august 2005 .
r. dougherty , c. freiling , and k. zeger , `` linear network codes and systems of polynomial equations '' , _ ieee transactions on information theory _ vol .
5 , pp . 2303 - 2316 , may 2008 .
r. appuswamy , m. franceschetti , n. karamchandani , and k. zeger , `` network coding for computing : cut - set bounds '' , _ to appear in ieee transactions on information theory _ , feb .
r. appuswamy , m. franceschetti , n. karamchandani , and k. zeger , `` network coding for computing : linear codes '' , _ submitted to ieee transactions on information theory _ , 2010 .
j. t. schwartz , `` fast probabilistic algorithms for verification of polynomial identities '' , _ j. acm . _ , vol .
701 - 717 , 1980 .
hungerford , `` algebra '' , _
springer - verlag _ , 1997 . | we consider the scenario in which a set of sources generate messages in a network and a receiver node demands an arbitrary _ linear function _ of these messages .
we formulate an algebraic test to determine whether an arbitrary network can compute linear functions using _
linear codes_. we identify a class of linear functions that can be computed using linear codes in every network that satisfies a natural cut - based condition .
conversely , for another class of linear functions , we show that the cut - based condition does not guarantee the existence of a linear coding solution . for linear functions over the binary field ,
the two classes are complements of each other
. [ section ] [ theorem ] lemma [ theorem ] corollary [ theorem ] proposition [ theorem ] remark [ theorem ] algorithm [ theorem ] conjecture [ theorem ] definition [ theorem ] example |
in 1996 yamazaki et al . discovered a bound state formed out of a @xmath0 meson and a @xmath1pb nucleus @xcite .
the negatively charged pion was bound with the lead nucleus by means of the coulomb interaction and the interplay between an attractive coulomb and a repulsive strong potential lead to the very narrow state with the width of less than one mev . due to the nature of the binding
we can consider the state as a pionic atom .
it is also conceivable that a neutral meson could be bound to a nucleus . in this case
the binding is exclusively due to the strong interaction and hence such object can be called a _
mesic nucleus_. here the most promissing candidate is the @xmath2-mesic nucleus since the @xmath2n interaction is attractive and since there exists a baryonic resonance ( n@xmath3 ) which couples predominantly to the @xmath2n channel @xcite
. we may picture @xcite the formation and decay of the @xmath2-mesic nucleus as the @xmath2 meson absorption by one of the nucleons and then its propagation in the nucleus via subsequent excitation of nucleons to the n@xmath3 state until the resonance decays into the pion - nucleon pair which escapes from the nucleus .
predicted values of the width of such states range from @xmath47 to @xmath440 mev @xcite .
the search of the @xmath2 - mesic nucleus was conducted in many inclusive experiments @xcite via reactions induced by pions @xcite , protons @xcite , and photons @xcite .
many promissing indications of the existence of such an object were reported @xcite , but so far none was independently confirmed .
experimental investigations with higher statistical sensitivity and the detection of the n@xmath3 decay products are being continued at the cosy @xcite , jinr @xcite , and mami @xcite laboratories . in this contribution
we report on searches of the @xmath2 - mesic helium in exclusive measurements carried out at the cooler synchrotron cosy by means of the wasa - at - cosy @xcite and cosy-11 @xcite detector setups .
we consider the study of the @xmath2-mesic nuclei as interesting on its own account , but additionally it is useful for investigations of ( i ) the @xmath2n interaction , ( ii ) the n@xmath3 properties in nuclear matter @xcite , ( iii ) the properties of the @xmath2 meson in the nuclear medium @xcite , and ( iv ) the flavour singlet component of the @xmath2 meson @xcite .
in 1985 bhalerao and liu @xcite performed a coupled - channel analysis of the @xmath5 , @xmath6 and @xmath7 reactions and discovered that the interaction between the nucleon and the @xmath2 meson is attractive . based on this finding
haider and liu postulated the existence of the @xmath2mesic nuclei @xcite , in which the electrically neutral @xmath2 meson might be bound with the nucleons by the strong interaction .
the formation of such a bound state can only take place if the real part of the @xmath2-nucleus scattering length is negative ( attraction ) , and the magnitude of the real part is greater than the magnitude of the imaginary part @xcite . in the 1980 s the @xmath2mesic nuclei
were considered to exists for a @xmath8 12 only @xcite due to the relatively small value of the @xmath2n scattering length ( @xmath9 fm @xcite ) .
however , recent theoretical investigations of hadronic- and photo - production of the @xmath2 meson result in values of @xmath10 which depending on the analysis method range from @xmath11 fm up to @xmath12 fm @xcite , and which do not exclude the formation of a bound @xmath2-nucleus states for such light nuclei as helium @xcite or even for deuteron @xcite . according to the calculations including multiple scattering theory @xcite or skyrme model
@xcite an especially good candidate for binding is the @xmath13 system .
recent calculations by haider @xcite or tryasuchev and isaev @xcite also indicate the binding in the @xmath13 system , while they rather exclude the existence of the @xmath14 state .
on the other hand there are promissing experimental signals which may be interpreted as indications of the the @xmath15 bound state .
for example the shape of the excitation function for the @xmath16 reaction @xcite , determined by the spes-4 @xcite , spes-2 @xcite , cosy-11 @xcite , and cosy - anke @xcite collaborations ( fig . [ fig1](left ) )
. reaction as determined by the cosy - anke @xcite ( open circles ) and the cosy-11 @xcite ( full dots ) and @xcite ( triangles ) .
the solid and dashed lines represent the scattering length fit to the cosy-11 and cosy - anke data , respectively .
( right ) angular asymmetry parameter @xmath17 .
closed and open circles represent results of cosy - anke @xcite and cosy-11 @xcite , respectively .
the dashed and solid lines denote results @xcite of the fit ( allowing for the phase variation ) to the cosy-11 and cosy - anke data , respectively .
the dotted line denotes result of the fit without the phase variation .
the figure is adopted from ref .
@xcite.,title="fig : " ] it has been indicated by wilkin @xcite that a steep rise of the total cross section in the very close - to - threshold region followed by a plateau may be due to the existence of a pole of the @xmath18 scattering amplitude in the complex excess energy plane @xmath19 with @xmath20 @xcite .
this reference shows that the occurence of the pole changes the phase and the magnitude of the s - wave production amplitude .
and indeed the momentum dependence of the asymmetry in the angular distributions of @xmath21 expressed in terms of a parameter @xmath17 @xcite , can only be satisfactorily described ( solid and dashed lines in fig .
[ fig1 ] ( right ) ) if a very strong phase variation associated with the pole is included in the fits @xcite .
otherwise there is a significant discrepancy between the experimental data and the theoretical description ( dotted line in fig .
[ fig1 ] ( right ) ) .
details of the experimental technique and the comprehensive description of results concerning the search of the @xmath2-mesic @xmath23 nucleus conducted by the cosy-11 group were described elsewhere @xcite .
therefore , here we only briefly summarize the main results .
the measurements were carried out using the deuteron beam of cosy which was circulating through the stream of the internal hydrogen target of the cluster - jet type @xcite .
data were taken during a slow acceleration of the beam from 3.095 gev / c to 3.180 gev / c , crossing the kinematical threshold for the @xmath2 production in the @xmath24 reaction at 3.141 gev / c .
the determined excitation function for the @xmath25 process does not show any structure which could originate from a decay of @xmath26 bound state .
the estimated upper limit for the cross section of the @xmath27 reaction chain is equal to 70 nb @xcite .
similarly , the analysis of the @xmath28 reaction resulted only in an upper limit of 270 nb for the total cross section of the @xmath29 reaction @xcite .
the installation of the wasa detector at cosy opened a unique possibility to search for the @xmath31 bound state with high statistics and high acceptance .
we have conducted a search via an exclusive measurement of the excitation function for the @xmath32 reaction varying continuously the beam momentum around the threshold for the @xmath33 reaction .
ramping of the beam momentum and taking advantage of the large acceptance of the wasa detector reaction the acceptance of the wasa - at - cosy detector equals to almost 70% . ] allows to minimize systematical uncertainities making the wasa - at - cosy a unique facility @xcite for such kind of exclusive experiments .
the @xmath34 bound state should manifest itself as a resonant like structure below the threshold for the @xmath35 reaction .
if a peak below the @xmath36 threshold is found , then the profile of the excitation curve will allow to determine the binding energy and the width of the @xmath37 bound state .
if , however , only an enhancement around the threshold is found , then it will enable to establish the relation between width and binding energy @xcite .
finally , if no structure is seen the upper limit for the cross section of the production of the @xmath2-helium nucleus will be set at few nanobarns .
in addition , when searching for the signal of the @xmath2-mesic state we may take advantage of the fact that the distribution of the relative angle between the @xmath38 pair for the background ( due to the prompt @xmath39 reaction ) is much broader than the one expected from the decay of the bound state .
this is because the relative angle between the outgoing @xmath38 pair originating from the decay of the n@xmath3 resonance is equal to 180@xmath40 in the n@xmath41 reference frame and it is smeared only by about 30@xmath40 in the reaction center - of - mass frame due to the fermi motion of the nucleons inside the @xmath42 nucleus .
this allows for an additional control of the background by comparing excitation functions corresponding to the `` signal - rich '' and `` signal - poor '' regions .
in the first experiment conducted in june 2008 , we used a deuteron pellet target and the cosy deuteron beam with a ramped momentum corresponding to a variation of the excess energy for the @xmath37 system from -51.4 mev to 22 mev . at present
the data are evaluated and some preliminary results were reported @xcite . the experiment will be continued in 2010 @xcite .
two weeks of cosy beamtime were already recommended by the cosy program advisory committee .
the systematic uncertainties in establishing the shape of the excitation functions discussed in the previous sections are significantly reduced by using the momentum ramping technique since the energy range of interest for the search of the @xmath2-mesic nucleus is scanned in each cosy cycle .
a scan of the energy can also be achieved in the case of the fixed beam energy available for the external beam experiments .
this may be realised by means of the quasi - free reactions as already successfully used at cosy for the study of the meson production in quasi free proton - neutron collisions @xcite .
for instance , a search of the @xmath2-mesic tritium may be realized by studying the exitation function of the @xmath43 reaction using a deutron beam and tagging the nd reactions by the measurement of the spectator protons ( @xmath44 ) from the @xmath45 reaction .
the fermi motion of nucleons inside the deuteron beam will allow to scan a large range ( in the order of 100 mev ) of excess energies in the nd reaction .
the work was supported by the european community - research infrastructure activity under the fp6 program ( hadron physics , rii3-ct-2004 - 506078 ) , by the german research foundation ( dfg ) , by the ffe grants from the research center jlich , and by the polish ministry of science and higher education ( 3240/h03/2006/31 and 1202/dfg/2007/03 ) .
abc t. yamazaki _ et al .
_ , _ z. phys .
_ * a355 * , 219 ( 1996 ) . c. amsler _ et al .
_ , _ phys .
lett . _ * b667 * , 1 ( 2008 ) . g. a. sokol et al . , arxiv : nucl - ex/0106005 . c. garcia - recio _ et al .
_ , _ phys .
lett . _ * b550 * , 47 ( 2002 ) . q. haider , l.c .
liu , _ phys .
rev . _ * c66 * , 045208 ( 2002 ) .
q. haider , l.c .
liu , _ acta phys . pol .
* 2 * , 121 ( 2009 ) .
r. e. chrien _ et al .
lett . _ * 60 * , 2595 ( 1988 ) .
b. j. lieb _ et al .
_ , _ proc .
conf . , sao paulo , brazil _ ( 1989 ) .
g. a. sokol _ et al .
_ , _ arxiv : nucl - ex/9905006 _ ( 1999 ) . m. kh
. anikina _ et al .
_ , _ arxiv : nucl - ex/0412036 _ ( 2004 ) .
a. gillitzer , _ acta phys . slovaca _ * 56 * , 269 ( 2006 ) .
a. budzanowski _ et al .
_ , _ phys .
rev . _ * c79 * , 061001(r ) ( 2009 ) .
m. pfeiffer , _ et al .
_ , _ phys .
_ * 92 * , 252001 ( 2004 )
. p. moskal _
et al . _ ,
cosy proposal no .
186.1 ( 2008 ) .
w. krzemien , p. moskal , j. smyrski , _ acta phys . pol .
_ * 2 * , 141 ( 2009 ) .
b. krusche , i. jaegle , _ acta phys . pol .
* 2 * , 51 ( 2009 ) .
j. smyrski , p. moskal , w. krzemien , _ acta phys . pol .
* 2 * , 133 ( 2009 ) .
w. krzemie _ et al .
phys . _ * a24 * , 576 ( 2009 ) . j. smyrski _ et al .
_ , _ nucl .
phys . _ * a790 * , 438 ( 2007 ) . j. smyrski _ et al .
_ , _ acta phys
. slovaca _ * 56 * , 213 ( 2006 ) .
d. jido , h. nagahiro , s. hirenzaki , _ phys .
rev . _ * c66 * , 045202 ( 2002 ) .
t. inoue , e. oset , _ nucl .
_ * a710 * , 354 ( 2002 ) .
s. d. bass , a. w. thomas , _ phys .
lett . _ * b634 * , 368 ( 2006 ) .
r. s. bhalerao , l.c .
liu , _ phys .
lett . _ * 54 * , 865 ( 1985 )
. q. haider , l.c .
liu , _ phys .
lett . _ * b172 * , 257 ( 1986 ) .
a. sibirtsev _ et al .
_ , _ phys .
rev . _ * c70 * , 047001 ( 2004 ) .
a. m. green , s. wycech , _ phys .
_ * c71 * , 014001 ( 2005 ) . c. wilkin , _ phys .
* c47 * , r938 ( 1993 ) . s. wycech , a. m. green and j. a. niskanen , _ phys .
rev . _ * c52 * , 544 ( 1995 ) .
a. m. green _ et al .
_ , _ phys .
rev . _ * c54 * , 1970 ( 1996 ) .
n. n. scoccola and d. o. riska , _ phys .
lett . _ * b444 * , 21 ( 1998 ) .
v. a. tryasuchev , a. v. isaev , _ arxiv:0901.3242 _ ( 2009 ) .
j. berger , _ et al .
_ , _ phys . rev
. lett . _ * 61 * , 919 ( 1988 ) .
b. mayer , _ et al .
_ , _ phys .
rev . _ * c53 * , 2068 ( 1996 ) .
j. smyrski _ et al .
_ , _ phys .
* b 649 * , 258 ( 2007 ) .
t. mersmann _ et al .
_ , _ phys .
lett . _ * 98 * , 242301 ( 2007 ) .
adam _ et al .
_ , _ phys .
_ * c75 * , 014004 ( 2007 ) . c. wilkin _ phys .
* b654 * , 92 ( 2007 ) . c. wilkin _ aip conf
_ * 950 * , 23 ( 2007 ) .
s. brauksiepe _ et al .
_ , _ nucl .
instr . & meth .
_ * a376 * , 397 ( 1996 ) .
adam _ et al .
_ , arxiv : nucl - ex/0411038 .
p. moskal _ et al .
_ , _ phys .
rev . _ * c79 * , 015208 ( 2009 ) , and references therein .
m. abdel - bary _ et al .
_ , _ eur .
j _ * a29 * , 353 ( 2006 ) . | we review status and perspectives of the search of the eta -mesic helium at the cooler synchrotron cosy . |
ob associations were first recognized by @xcite and @xcite as extended moving groups of blue luminous stars .
they are defined as loose stellar systems ( stellar mass density of @xmath5 ) containing o- and/or early b - type stars ( for a recent review see * ? ? ?
* ) . at such low densities ,
the associations are unstable against galactic tidal forces , and therefore it follows from their definition that ob associations must be young ( @xmath6 myr ) entities
. most of their low - mass members are therefore still in their pre - main sequence ( pms ) phase .
there are different models for the origin of ob associations .
one possibility is that they start as initially dense clusters , which get unbound and expand quickly as soon as the massive stars expel the gas ( see * ? ? ?
an alternative model assumes that ob associations originate from _ unbound _ turbulent giant molecular clouds ( see * ? ? ?
* ) , i.e. start already in a spatially extended configuration and thus form in a fundamentally different way than dense , gravitationally bound clusters .
many well investigated ob associations show remarkably small internal velocity dispersions ( often @xmath7 km / sec ) , which are in some cases much smaller than required to explain the large present - day size ( typically tens of parsecs ) by expansion over the age of the association .
this excludes the expanding cluster model and provides strong support for an origin as an extended unbound cloud for these associations .
the considerable number of ob associations in the solar neighborhood ( e.g. , * ? ? ?
* ) suggests that they account for a large , maybe the dominant , fraction of the total galactic star formation .
a good knowledge of their stellar content is thus essential in order to understand the nature of the star formation process not only in ob associations but also on galactic scales . for many years
star formation was supposed to be a bimodal process ( e.g. * ? ? ?
* ; * ? ? ?
* ) according to which high- and low - mass stars should form in totally different sites .
although it has been long established that low - mass stars _ can _ form alongside their high - mass siblings in nearby ob associations ( e.g. * ? ? ?
* ) , it is still not well known _
what quantities _ of low - mass stars are produced in ob environments .
there have been many claims that high - mass star forming regions have a truncated initial mass function ( imf ) , i.e. contain much smaller numbers of low - mass stars than expected from the field imf ( see , e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
possible explanations for such an effect are often based on the strong radiation and winds from the massive stars .
for example , increased radiative heating of molecular clouds may raise the jeans mass ; lower - mass cloud cores may be completely dispersed by photoevaporation before low - mass protostars can even begin to form ; the radiative destruction of co molecules should lead to a change in the equation of state of the cloud material , affecting the fragmentation processes and ultimately leading to the formation of a few massive stars rather than the `` normal '' imf which is dominated by low - mass stars ( e.g. * ? ? ?
* ) . however , several well investigated massive star forming regions show _ no _ evidence for an imf cutoff ( see , e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* for the cases of ngc 3603 , 30 dor , and ngc 346 , respectively ) , and notorious difficulties in imf determinations of distant regions may easily lead to wrong conclusions about imf variations ( see , e.g. , discussion in * ? ?
* ; * ? ? ?
if the imf in ob associations is not truncated and similar to the field imf , it would follow that most @xmath8 of the total stellar mass is found in low - mass @xmath9 stars .
this would then imply that most of the current galactic star formation is taking place in ob associations ( as initially suggested by * ? ? ?
* ) , and the _ typical _ environment for forming stars ( and planets ) would be close to massive stars and not in isolated regions like taurus .
the presence of nearby massive stars affects the evolution of young stellar objects and their protoplanetary disks in ob environments .
for example , photoevaporation by intense uv radiation can remove a considerable amount of circumstellar material around young stellar objects ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , and these objects will therefore ultimately end up with smaller final masses than if they were located in isolated regions @xcite . although generally considered to be a threat for forming planetary systems , photoevaporation may actually help to form planets , as it seems to play an important role in the formation of planetesimals @xcite .
ob associations provide excellent targets to investigate these effects on the formation and evolution of low - mass stars ( and their forming planetary systems ) during ages between a few myr and a few ten myr .
however , before one can study the low - mass members , one first has to find them .
although this statement sounds trivial , the major obstacle on the way towards a reliable knowledge of the low - mass population in ob associations is the problem to identify the individual low - mass members . unlike stellar clusters , which can be easily recognized on the sky ,
ob associations are generally very inconspicuous : since they extend over huge areas in the sky ( often several hundred square - degrees for the nearest examples ) , most stars in the area actually are unrelated foreground or background stars . finding the association members among these field stars
is often like finding needles in a haystack .
as the low - mass members are often too faint for proper - motion studies , the only reliable sign to discern between low - mass association members and unrelated , much older field stars is the strength of the 6708 lithium line in the stellar spectrum : at ages of @xmath10 myr , the low - mass association members still have most of their initial li preserved and show a strong li line , whereas the older foreground and background field stars do not show this line since they have already depleted their primordial li ( e.g. , * ? ? ?
. however , an accurate measurement of li line width requires at least intermediate resolution spectroscopy , and thus the observational effort to identify the widespread population of pms stars among the many thousands of field stars is huge .
many empirical imf determinations are therefore based on photometric data only ; while this strategy rather easily provides a complete spatial coverage and allows one to work with large samples , photometry alone can not give completely reliable membership information .
most studies with spectroscopically identified member samples , on the other hand , include only very small fractions of the total stellar population and are strongly affected by small number statistics and the necessity of using large extrapolation factors .
therefore , most studies dealing with ob associations have been restricted to estimating the number - ratio of low - mass versus high - mass members ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
during the last years , new and very powerful multiple - object spectrographs like 2df at the anglo - australian telescope ( see * ? ? ?
* ) made large spectroscopic surveys for low - mass pms members feasible . in combination with the hipparcos results , which allowed the complete identification of the high- and intermediate - mass @xmath11 stellar population in many nearby associations ( see * ? ? ?
* ) , studies of the _ complete _ stellar population in ob associations are now possible , enabling us to investigate in detail the spatial and temporal relationships between high- and low - mass members . in this chapter , we review the state of knowledge regarding the stellar populations of the nearest ob association : scorpius - centaurus ( sco ob2 ) .
we structure this chapter in the following manner .
section 2 gives a general , and historical , overview of the association and its subgroups .
sections 3 , 4 , and 5 discuss the surveys for the members of the three primary subgroups of sco - cen : upper scorpius , upper centaurus - lupus , and lower centaurus - crux , respectively .
readers uninterested in the individual surveys may want to skip ahead to section 6 , where we discuss the astrophysical ramifications , and interpretations , of studies of the sco - cen members .
the scorpius - centaurus ( sco - cen ) association is the ob association nearest to the sun ( see fig . [ nearby_ob.fig ] ) .
it contains at least @xmath3150 b stars which concentrate in the three subgroups upper scorpius ( us ) , upper centaurus - lupus ( ucl ) , and lower centaurus - crux ( lcc ; fig .
[ scocen_map.fig ] ; cf . * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
there is one o - type star associated with sco - cen : the runaway o9v star @xmath12 oph .
the currently recognized sco ob2 subgroups were defined by @xcite : subgroup # 2 is us , # 3 is ucl , and # 4 is lcc . derived ages for the b - type stars in the different subgroups from the main sequence turnoff in the hr diagram and found that upper scorpius is the youngest subgroup ( @xmath13 myr ) , whereas lower centaurus crux ( @xmath14 myr ) and upper centaurus - lupus ( @xmath15 myr ) are considerably older myr for lower centaurus - crux and @xmath16 myr for upper centaurus - lupus . ] . although the focus of our review is the low - mass membership of sco - cen , it is worth reviewing the history regarding the unveiling of the group s membership . for most of the past century
`` sco - cen '' was studied as a single entity , and only the highest mass members ( earlier than b5-type ) were called `` members '' with any degree of confidence . the discovery of the low - mass population ( fgkm stars ) in large numbers has only recently become technically feasible .
many early studies of sco - cen involved isolating the members from the field population through kinematic and photometric means .
sco - cen contained a large sample of b - type stars with large , convergent proper motions , which provided a critical lower rung in the galactic ( and cosmic ) distance ladder ( e.g. * ? ? ? * ) .
our brief historical review here is not exhaustive , but complements the recent summary by ( * ? ? ?
* their sect .
4.1 ) . around 1910 ,
large compilations of proper motion and spectral type data were becoming available for galactic structure investigations
primarily lewis boss s ( 1910 ) _ preliminary general catalog _ and the spectral type compilations from the henry draper memorial project ( e.g. * ? ? ?
@xcite appears to have been the first to make a convincing case that the b - type `` helium '' stars in the scorpius - centaurus region demonstrate convergent proper motions , and constituted a moving group .
subsequent work on the association during the early 20th century revolved around testing the reality of the group , ascertaining its membership , and ( most importantly for the rest of the astronomical community ) estimating the group s distance .
the importance of understanding , and exploiting , this conspicuous group of bright b - type stars was not lost on @xcite : `` _ the real question of importance is this : is the parallelism and equality of motion in this part of the sky [ sco - cen ] of such a nature that we can derive individual parallaxes ? _ ''
the reality of sco - cen as a moving group was challenged , most notably , by @xcite , and @xcite .
some primary objections raised by these authors were that ( 1 ) the convergent point of the sco - cen proper motions was so close to the solar antapex that it could be construed simply as solar reflex motion on unrelated b - type stars in the field , ( 2 ) the group was so dispersed that it should disintegrate on a short time scale , and ( 3 ) the sco - cen space motion varied as one isolated different subsamples within the group , enough so to cast doubt on the derived cluster parallaxes for the constituent stars .
many of the kinematic objections by smart and others were addressed by adriaan @xcite in his phd thesis , which concluded that sco - cen was indeed a true moving group . perhaps blaauw s ( 1946 )
most amusing retort was in telling the regius chair of astronomy at glasgow to simply look up ( p. 19 ) : `` _ smart has not paid attention to the fact that the existence of the cluster is evident from the apparent distribution of the bright stars in the sky .
_ '' in hindsight , the first objection is understandable as the young stars in the solar neighborhood form from molecular gas which itself has small peculiar motions with respect to the lsr @xcite .
the last two objections are symptomatic of ob associations in general , when compared to the more coherent kinematic groups like the hyades cluster . as one of the few nearby ob associations with appreciable proper motion , the kinematic studies of blaauw , and others , were critical to understanding the dynamical state of ob associations in general .
the modern consensus ( e.g. * ? ? ?
* ; * ? ? ?
* ) is that sco - cen constitutes a moving group , but that it has subgroups isolated by position , age , and space motion , and that these structures are young and unbound .
investigations of the expansion of the sco - cen subgroups have been undertaken by @xcite , @xcite , and @xcite . a modern study using the radial velocities of the low - mass members
is sorely needed to confirm , and build upon , these findings .
the most recent investigation of the membership of high - mass stars in scorpius - centaurus was carried out by @xcite .
they used hipparcos proper motions and parallaxes in conjunction with two moving group methods in order to accurately establish the high - mass ( and sometimes intermediate mass ) stellar content of 12 nearby ob associations . in sco - cen the membership of nearly 8000 hipparcos catalogue stars was investigated .
a total of 120 stars in us , 221 stars in ucl , and 180 stars in lcc were identified as high probability members .
the mean distances of the subgroups , derived from the hipparcos parallaxes , are 145 pc for us , 140 pc for ucl , and 118 pc for lcc .
the vicinity of sco - cen is rich with well - studied sites of current , and recently terminated , star - formation .
besides the ophiuchus and lupus star - forming regions , there are other dark clouds , t associations , and somewhat older `` outlying '' stellar groups which appear to be genetically related to sco - cen by virtue of their ages , positions , and space motions @xcite .
several of these are described elsewhere in this volume .
these neighboring regions demonstrate a broad evolutionary spectrum of star formation .
they include dark clouds with little , if any , star - formation activity ( e.g. musca , coalsack , pipe nebula , cha iii ) , molecular cloud complexes that are currently forming stars ( e.g. cha i & ii , cra ) , and several recently discovered groups of @xmath17 myr - old stars with little or no trace of the dark clouds from which they formed ( e.g. the tw hya , @xmath18 pic , @xmath19 cha , @xmath20 cha groups ) . in an investigation of the origins of the @xmath19 cha cluster , @xcite and @xcite found that many nearby , young groups , and isolated young stars , in the southern hemisphere within @xmath3200pc ( e.g. @xmath19 cha , tw hya , @xmath18 pic , @xmath20 cha , cra , etc . ) are not only spatially close to the sco - cen ob association , but _ moving away from the subgroups_. the star - forming clouds in oph , cra , and cha i manifest head - tail morphologies , with the star - forming `` heads '' on the side facing sco ob2 .
more detailed investigations of the space motions and star - formation histories of the associations in this region are needed , however the preliminary results suggest that star - formation in these small `` satellite '' groups near sco - cen may have been triggered by the massive star - formation event in the primary sco - cen subgroups ( see , e.g. the comprehensive model scenario for the formation of the sco - cen association and the young stellar groups proposed by * ? ? ?
the sco - cen region , including the ob subgroups , molecular clouds , and outlying associations , may be thought of as a small star - forming _ complex _ @xcite .
the _ sco - cen complex _ has been variously referred to as the _ oph - sco - cen association _
( osca ; * ? ? ?
* ) , _ greater sco - cen _ @xcite , or , perhaps with tongue in cheek , the _ oph - sco - lup - cen - cru - mus - cha star - formation region _ @xcite . throughout this review
, we will mostly limit our discussions to the 3 subgroup regions outlined by fig .
2 ( us , ucl , lcc ) , _ but excluding the regions associated with the ophiuchus and lupus molecular cloud complexes _
( see chapters by wilking et al . and
comern ) .
both complexes are within @xcite projected boundaries of us and ucl , respectively , and are approximately co - distant with the subgroups ( @xmath21 pc ) .
there is no evidence for ongoing star formation activity in the ob subgroups of sco - cen itself .
this makes it an ideal target for an investigation of the _ outcome of the recently completed star formation process_. the area is essentially free of dense gas and dust clouds , and the association members show only very moderate extinctions ( @xmath22 mag ) .
this is probably the consequence of the massive stellar winds and several supernova explosions , which have cleared the region from diffuse matter and created a huge system of loop - like hi structures around the association .
these loop structures have a total mass of about @xmath23 and seem to be the remnants of the original giant molecular cloud in which the ob subgroups formed ( cf . * ? ? ?
* ) . for historical completeness , we mention some candidate stellar groups in the vicinity of sco - cen which have been proposed , but later refuted . was unable to verify the existence of kinematic groups in blaauw s ( 1946 ) areas # 1 ( cra region ) and # 5 ( car - vel region ) .
blaauw s areas # 6 and # 7 correspond to the modern - day vel ob2 and collinder 121 associations , however they were sufficiently detached from groups # 2-#4 in position , distance , and velocity , that @xcite did not include them in his final census of sco - cen groups .
@xcite claimed to have discovered a moving group of x - ray bright stars adjacent to lcc in carina - vela , in essentially the same region as blaauw s subgroup # 5 .
they claimed that the new group is a `` near extension of the sco - cen complex '' and that the open cluster ic 2602 was part of this group .
the status of car - vel as a coherent group is very unlikely , let alone any relation to sco - cen .
@xcite show that the inferred `` kinematic '' parallaxes for their proposed car - vel membership show a disturbing `` finger - of - god '' effect with distances ranging from @xmath24 pc , with a large gap in the distribution .
closer examination of the kinematic and spectroscopic data for these stars by ( * ? ? ?
* and in prep . )
show that the objects appear to constitute a heterogeneous sample of low - mass members of ic 2602 and the @xmath25 car cluster (= platais 8) , and probable gould belt stars with a large range of distances .
@xcite claim that a subsample of the car - vel stars at d @xmath26 pc constitute a previously unknown nearby @xmath27 myr - old group , which they dub `` carina - near '' .
regardless , the carina stars appear to be unrelated to sco - cen .
the consensus from studies of the high mass and low mass populations appears to be that the western `` edge '' of sco - cen lies near galactic longitude 290@xmath28 .
recently , @xcite and @xcite suggested that an ob association spatially contiguous to lcc might exist to the south in chamaeleon . while there are a total of 4 known b stars associated with the cha i ( 2 ) , @xmath20 cha ( 1 ) , and @xmath19 cha ( 1 ) kinematic groups in chamaeleon
, @xcite argued that the kinematic and stellar density data are inconsistent with the idea of an ob association in cha , at least of the size proposed by ( * ? ? ? * b - type stars ) .
their sample appears to be dominated by field stars completely unrelated to the chamaeleon molecular clouds or sco - cen .
the upper scorpius association is the best studied part of the sco - cen complex . despite its rather young age ( @xmath29 myr ) and the neighborhood to the @xmath30 oph molecular cloud ( see chapter by wilking et al . in this book ) , which is located in front of the southeastern edge of us and is well known for its strong star formation activity , there are no indications for ongoing star formation in us . below we will make a distinction between the high- and the low - mass stellar population .
the former refers to stars of spectral types f and earlier ( @xmath31 ) whose membership of us was established using hipparcos data @xcite .
the latter refers to g , k , and m stars in the mass range @xmath32 m@xmath33 for which membership was established from their pms character . investigated the membership of 1215 stars in us listed in the hipparcos catalogue ; 120 of these were identified as genuine members .
the spectral types of the members on the ( pre-)main - sequence range from b0.5v to g5v , and there are some evolved stars with giant luminosity classes ( including the m1.5 supergiant antares [ @xmath34 sco ] ) .
the most massive star in us was presumably a @xmath35 o5o6 star , which exploded as a supernova about 1.5 myr ago .
@xcite suggested that the pulsar psr j1932 + 1059 is the remnant of this supernova and that the runaway star @xmath12 oph was the previous binary companion of the supernova progenitor and was ejected by the explosion .
however , the new parallax of psr j1932 + 1059 determined by @xcite challenged this scenario .
the new data suggest that the pulsar was probably _ not _ the former binary companion of @xmath12 oph , but it is still possible that psr j1932 + 1059 was created in us @xmath36 myr ago .
the 120 kinematic members of us cover an area of about 150 deg@xmath37 on the sky .
the large intrinsic size of the association suggests that the spread of individual stellar distances can not be neglected .
the projected diameter on the sky is @xmath38 which at the distance of us corresponds to @xmath39 pc .
however , while the hipparcos data allow the determination of a very accurate mean distance of @xmath40 pc for us @xcite , the errors on the trigonometric parallaxes ( @xmath41 mas ) are too large to resolve the internal spatial structure .
the only conclusion that can be drawn directly from the hipparcos parallaxes is that the line - of - sight depth of us can not be much larger than @xmath42 pc .
this prompted @xcite to carry out a more detailed investigation of the hipparcos members of scorpius - centaurus by performing a careful kinematic modeling of the proper motion and parallax data .
he used a maximum likelihood scheme based on a generalized moving - cluster method to derive secular ( or ` kinematically improved ' ) parallaxes for the association members .
he showed that the method is robust and that the secular parallaxes are a factor of @xmath43 more precise than the hipparcos trigonometric parallaxes .
the secular parallaxes for the members of us show a much reduced line - of - sight dispersion , confirming that the dominant part of the scatter in the hipparcos distances is caused by the trigonometric parallax errors and not by a large intrinsic dispersion of individual stellar distances .
the distribution of secular parallaxes for us is not resolved , which means that the distance spread can not be larger than @xmath44 pc .
hence one can assume that us has a roughly spherical shape , i.e. that the intrinsic spread of distances is about @xmath45 pc from the mean value of 145 pc .
previous investigations of the imf of us focused on the high- to intermediate mass stellar content . established the membership of stars in us using walraven multi - color photometry , and determined their physical parameters ( @xmath46 and @xmath47 ) , from which they derived stellar masses .
@xcite used the preliminary results on membership from the hipparcos data to determine the luminosity function for the high - mass stars using the hipparcos parallaxes , and then transformed a smoothed version of this into a mass function using the mass - luminosity relation listed in @xcite . adopting a conservative completeness limit of @xmath48 , which corresponds to masses of about @xmath49 , he concluded that down to this mass limit the imf is consistent with a single power - law @xmath50 with slope @xmath51 .
there are , however , several problems with this determination of the imf , most importantly the fact that a mass - luminosity relation had been used that might not be appropriate for young stars ( see discussion in brown 1998 ) .
finally , @xcite attempted to extend the completeness of the kinematic studies of the membership of us toward lower masses by making use of the trc and act astrometric catalogs @xcite , which are believed to be complete to @xmath52 .
he selected some 250 candidate members with @xmath53 from these catalogs by searching for stars with proper motions consistent with those of the hipparcos association members , and which lie @xmath54 mag below and @xmath55 mag above the main sequence in the color - magnitude diagram .
however , we note that the second selection criterion actually excludes most of the late - type association members , which , at an age of about 5 myr , lie well above the main sequence .
numerous studies have tried to reveal low - mass stars in us .
most of these , however , focused on very small subregions of the association .
for example , @xcite studied iras sources in a @xmath56 deg@xmath37 field near @xmath57 sco and found 4 young stars .
@xcite used deep pointed rosat x - ray observations to search for pms stars in a @xmath58 deg@xmath37 area and found several candidates for pms stars .
@xcite analyzed pointed rosat observations in the vicinity of the @xmath30 oph star forming region and found a number of additional pms stars in this area .
the first systematic search for low - mass members covering a significant part of us was performed by @xcite , who obtained spectroscopy and photometry for the optical counterparts of x - ray sources detected in 7 individual einstein fields .
they classified 28 objects as low - mass pms stars and placed them into the hr - diagram .
they found a remarkably small dispersion in stellar ages pc derived from the hipparcos data for the massive members . ] and interpreted this as an indication that the formation of these stars was triggered by some external event . in another study ,
m. kunkel investigated the optical counterparts of more than 200 rosat all sky survey ( rass ) x - ray sources in a @xmath59 deg@xmath37 area in us and ucl ( see * ? ? ? * for a list of these stars ) .
32 objects in this sample that are located in us can be classified as new low - mass members ( cf .
a deep search for very - low mass pms star and brown dwarf candidates in us was presented by @xcite .
their photometric survey covered an area of 14 deg@xmath37 and yielded some 100 candidate members .
low - resolution spectroscopy for some of these candidates led to the classification of 20 stars with strong h@xmath34 emission as potential association members . for eleven of these candidates @xcite and @xcite performed high - resolution optical spectroscopy and derived stellar parameters .
they showed that five of these objects have masses @xmath60m@xmath33 , i.e. are brown dwarfs .
@xcite presented low - resolution optical spectroscopy of further candidate very low - mass members of us . their analysis indicated that 28 of these objects are most likely members of us , and 18 objects have spectral types in the range m6.5m9 , i.e. are likely young brown dwarfs .
@xcite analyzed deep _
xmm - newton _ x - ray observations of two fields in us . among the 224 detected x - ray sources they identified 22 stars as photometric member candidates , 13 of which were not known to be association members before .
@xcite presented a wide - field optical / near - infrared photometric survey of us .
follow - up spectroscopy of selected stars led to the identification of 43 new low - mass members with estimated masses in the @xmath61 range , 30 of which are likely new brown dwarf members of us .
finally , @xcite identified about a dozen additional new likely low - mass members of us from an analysis of ukirt infrared deep sky survey early data release data of a 9.3 deg@xmath37 field in us and follow - up observations . in continuation of this work , @xcite used ukidss galactic cluster survey data of a 6.5 deg@xmath37 region in us and identified 129 members by photometric and proper motion criteria .
the estimated masses of these objects are in the range between @xmath62 and @xmath63 and they conclude that the sample contains a dozen new brown dwarf candidates below 15 jupiter masses . during the last couple of years ,
extensive spectroscopic surveys for low - mass members of us were performed with wide - field multi - object spectrographs at the anglo - australian observatory .
the first step was a survey with the wide - field multi - object spectrograph flair at the 1.2 m united kingdom schmidt telescope to reveal pms stars among rosat all sky survey x - ray sources in a 160 deg@xmath37 area @xcite . in this spatially complete , but flux - limited survey covering nearly the full area of the association , 39 new pms stars were found .
@xcite investigated the star formation history in us . in a detailed analysis of the hr diagram , properly taking into account the uncertainties and the effects of unresolved binaries , they found that the low - mass pms stars have a mean age of about 5 myr and show no evidence for a large age dispersion .
the pms sample of @xcite ( see table [ pz99tab ] ) is statistically complete for stars in the mass range @xmath64m@xmath33 to @xmath56m@xmath33 .
the next step was to reveal the _ full _ population of low- and very - low mass @xmath65m@xmath33 ) stars in a representative area of us , in order to allow a direct determination of the full imf of this ob association .
the multi - object spectrograph 2df at the 3.9 m anglo - australian - telescope was used to obtain intermediate resolution spectra of more than 1000 stars with magnitudes @xmath66 in a 9 deg@xmath37 area . among these
, 166 new pms stars were found , nearly all of them m - type stars , by their strong li absorption lines .
the results of these observations were reported in @xcite and @xcite , and the newly revealed low - mass members are listed in table [ 2dfpmstab ] . combining these results with
the earlier investigation yielded a sample of 250 pms stars in the mass range @xmath67m@xmath33 to @xmath56m@xmath33 .
a map of the survey region showing the locations of the low - mass pms stars as well as the high - mass hipparcos members is shown in fig .
[ usco_map.fig ] .
one can see that the low - mass members are spatially coincident with the early type members of the us association .
@xcite studied the properties of the full stellar population in us on the basis of a large sample of 364 association members .
this sample was composed of the following parts : \(1 ) the 114 hipparcos members constitute a complete sample of all members with masses above @xmath68 and an ( incomplete ) sample of lower - mass members with masses down to @xmath69m@xmath33 , covering the full spatial extent of the association .
\(2 ) the 84 x - ray selected pms stars from @xcite provide a statistically complete sample of the @xmath70m@xmath33 members ( plus some lower mass stars ) in a 160 deg@xmath37 area .
\(3 ) the 166 low - mass pms stars identified with 2df @xcite constitute a statistically complete , unbiased sample of the member population in the @xmath71m@xmath33 mass range , which covers a 9 deg@xmath37 area .
figure [ hrd_full.fig ] shows the hr - diagram with all these us members . from fig .
[ hrd_full.fig ] one can see that not only the majority of the low - mass stars , but also most of the intermediate- and high - mass stars lie close to or on the 5 myr isochrone .
there clearly is a considerable scatter around this isochrone that may seem to suggest a considerable spread of stellar ages .
however , it is very important to be aware of the fact that the masses and especially the ages of the individual stars read off from their position in the hr - diagram are generally not identical to their true masses and ages . for the case of us ,
the most important factor is the relatively large spread of individual stellar distances ( @xmath72 pc around the mean value of 145 pc ; * ? ? ?
* and priv . comm . ) in this very nearby and extended region , which causes the luminosities to be either over- or under - estimated .
another important factor is the presence of unresolved binary companions , which cause over - estimates of the luminosity .
further factors include photometric errors and variability , and the uncertainties in the calibrations used to derive bolometric luminosities and effective temperatures .
detailed discussions and simulations of these effects can be found in @xcite and @xcite .
the net effect of the uncertainties is that in the observed hr diagram a ( hypothetical ) perfectly coeval population of stars will _ not _ populate just a single line ( i.e. the corresponding isochrone ) , but will always display a finite spread , mimicking an age spread . @xcite and @xcite found via statistical modeling of these effects that the observed hr - diagram for the low - mass stars in us is consistent with the assumption of a _ common stellar age of about 5 myr ; there is no evidence for an age dispersion , although small age spreads of @xmath36 myr can not be excluded by the data_. @xcite showed that the derived age is also robust when taking into account the uncertainties of the theoretical pms models .
the mean age of 5 myr and the absence of a significant age spread among the us stars has been confirmed in the independent studies of @xcite and @xcite .
it is remarkable that the age derived for the low - mass stars is very well consistent with previous independent age determinations based on the nuclear and kinematic ages of the massive stars @xcite , which also yielded 5 myr .
this very good agreement of the _ independent _ age determinations for the high - mass and the low - mass stellar population shows that _ low- and high - mass stars are coeval _ and thus have formed together .
furthermore , the absence of a significant age dispersion implies that all stars in the association have formed more or less simultaneously .
this means that the star - formation process must have started rather suddenly and everywhere at the same time in the association , and also must have ended rather suddenly after at most a few myr . the star formation process in us
can thus be considered as a _ burst of star formation_. this will be discussed in more detail in sect .
[ sf_in_usco ] upper scorpius is an ideal target for an investigation of the initial mass function ( imf ) for several reasons .
first , the star formation process is completed and thus stars of all masses are already present ; since the molecular cloud has already been dispersed , the members are also rather easily observable .
second , the age of @xmath73 myr is young enough that nearly all members are still present ; only very few of the most massive members have already evolved away from the main - sequence , and these stars can be accounted for individually . therefore ,
_ the present day mass function _
( with the addition of the one member that already exploded as a supernova ) _ is identical to the initial mass function .
_ figure [ imf2.fig ] shows the empirical mass function for us as derived in @xcite . the best - fit multi - part power law function for the probability density distribution
is given by @xmath74 or , in shorter notation , @xmath75 = -0.9\pm0.2 $ ] , @xmath76 = -2.8\pm0.5 $ ] , @xmath77 = -2.6\pm0.3 $ ] .
the derived us mass function is compared to two different representations of the field imf .
the first one is the field imf representation suggested by @xcite , given by @xmath78 = -1.2\pm0.3 $ ] , @xmath79 = -2.7\pm0.5 $ ] , @xmath80 = -2.3\pm0.5 $ ] , the second is the parameterization for the average galactic field star imf in the solar neighborhood derived by @xcite , which is given by @xmath81 = -0.3\pm0.7 $ ] , @xmath82 = -1.3\pm0.5 $ ] , @xmath83 = -2.3\pm0.3 $ ] .
although the observed us mass function is not identical to either the @xcite or @xcite field imf representations , it is well within the ranges of slopes derived for similar mass ranges in other young clusters or associations , as compiled in @xcite .
therefore , it can be concluded that , within the uncertainties , the general shape of the us mass function is consistent with recent field star and cluster imf determinations .
the _ total stellar population _ of us in the @xmath84m@xmath33 to @xmath85m@xmath33 mass range can be described by the ` best fit ' mass function ( eqn .
[ imf_usco ] ) .
the integration of this function yields 2525 stars with a total mass of @xmath86m@xmath33 .
75% of all stars have masses below @xmath87m@xmath33 and contribute 39% of the total mass .
only 3% of all stars have masses above @xmath88m@xmath33 , but they contribute 33% of the total mass .
all these numbers are based on the primary star mass function . for a reasonable estimate of the total stellar mass one
has to take into account that most of the stars are probably in multiple systems . the binary frequency ( i.e. the probability that a given object is multiple ; cf . * ? ? ?
* ) for late type stars is at least about 50% ( c.f .
* ; * ? ? ?
* see khler et al .
2000 for the case of us ) and even higher for early type stars @xcite .
the total mass of the companions in the multiple systems can be estimated as follows : assuming a binary frequency of 100% and random pairing of secondaries from the same underlying mass function ( eqn.[imf_usco ] ) as the primaries , the total mass of all companions in the @xmath89m@xmath33 mass range is 40% of the total mass of all primaries . for the estimate of the total stellar mass one
also has to include the most massive stars , that have already evolved away from the main - sequence : antares ( @xmath90 ) and its b2.5 companion ( @xmath91 ) , the supernova progenitor ( @xmath35 ) , and @xmath12 oph ( @xmath92 ) .
the total stellar mass is then : @xmath93 .
while us has been the focus of many investigations over the past century , its older siblings the ucl and lcc regions have received much less attention .
there appear to be several biases which made these two groups more difficult to study , or perhaps more easy to ignore , at least historically .
these are : ( 1 ) ucl and lcc are not as concentrated as us , ( 2 ) a significant fraction of their memberships are within 10@xmath2815@xmath28 of the galactic plane , making separating their membership from the galactic population more difficult , ( 3 ) most of ucl and lcc are inaccessible to telescopes in the northern hemisphere , ( 4 ) and early investigations of pms stars centered on dark and reflection nebulae , which are in abundance in and near us , but more lacking or absent in ucl and lcc . only over the past decade
has the availability of high quality astrometry ( _ hipparcos _ , tycho , etc . ) and the _ rosat _ all - sky survey , enabled the efficient identification of low - mass ucl and lcc members . given the numbers of b - type stars in ucl and lcc , there are probably _ thousands _ of low - mass members awaiting discovery .
there appear to be no post - ms members of ucl , however there is a fairly well - defined main sequence turn - off near spectral type b1.5 @xcite .
the notable turn - off stars are , in approximate order of mass : @xmath94 sco ( b1.5iv ) , @xmath34 lup ( b1.5iii ) , @xmath18 lup ( b2iii ) , @xmath95 lup ( b1.5iv ) , @xmath96 cen ( b2iv ) , @xmath97 sco ( b2iv ) , and @xmath19 lup ( b1.5v ; where mk spectral types are from * ? ? ?
the positions of these massive stars are plotted and labeled in the ucl map in fig .
[ ucl_map.fig ] .
all of these stars were selected as _ hipparcos _ members of ucl by @xcite except for @xmath94 sco and @xmath18 lup .
@xcite has noted that the longer baseline act and trc proper motions for these stars are more conducive to ucl membership .
the tycho-2 proper motion for @xmath18 lup is well - pointed toward the ucl convergent point defined by @xcite , however its magnitude is somewhat larger than most other ucl members . using the @xcite evolutionary tracks , @xcite estimated the main sequence turn - off age of ucl to be @xmath317myr .
it was @xcite who first calculated the number of supernovae to have exploded in ucl ( @xmath36@xmath983 ) , and showed that the sum of the total kinetic energy predicted to be imparted by the supernovae and stellar winds of the deceased ucl members ( @xmath310@xmath99 erg ) is roughly consistent with the kinetic energy of the @xmath3100pc - radius expanding shell of hi centered on ucl .
this satisfactory agreement supports the original scenario by @xcite , that the massive stars in the oldest sco - cen subgroups ( ucl and lcc ) destroyed the proto - sco - cen molecular cloud complex , and radially dispersed most of the gas into what is seen today as large , expanding , loop - like h i structures centered on sco - cen . a modern estimate by the authors of the number of exploded supernovae in ucl ( @xmath37 sne ) , using the @xcite membership list for ucl , and adopting a @xcite initial mass function , corroborates de geus s ( 1992 ) prediction .
ucl contains a herbig ae star ( hd 139614 ) and two accreting fe stars ( ak sco and hd 135344 ) .
both hd 139614 and hd 135344 are close to the western edge of the lupus clouds near @xmath100 @xmath101 332@xmath28 ( see fig
. 8 of the lupus chapter ; comern , this volume ) , and so may be younger than the mean age for ucl . _ the pre-_rosat _ era : _
a small number of pms members of ucl were found before the arrival of the _ rosat _ x - ray surveys .
in the course of a radial velocity survey of b - type members of sco - cen , @xcite found that many of their visual companions shared similar radial velocities , i.e. they probably constituted physical binaries .
some of these objects were of afgk spectral types , and include hd 113703b ( k0ve ; lcc ) , hd 113791b ( f5v ; lcc ) , hd 143099 ( g0v ; ucl ) , and hd 151868 ( f6v ; ucl ) .
the high - mass primaries are still considered secure members of us , ucl , and lcc @xcite .
@xcite proclaimed `` _ [ t]hese observations do in fact represent the first to establish the presence of stars later than type a0 in the group .
they include one k - dwarf which is presumably no older than the group to which @xcite assigns an age of 20 million years _ '' .
@xcite noted strong li in the spectrum of hd 113703b , confirming its extreme youth .
the first two stars , along with hd 129791b ( k5ve ; ucl ) and hd 143939b ( k3ve ; ucl ) , are included in the well - studied `` lindroos '' sample of post - t tauri companions to massive stars ( e.g. * ? ? ?
* ; * ? ? ?
in an effort to tie the absolute magnitude calibration of b - type stars to that of later type stars , as well as explore the luminosity function of sco - cen , @xcite identified 27 candidate a- and f - type stars in a @xmath102 survey of ucl .
of these 27 candidates , 24 have _ hipparcos _ astrometry , and @xcite retained 14 of these as kinematic members of ucl . _ the _ rosat _ era : the lupus region surveys : _
most of the low - mass members of ucl which have been identified over the past decade have been due to _ rosat _ x - ray pointed observations and its all - sky survey .
the focus of most of these surveys was not ucl , but the lupus molecular cloud complex . in a wide - field survey of x - ray - luminous stars in a 230 deg@xmath37 region around the lupus clouds , @xcite identified 136 candidate t tauri stars .
while 47 of these new t tauri stars ( tts ) were found in pointed _ rosat _ observations of the lupus clouds , the majority ( 89 ) were found scattered over a wider region with the _ rosat _ all - sky survey ( rass ) .
@xcite found that if the `` off - cloud '' t tauri stars were co - distant with the lupus clouds , then their mean isochronal age was significantly older than the `` on - cloud '' lupus t tauri stars ( @xmath37myr vs. @xmath31 - 3myr , respectively ) .
@xcite hypothesized that the off - cloud tts formed in the lupus clouds , but were dispersed either due to a large intrinsic velocity dispersion in the clouds , or due to ejection @xcite .
@xcite also conducted a spectroscopic survey of rass sources west of the lupus clouds in a strip between 325@xmath28 @xmath103 @xmath100 @xmath103 335@xmath28 and @xmath104 @xmath103 @xmath105 @xmath103 @xmath106 ( _ dotted line _ in fig .
[ ucl_map.fig ] ) .
they identified 48 li - rich stars , most of which were concentrated between galactic latitudes + 8@xmath28 and + 22@xmath28 .
@xcite hypothesized that the majority of these objects were `` gould belt '' members , with ages of @xmath10360myr .
the idea that there are older , dispersed rass ttss near the lupus clouds becomes somewhat less surprising when it is appreciated that the lupus clouds are adjacent to the us and ucl subgroups of sco - cen ( with ages of 5 and @xmath317 myr , respectively , and thousands of predicted members ) .
sco - cen and ucl are not mentioned in the @xcite , @xcite , or @xcite surveys , although in this region of sky the `` gould belt '' is essentially _ defined _ by the high - mass stars of ucl ( e.g. * ? ? ?
* ) and the gas associated with the lupus clouds .
two of the clumps of ttss seen by wichmann et al .
( 1997b ) seem to be co - spatial with the over - densities of high - mass ucl members , associated with @xmath107 lup and @xmath108 lup .
a more detailed kinematic study of the krautter - wichmann ttss is needed to disentangle the star - formation of the region , explore the relation between the modern - day lupus clouds and the `` completed '' star - formation in ucl .
recently , @xcite investigated the kinematics of the @xcite tts in the lupus region . after correcting for the distribution of individual stellar distances , the color - magnitude diagram revealed two separate stellar populations with clearly different ages : a young ( @xmath41 myr ) population of tts which are closely concentrated at the lupus dark clouds , and an older ( @xmath109 myr ) population of stars which are much more widely dispersed in the area .
based on kinematic arguments , makarov concludes that `` it is unlikely that the t tauri members were born in the same star - forming cores as the more compactly located classical t tauri stars . ''
we agree with this assessment , but add that the positions , motions , and _
mean _ age of the outlying members are very well consistent with the idea that the @xcite stars are in fact low - mass members of the us and ucl subgroups of sco - cen . rather than subsume all of the rass stars into a `` lupus association '' as @xcite elects , it is probably wise for astrophysical studies to separate the on- or near - cloud lupus members from the off - cloud ucl / us members .
we have visually attempted to do this with a dashed line box in fig . 6 , but clearly more study is warranted .
there is good kinematic evidence that the @xcite rass stars are mostly ucl members . a preliminary kinematic analysis by e.m .
suggests that for the 33 @xcite _ rosat _ stars with either tycho-2 @xcite or ucac2 @xcite proper motions , _ all but four have motions consistent with ucl membership_. the _ rosat _ x - ray stars rx j1412.2 - 1630 , j1419.3 - 2322 , j1509.3 - 4420 , and j1550.1 - 4746 can all be rejected as kinematic ucl members .
the first two objects are also well outside of the ucl region defined by @xcite , so their rejection is perhaps unsurprising .
the inferred cluster parallax distances for the rest range from @xmath390pc to @xmath3200pc , suggesting considerable depth to the ucl subgroup . of note for future li depletion studies of ucl members , we find that the m - type wichmann et al .
stars rx j1512.8 - 4508b ( m1 ) , rx j1505.4 - 3716 ( m0 ) , and rx j1457.3 - 3613 all have proper motions consistent with membership in ucl .
the components of the wide m - type binary rx j1511.6 - 3249a and b ( m1.5+m1.5 ; @xmath110 separation ) also have proper motions consistent with ucl membership .
_ the rosat era : the mamajek et al .
survey : _
@xcite conducted the first wide - field , spectroscopic survey searching explicitly for pms gk - type members of ucl .
they selected 56 ucl candidates by cross - referencing proper motion - selected stars from the kinematic study of ( * ? ? ? * but with color - magnitude constraints more amenable to identifying pms g- and k - type members from the act and trc astrometric catalogs ) with x - ray sources from the _ rosat _ all - sky survey bright star catalog @xcite .
they also measured optical spectra of 18 gk - type _ hipparcos _ stars selected as probable kinematic members by @xcite .
blue and red optical spectra of the candidates were taken with the dbs spectrograph on the siding springs 2.3-m telescope , with resolution of @xmath32.8 in the blue , and @xmath31.3 in the red .
stars were classified as pms by virtue of having strong li ( stronger than the loci of @xmath330100 myr - old clusters ) and subgiant luminosity classes ( measured via the sr ii @xmath1074077 line ) . between ucl and lcc ,
the x - ray and proper motion selection was @xmath393% efficient at selecting pms stars , while 73% of the kinematic candidates selected by @xcite were pms ( well in line with the @xmath330% contamination rate predicted by those investigators ) .
@xcite confirmed 12 out of 18 of the @xcite proper motion - selected ucl candidates as pms , and found 50 li - rich ucl members from the x - ray and proper motion - selected sample .
none of the ucl pms stars identified could be considered classical t tauri candidates ( i.e. accreting ) , however a few demonstrated small h@xmath34 emission excesses ( all @xmath1032 ) , probably due to enhanced chromospheric activity . although @xcite found no new classical t tauri ucl members ,
there are two known f - type accretors ( ak sco and hd 135344 ; e.g. * ? ? ?
the mean age of the pms ucl members in the @xcite survey depends , unsurprisingly , on one s choice of evolutionary tracks , but ranged from 1522 myr ( see fig .
[ hrd_ucl_lcc.fig ] ) . using the evolutionary tracks of (
* where appropriate ) , and accounting for the magnitude limit of the @xcite survey ( see sect . 7.1 of their paper )
, one finds a median age for the ucl pms stars of 16 myr . while mamajek et al.s survey of ucl was very broad , admittedly it was not very deep , and more members could be easily identified with existing catalog data ( e.g. rass faint source catalog , ucac2 , sacy , etc . ) .
@xcite claim that the majority ( @xmath380% ) of the @xmath31.11.4m@xmath33 members have probably been identified , but that @xmath32000 @xmath1031m@xmath33 members likely await discovery .
the lowest - mass ucl candidates which have been found ( and which are not in the immediate vicinity of the lupus star - forming clouds ) are the @xcite li - rich early - m - type stars rxj1514.0 - 4629a ( lr lup ; m2 ) and rxj1523.5 - 3821 ( mn lup ; m2 ) .
the equivalent widths of the lii @xmath1076707.7 line for these m2 stars ( @xmath30.38 ) , as measured in low - resolution spectra by @xcite , _ suggests that modest li depletion is taking place among the early - m stars in ucl_. a high resolution spectroscopic survey of li - rich m - type members could place independent constraints on the age of ucl , as well as allow an interesting comparison between li depletion ages and main sequence turn - off ages . _ a list of low - mass ucl members : _ we have assembled a membership list of candidate low - mass members of ucl from the previously mentioned literature ( table [ table_ucl_list ] ) .
the list is meant to be rather exclusive in that it selects only those low - mass ( gkm ) stars that are li - rich _ and _ have kinematics consistent with group membership .
all are known x - ray sources , except for a small number of proper motion candidates from @xcite which were confirmed to be li - rich by @xcite .
the f stars in ucl have not been thoroughly investigated spectroscopically .
given the very high efficiency of selecting pms stars amongst x - ray and proper motion selected gk stars ( * ? ? ?
* % ) , we list only the proper motion - selected f - type candidate members from the literature which have rosat all - sky survey x - ray counterparts within 40 of their optical positions .
the sample is not meant to be complete , by any means , but should represent a relatively clean sample of @xmath1112m@xmath33 ucl members .
the incidence of non - members is probably well below @xmath1115% .
two confirmed pms companions to high mass ucl members ( hd 129791b and hd 143939b ) were included , using data from @xcite , @xcite , @xcite , and distances ( for the primaries ) from @xcite .
although the majority of @xcite and @xcite stars appear to be bona fide ucl members ( by virtues of their positions , proper motions , appropriate hr - d positions , etc . ) , we have not included them in our membership list , but refer the reader to those papers . in table
[ table_ucl_list ] we have flagged the ucl members which are near the lupus molecular clouds .
we have defined a box around the lupus clouds that are actively forming stars : 335.5@xmath28 @xmath103 @xmath100 @xmath103 341@xmath28 and + 7@xmath28 @xmath103 @xmath105 @xmath103 + 18@xmath28 .
_ the population of stars in the off - cloud region in that box may represent a mix of stars that have recently formed in the lupus clouds , along with older ucl pms stars_. the hrd positions of the flagged ucl stars in table [ table_ucl_list ] are consistent with having ages of @xmath3723 myr , so they are probably ucl members projected against ( or behind ? ) the lupus clouds .
lcc straddles the galactic equator in crux , stretching from galactic latitude to @xmath112 between centaurus , crux , and musca ( see fig .
[ lcc_map.fig ] ) .
although lcc is the closest recognized ob association subgroup to the sun ( @xmath113 pc ) , it is the least studied of the sco - cen regions .
there is some hint of substructure in the group , and it appears that the northern part of the group is somewhat more distant , older , and richer ( @xmath317 myr , 120 pc ) than the southern part of the group ( @xmath312 myr , 110 pc ) . the upper main sequence of lcc is poorly defined , and so the group has the least secure turn - off age of the sco - cen subgroups . in approximate order of mass ,
the lcc turnoff stars with fairly secure kinematic membership @xcite are : @xmath95 cru ( b2iv ) , @xmath57 cen ( b2v ) , @xmath34 mus ( b2iv - v ) , @xmath94 cru ( b2iv - v ) , and @xmath114 cen ( b1.5v ; mk types from * ? ? ?
there are other massive stars in this region , whose membership in lcc has been debated , but whose _ hipparcos _ proper motions were _ inconsistent _ with membership @xcite . among these systems
are the `` _ super cen - crux six _ '' : @xmath34 cru ( b0.5iv+b1v+b4iv ) , @xmath18 cru ( b0.5iii ) , @xmath18 cen ( b1iii ) , @xmath95 cen ( b2ivne ) , @xmath20 cen ( b1iii ) , and @xmath30 cen ( b3v ; * ? ? ?
* ) . as noted by @xcite for the most massive subset of these , _ these stars are @xmath31020m@xmath33 , with inferred ages of @xmath3515 myr and distances of @xmath3100150pc .
such stars are extremely rare , and their presence in the lcc region appears to be more than coincidental . _ as noted by @xcite , all of these objects are flagged as having unusual motions due to either binarity or variability in the _ hipparcos _ catalog , however they only consider @xmath95 cen , @xmath18 cru , and @xmath30 cen to be probable lcc members whose _ hipparcos _ proper motions excluded them from kinematic selection . in a follow - up paper , @xcite demonstrated that the long - baseline proper motions for @xmath30 cen ( act catalog ) and @xmath18 cru ( trc catalog ) are more consistent with lcc membership , however they did not address the remainder of the `` six '' . a more thorough kinematic investigation of the center - of - mass motions of these systems is needed .
although the sco - cen subgroup memberships seem to have a @xmath311.5 km s@xmath115 internal velocity dispersion @xcite , if significant dynamical evolution has taken place among the massive multiple systems over the past @xmath31015 myr ( i.e. due to supernovae , or ejections from short - lived trapezia ) , then the resultant kicks could explain the presence of these massive multiple systems with deviant motions in the vicinity of the sco - cen subgroups .
the results from a kinematic study of these `` super six '' , as well as the runaway star and pulsar candidates in the vicinity ( e.g. * ? ? ?
* ) could have important implications for the imf and star - formation history of lcc .
lcc contains two known herbig ae / be stars : hd 100453 ( a9ve ) and hd 100546 ( b9ve ) . although sometimes labeled `` isolated '' haebe stars , their positions , proper motions , and distances ( @xmath116 pc ) are all consistent with lcc membership .
estimated an age of @xmath11710 myr for hd 100546 , consistent with other lcc members .
circumstellar pah emission was recently resolved around both stars using vlt @xcite , while the disk for hd 100546 was resolved in the thermal ir at magellan with mirac @xcite . @xcite
recently identified a faint candidate companion to hd 100453 at @xmath118 separation consistent with being an m - type pms star , however it has not been spectroscopically confirmed .
hd 100546 has a gap in its disk consistent with the presence of a substellar object at @xmath119 au @xcite .
the low - mass population of lcc has been investigated even less than that of ucl .
low - mass members were identified serendipitously through surveys of the regions near @xmath18 cru @xcite and the tw hya association @xcite , which is near the northwest corner of the lcc box defined by @xcite , but at roughly half the distance ( @xmath350pc ; * ? ? ?
* ) as lcc ( @xmath3118pc ) .
@xcite conducted the only systematic survey , thus far , whose goal was to identify low - mass members over the whole lcc region . _ the park - finley _ rosat _ stars near @xmath18 cru : _
@xcite identified 6 unknown , variable x - ray sources near the b0.5iii star @xmath18 cru in a _
pspc pointing . by virtue of their x - ray spectral fits , variability , and x
- ray to optical flux ratios , @xcite conjectured that the 6 objects were t tauri stars _
( simbad ) or _ cru _ @xcite .
we adopt the simbad nomenclature . ] , and that they had discovered a previously unknown star - forming region centered on @xmath18 cru .
a low - resolution spectroscopic survey by @xcite found that the x - ray stars were li - rich , late - type stars , consistent with classification as weak - lined t tauri stars . after rejecting several hypotheses regarding the origins of these young stars
, @xcite argued that park & finley serendipitously uncovered the first known `` isolated '' low - mass members of lcc .
@xcite conducted a high - resolution spectroscopic study of the park - finley stars , and confirmed that 4 of the 6 are sufficiently li - rich to be classified as pms .
@xcite also find that the radial velocities and isochronal ages ( @xmath3510 myr ) = 110pc ( the distance to @xmath18 cru ) .
the reason for the discrepancy in ages between the park - finley sample and the mamajek et al .
sample is unknown , but could be due to the unknown distances to the park - finley stars . ] for the 4 most li - rich stars are roughly consistent with lcc membership .
the 4 pms stars are k5-m4 in spectral type , and have li abundances intermediate between those of t tauri stars and the @xmath350-myr - old ic 2602 cluster . _
the discovery of more m - type sco - cen members may prove to be an interesting means of testing pms evolutionary tracks and li depletion models_. _ twa _ rosat _ stars in lcc : _ in their attempt to identify new members of the nearby tw hya association ( twa ; age @xmath101 10myr ; @xmath120 @xmath101 50pc ) ,
@xcite conducted a spectroscopic survey of rass bsc x - ray sources in the vicinity of the famous debris disk star hr 4796 .
they identified eight new t tauri stars , however they were different from the rest of the twa members thus far found : their optical , infrared , and x - ray fluxes were significantly dimmer .
@xcite claimed that the new twa stars ( # 14 - 19 ) were further away from the original twa # 1 - 13 sample , with distances of perhaps 70 - 100pc .
as pointed out in @xcite , the positions of twa 14 - 19 overlap with the lcc region defined by @xcite .
@xcite have presented evidence that the distribution of rotational periods of twa 1 - 13 and 14 - 19 are very different , suggesting two different populations . in a detailed kinematic investigation of ` twa' stars in the literature ,
@xcite concluded that the twa stars are dominated by two populations : a group of two dozen stars with distances of @xmath349@xmath9812pc ( what probably constitutes the true `` tw hya association '' ) , and a subset of objects with distances of @xmath3100150pc , which are likely lcc members ( partially corroborating the lawson & crause findings ) .
@xcite claims that twa 12 , 17 , 18 , 19 , and 24 are very likely lcc members , with twa 14 a borderline case .
twa 19 (= hip 57524 ) was identified as an lcc member by @xcite , as was twa 24 (= mml 5 ) by @xcite . _
the mamajek et al .
rosat survey : _ as discussed at length in sect .
4.2 , @xcite conducted a search for @xmath31m@xmath33 members of ucl and lcc amongst an x - ray and proper motion - selected sample .
they identified 37 lcc members , and confirmed the youth of 10 ( out of 12 ) of the @xcite _ hipparcos _ g - type candidates they observed .
@xcite claimed that the classical t tauri star pds 66 @xcite is actually a lcc member , and appears to be the only known accreting low - mass star in lcc .
the mean age of the pms lcc members in the @xcite survey ranged from 17 to 23 myr , depending on the choice of evolutionary tracks . using the evolutionary tracks of (
* where appropriate ) , and accounting for the magnitude limit of the mamajek et al .
survey , one finds a median age for the lcc pms stars of 18 myr . _ torres et al .
`` sacy '' rosat survey : _
@xcite presented results for a high resolution spectroscopic survey of 1151 stars in the southern hemisphere with _
rosat _ all - sky survey x - ray counterparts .
results for the `` sacy '' ( search for associations containing young stars ) survey regarding nearby young low - density stellar groups are reported elsewhere in this volume ( torres et al . ) .
there are numerous li - rich late - type sacy stars in the sco - cen region , and most have proper motions and radial velocities suggestive of membership to the sco - cen groups . for this review ,
we have only attempted to assign membership of sacy stars to the lower centaurus - crux group , although the sacy catalog no doubt contains many new ucl and us members as well .
the selection of lcc members from the sacy catalog will be discussed in more detail by mamajek ( in prep . ) . to construct a sample of lcc members in the sacy catalog ,
we start with the 138 sacy stars that lie within the @xcite boundary for lcc . of these objects ,
45 are previously known lcc members found either by @xcite or @xcite .
we further prune the sacy sample by removing giants , and selecting only those that are li - rich ( ew(li 6707 ) @xmath117 100m ) and that have proper motions within 25 masyr@xmath115 of the @xcite mean value for lcc .
lastly we run a convergent point algorithm from @xcite on the full sample of @xcite and @xcite members along with the remaining 49 sacy objects .
following @xcite , we assume an internal velocity dispersion of 1.14 kms@xmath115 , and calculate cluster parallax distances using the velocity vector of @xcite . in total
we identify 45 sacy stars as probable new members of lcc .
the median rv for the sacy stars selected as lcc members is @xmath121 km s@xmath115 , which is nearly identical to the subgroup mean rv from @xcite ( @xmath122 km s@xmath115 ) . _ a list of low - mass lcc members :
_ table [ table_lcc_list ] provides a modern catalog of low - mass lcc members , and was constructed in a similar manner as that for ucl ( sect .
[ ucl_low_mass ] ) .
the positions of known lcc members are plotted in fig .
[ lcc_map.fig ] .
individual cluster parallax distances were adopted from @xcite or @xcite .
where no published cluster parallax distance was available , we calculate new values using the @xcite motion vector for lcc , while using the proper motions for the twa objects listed in @xcite and from the sacy catalog @xcite .
the us group has been particularly well - studied , and can give us some quantitative insight into the star formation history as well as constraints on the mechanism responsible for triggering the star formation .
as described in detail in sect . [ us_ages ] , the populations of the high - mass as well the low - mass stars both have a common age of 5 myr .
there is no evidence for a significant age spread , and the data are consistent with the idea that all stars have formed within a period of no more than @xmath123 myr .
another important aspect is the initial configuration of the region at the time when the stars formed .
today , the bulk ( 70% ) of the hipparcos members ( and thus also the low - mass stars ) lie within an area of 11 degrees diameter on the sky , which implies a characteristic size of the association of 28 pc .
showed that the internal 1d velocity dispersion of the hipparcos members of us is only 1.3 km / s . in combination with the well determined age of us and the present - day size
, this strikingly small velocity dispersion clearly shows that us can not have originated in a compact cluster configuration that expanded later , but must have been in a spatially extended configuration from the beginning .
us seems to have formed in an extended , unbound giant molecular cloud , similar to the models considered by @xcite .
the initial size of the association can be estimated by assuming that the stars expanded freely from their initial positions . with a gaussian velocity distribution characterized by the measured velocity dispersion , a single point in space
would have expanded to a size of about 13 pc in 5 myr .
subtracting this number from the current characteristic size of 28 pc in quadrature leads to an initial size of 25 pc .
this implies an lateral stellar crossing time of 25 pc / 1.3 km / s @xmath3 20 myr in the initial configuration .
this large crossing time is in remarkable contrast to the upper limit on any possible age spread among the association members of only @xmath123 myr .
the fact that the lateral stellar crossing time is much ( about an order of magnitude ) larger than the age spread of the association members clearly suggests that some external agent must have coordinated the onset of the star formation process over the full spatial extent of the association . in order to account for the small spread of stellar ages ,
the triggering agent must have crossed the initial cloud with a velocity of at least @xmath124 km / s .
also , some mechanism must have terminated the star formation process about 1 myr after it started .
finally , we note that the us region does no longer contain significant amounts of molecular cloud material . the original molecular cloud ,
in which the stars formed , has been nearly completely dispersed . today
, most of this material appears to be situated in an expanding , ( mostly ) atomic hi superbubble centered on us . estimated the mass of this superbubble to be @xmath125 . comparing this to the estimated total mass of all stars in us ( @xmath126 ;
see sect .
[ us_imf ] ) suggests that only a small fraction of the initial cloud mass was transformed into stars .
these findings can be understood as consequences of the feedback from massive stars .
high - mass stars , above about ten solar masses , profoundly affect their environment in several ways .
their strong ionizing radiation can photoevaporate molecular cloud clumps ( e.g. see hester et al .
1996 and mccaughrean & andersen 2002 for the case of m16 ) and circumstellar matter around young stellar objects ( see * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
their powerful stellar winds deposit considerable amounts of momentum and kinetic energy into the surrounding medium .
finally , supernova explosions cause strong shock waves that transfer typically some @xmath127 erg of kinetic energy into the ambient interstellar medium .
the supernova blast wave will initially expand within the wind - blown bubble formed by the supernova progenitor ; as it catches up with the bubble shock front , it will accelerate the expansion of the bubble ( see e.g. * ? ? ?
* ) , further disrupt the parental molecular cloud ( see e.g. * ? ? ?
* ) and sweep up a massive shell of dust and gas . in general , massive stars have a very destructive effect on their nearby environment ; they can disrupt molecular clouds very quickly and therefore prevent further star formation in their surroundings . at somewhat larger distances , however , the wind- and supernova - driven shock waves originating from massive stars can have a constructive rather than destructive effect by driving molecular cloud cores into collapse .
several recent numerical studies ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) have found that the outcome of the impact of a shock wave on a cloud core mainly depends on the type of the shock and its velocity : in its initial , adiabatic phase , the shock wave is likely to destroy ambient clouds ; the later , isothermal phase , however , is capable of triggering cloud collapse if the velocity is in the right range .
shocks traveling faster than about 50 km / s shred cloud cores to pieces , while shocks with velocities slower than about 15 km / s usually cause only a slight temporary compression of cloud cores .
shock waves with velocities in the range of @xmath128 km / s , however , seem to be well able to induce collapse of molecular cloud cores .
a good source of shock waves with velocities in that range are expanding superbubbles driven by the winds and supernova explosions of massive stars at distances between @xmath124 pc and @xmath129 pc ( see * ? ? ?
observational evidence for star forming events triggered by shock waves from massive stars has , for example , been discussed in @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , and @xcite ; see also the discussions in @xcite and @xcite . for the star burst in us , a very suitable trigger is the shock - wave of the expanding superbubble around the ucl group , which is driven by the winds of the massive stars and several supernova explosions that started to occur about 12 myr ago .
the structure and kinematics of the large h i loops surrounding the scorpius - centaurus association suggest that this shock wave passed through the former us molecular cloud just about 5 myr ago @xcite .
this point in time agrees very well with the ages found for the low - mass stars as well as the high - mass stars in us .
furthermore , since the distance from ucl to us is about 60 pc , this shock wave probably had just about the right velocity ( @xmath130 km / s ) that is required to induce star formation according to the modeling results mentioned above .
thus , the assumption that this wind- and supernova - driven shock wave triggered the star formation process in us provides a self - consistent explanation of all observational data .
+ a scenario for the star formation history of us consistent with the observational results described above is shown in fig .
[ sfm ] . the shock - wave crossing us
about 5 myr ago initiated the formation of some 2500 stars , including 10 massive stars upwards of @xmath131 .
when the new - born massive stars ` turned on ' , they immediately started to destroy the cloud from inside by their ionizing radiation and their strong winds .
this affected the cloud so strongly that after a period of @xmath132 myr the star formation process was terminated , probably simply because all the remaining dense cloud material was disrupted .
this explains the narrow age distribution and why only about 3% of the original cloud mass was transformed into stars .
about 1.5 myr ago , the most massive star in us , presumably the progenitor of the pulsar psr j1932 + 1059 ( see * ? ? ? * ; * ? ? ?
* ) , exploded as a supernova and created a strong shock wave , which fully dispersed the us molecular cloud and removed basically all the remaining diffuse material .
finally , it is interesting to note that this shock wave must have crossed the @xmath30 oph molecular cloud within the last 1 myr @xcite .
the strong star formation activity we witness right now in the l 1688 cloud of the @xmath30 oph region might therefore be triggered by this shock wave ( see * ? ? ?
* ) and would represent the third generation of sequential triggered star formation in the scorpius - centaurus - ophiuchus complex .
furthermore , we note that the lupus 1 & 2 dark clouds are also located near the edge of the shell around us .
these two dark clouds ( see chapter by comern in this book ) contain numerous young stellar objects with estimated ages of @xmath132 myr ; their recent star formation activity may also well have been triggered by the passage of the shock related to the expanding shell around us ( see , e.g. , * ? ? ?
_ an updated model for triggered cloud and star formation : _ while the scenario described above provides a good explanation of the star formation history , a potential problem is its implicit assumption that the us and @xmath30 oph molecular clouds existed for many myr without forming stars before the triggering shock waves arrived ( otherwise one should see large age spreads in the stellar populations , contrary to the observational evidence ) .
numerous recent studies ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) have provided increasing and convincing evidence that the lifetime of molecular clouds is much shorter than previously thought , and the whole process of molecular cloud formation , star formation , and cloud dispersal ( by the feedback of the newly formed stars ) occurs on timescales of only a few @xmath133 ) myr .
it is now thought that molecular clouds form by the interaction of flows in the interstellar medium that accumulate matter in some regions .
as soon as the column density gets high enough for self - shielding against the ambient uv radiation field , the atomic gas transforms into molecular clouds .
this new paradigm for the formation and lifetime of molecular clouds seems to invalidate the idea of a shock wave that hit a pre - existing molecular cloud and triggered star formation .
nevertheless , the basic scenario for the sequence of processes in sco - cen may still be valid .
as pointed out by @xcite , wind and supernova shock waves from massive stars are an important kind of driver for ism flows , and are especially well suited to create _ coherent large - scale flows_. only large - scale flows are able to form large molecular clouds , in which whole ob associations can be born . an updated scenario for sco - cen could be as follows : initially , the winds of the ob stars created a superbubble around ucl .
the interaction of this expanding superbubble with flows in the ambient ism started to sweep up clouds in some places .
when supernovae started to explode in ucl ( note that there were presumably some 6 supernova explosions in ucl up to today ) , these added energy and momentum to the wind - blown superbubble and accelerated its expansion . the accelerated shock wave ( now with @xmath134 km / sec ) crossed a swept - up cloud in the us area , and the increased pressure due to this shock triggered star formation .
this scenario does not only explain the temporal sequence of events in a way consistent with the ages of the stars and the kinematic properties of the observed hi shells .
the following points provide further evidence : ( 1 ) the model ( see fig . 3 in * ?
* ) predicts that stellar groups triggered in swept - up clouds should move away from the trigger source .
a look at the centroid space motions of the sco - cen subgroups @xcite actually shows that us moves nearly radially away from ucl with a velocity of @xmath135 km / sec .
furthermore , ( 2 ) a study by @xcite revealed that several young stellar groups , including the @xmath19 cha cluster , the tw hydra association , and the young stars associated with the cra cloud , move away from ucl at velocities of about 10 km / sec ; tracing their current motions back in time shows that these groups were located near the edge of ucl 12 myr ago ( when the supernova exploded ) .
( 3 ) the model also predicts that molecular clouds are most efficiently created at the intersection of two expanding bubbles .
the lupus 1 cloud , which is located just between us and ucl , seems to be a good example of this process .
its elongated shape is very well consistent with the idea that it was swept up by the interaction of the expanding superbubble around us and the ( post - sn ) superbubble created by the winds of the remaining early b stars in ucl . in the ucl and lcc subgroups
, @xcite found that the hr diagram positions of the stars in both groups are consistent with 95% of their low - mass star - formation occurring within a @xmath136 myr span .
the _ upper limit _ on the duration of star - formation is relatively large due to the fact that most of the stars from the @xcite survey were on the radiative portion of the pms evolutionary tracks ( where the isochrones are packed closely together ) , and the binarity of these stars is unstudied .
although @xcite did account for the spread in distances of the subgroup members by using cluster parallaxes , the previously mentioned effects conspire to make it difficult to be more precise about the group star - formation history at this time .
the mean ages of the pms populations , using the @xcite tracks ( 16 myr for ucl , 18 myr for lcc ) , and other sets of models ( 15 - 22 myr for ucl , 17 - 23 myr for lcc ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) agree fairly well with the revised turn - off ages ( 17 myr for ucl , 16 myr for lcc ; using tracks of * ? ? ?
we note two age trends in lcc that are worthy of future investigation .
first , there appears to be a mass - dependence on the derived age for lcc .
the turn - off age for the b - stars is @xmath316 myr ( but see discussion in mamajek et al .
2002 ) , while the median ages of the cool members listed in table 4 are 17 myr for the f stars , 17 myr for the g stars , 12 myr for the k stars , and 4 myr for the m stars .
the 8 m - type members of lcc listed in table 4 all have isochronal ages in the range 2 - 7 myr .
the isochronal age for the m - type members is inconsistent with the positions of the main sequence turn - on ( @xmath3f5 ) and turn - off points ( @xmath3b1 ) .
we do not believe that this is evidence for non - coeval formation for the low mass vs. high mass stars , but instead attribute the difference to errors with the evolutionary tracks and/or incompleteness of the member sample among the faintest ( i.e. apparently older ) low - mass members .
there also appears to be a trend in median age vs. galactic latitude in lcc , which is probably due to substructure .
the median age of the low - mass stars south of the galactic equator is 12@xmath982 myr , while the stars north of the equator have a median age of 17@xmath981 myr .
the observed dispersion in the ages , characterized by the 68% confidence intervals , is @xmath37 myr , however much of this is probably due to observational uncertainties and binarity .
the _ hipparcos _
parallax data for the @xcite members of lcc show that the mean distance to the southern part of the group is closer ( 109@xmath981 pc ) than the northern part ( 123@xmath982 pc ) .
immediately south of lcc at the same distance ( @xmath3110 pc ) is the @xmath20 cha region which formed stars as recently as @xmath33 - 6 myr ago .
could we be seeing evidence for a north - south axis of triggered star - formation ?
in this picture , star - formation would have started @xmath317 myr ago in northern lcc near @xmath57 , @xmath30 , and @xmath95 cen , progressed southward through the southern cross forming @xmath34 and @xmath18 cru and the southern part of lcc some @xmath312 myr ago , and progressing further southward to form the @xmath20 and @xmath19 cha groups some @xmath36 myr ago . in comparison to the more elegant picture of star - formation in the smaller upper sco region , ucl and lcc
probably have more complex histories .
the spatial distributions of their members suggest that they contain substructure ( and may warrant further subdivision ) , and are inconsistent with being the evaporating , expanding remnants of _ two _ massive embedded clusters . instead , we suspect that the bulk of star - formation in ucl and lcc proceeded @xmath310 - 20 myr ago in a series of multiple embedded clusters and filaments , containing tens to hundreds of stars each , in a dynamically unbound giant molecular cloud complex
. members of the sco - cen association have been the target of several studies concerning the multiplicity of the stars . @xcite and
@xcite carried out k - band speckle observations of more than 100 x - ray - selected weak - line t tauri stars in sco - cen .
the derived multiplicity of the stars they observed is @xmath137 times higher than typical for main - sequence stars ; almost all of the brighter m - type pms stars are binaries .
@xcite examined 115 b - type stars in sco - cen for the existence of visual companions in the near - infrared .
they derived a @xmath137 times higher binary fraction than found for low - mass dwarfs in the solar neighborhood and in open clusters in the same separation range .
@xcite performed near - infrared adaptive optics surveys of 199 a - type and late b - type hipparcos members of sco - cen with the aim to detect close companions .
they found 176 visual companions among these stars and conclude that 80 of these are likely physical companion stars .
the estimated masses of these companions are in the range from @xmath138 to @xmath139 .
the comparison of their results with visual , spectroscopic , and astrometric data on binarity in sco - cen suggests that each system contains on average at least 0.5 companions ; this number is a strict lower limit due to the likely presence of further , yet undiscovered companion stars .
@xcite performed a high - resolution imaging survey of 12 brown dwarfs and very low - mass stars in us with the advanced camera for surveys on hst .
this survey discovered three new binary systems , and lead to an estimated binary fraction of @xmath140 , which is consistent with that inferred for higher mass stars in us .
@xcite searched for planetary - mass objects and brown dwarfs around nine low - mass members of us , but found no convincing cases of such companions .
@xcite discovered a wide ( @xmath141 au ) , low - mass binary system in us with estimated masses of @xmath142 for both components .
he concluded that this new system further establishes that the formation of low - mass stars and brown dwarfs does not require ejection from multiple systems and that wide , low - mass binaries can form in ob associations as well as in smaller clusters .
@xcite performed an adaptive - optics imaging survey for multiple systems among 58 m - type members of us and resolved nine pairs with separations below @xmath118 .
one of these systems is probably a young brown dwarf binary system .
there appear to be several sco - cen objects which may be of interest in the broader context of testing stellar evolution theory , as well as refining estimates for the age and metallicity of the subgroups .
the most interesting objects in this regard are well - characterized binary systems ( mainly eclipsing systems ) , and low - mass stars which have demonstrated li depletion .
ucl has at least two eclipsing binary systems : @xmath94 sco and gg lup .
the eclipsing binary @xmath94 sco is a well - studied massive binary ( 12.8m@xmath33 + 8.4m@xmath33 ; b1.5v+b3 ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , and has been recognized as a sco - cen member for many decades .
we will briefly discuss some low - mass binaries in more detail .
although the eclipsing binary gg lup ( b7v+b9v ; * ? ? ? * ) was assigned ucl membership by @xcite , its membership has not been recognized in the binary modeling literature . within 0.5@xmath28 of gg lup , and having nearly identical proper motions ( also suggestive of ucl membership ) , are the _ rosat _ t tauri stars tts 26 and 27 ( g8 and k0 types ; * ? ? ?
* ) , the a8/9v star hd 135814 , and the ms turn - off star @xmath95 lup ( b2iv ) . by comparing the parameters of gg lup to the evolutionary tracks of @xcite , @xcite claim that the system must have a somewhat sub - solar metallicity ( z @xmath101 0.15 ) , with an age of @xmath320myr .
@xcite derived ages of 15@xmath986 myr ( no convective overshoot ) and 17@xmath143 myr ( with overshoot ) for the system , however work by @xcite was only able to constrain the age to be @xmath10350myr , with a wide range of possible metallicities ( z @xmath144 0.014 - 0.037 ) .
the agreement between the model ages for gg lup with the quoted mean age of the low - mass pms ucl members ( 1522 myr ; * ? ? ?
* ) , and recent estimates of the ucl turn - off age ( @xmath317 myr ; * ? ? ?
* ) , are surprisingly good . further observations and modeling of gg lup and its ucl neighbors could further improve our knowledge of the age and metallicity of ucl .
another well - constrained binary , however non - eclipsing , is ak sco @xcite .
@xcite use the _ lack _ of observed eclipses for this equal mass system ( two f5 stars ) , along with the _ hipparcos _ parallax ( @xmath145 = [email protected] ) , to constrain the masses of the components ( [email protected] @xmath146 ) .
the hr - d positions for the ak sco a and b , individually , correspond to ( logt@xmath147 , logl / l@xmath33 = 3.807 , 0.76 ) . using the evolutionary tracks of @xcite ,
this corresponds to an age of 11 myr and 1.5m@xmath33 .
the mass is some 2@xmath57 higher than alencar et al.s dynamic estimate , but still statistically consistent .
all the more remarkable , ak sco demonstrates spectroscopic and photometric evidence for a circumstellar accretion disk .
ak sco , along with pds 66 , appear to be rare examples of stars which can retain accretion disks for @xmath11710myr .
@xcite reported the discovery of a low - mass spectroscopic binary , uscoctio 5 ( spectral type m4 ; * ? ? ?
* ) , in us . the lower limit to the system mass derived from the orbit fit is higher than the mass suggested by pms models based on the luminosity and effective temperature , suggesting either problems of the theoretical pms evolutionary models or uncertainties in the empirical spectral type and/or temperature determination .
another important aspect is how circumstellar disks evolve and whether their evolution is affected by environmental conditions .
some interesting results on the time scales for mass accretion and disk dissipation have been obtained for sco - cen members .
first , it was found that the fraction of accreting stars ( as traced by strong h@xmath34 emission ) is @xmath148 in us @xcite , but @xmath103few% in ucl and lcc @xcite .
a recent _ spitzer _ study by @xcite of us members across the whole stellar mass spectrum has found that circumstellar disk evolution appears to be very mass dependent .
roughly 20% of the @xmath149 ( k / m - type ) members of us were found to have infrared seds consistent with primordial accretion disks , whereas _ none _ of the @xmath150 ( f / g - type ) stars , and only one ( of 61 ) of the massive a / b - type , showed signs of having accretion disks .
a small fraction ( @xmath320% ) of the a / b - type massive stars did show evidence for small excesses at 16@xmath151 m , which could be due to cold , dusty debris disks .
a _ spitzer _
study by @xcite investigated infrared excesses tracing dust at @xmath3 340 au from f / g - type stars from @xcite .
the mips observations of 40 f- and g - type proper motion members of sco - cen detected 14 objects that possess @xmath152 m fluxes @xmath153 larger than their predicted photospheres , tentatively corresponding to a disk fraction of @xmath154 , including seven objects that also possess @xmath155 m excesses @xmath156 times larger than their predicted photospheres .
20% of the targets in us , 9% in ucl , and 46% in lcc display @xmath152 m excesses .
another critical area of stellar evolution theory which studies of sco - cen membership could address is li depletion .
li depletion in low - mass stars is predicted to take place the fastest for stars of mass @xmath30.6m@xmath33 @xcite , which corresponds to spectral type @xmath3m2 for @xmath310 - 20myr - old objects .
only a small sample of early - m - type stars are known in the ucl and lcc subgroups , primarily among the krautter - wichmann and park - finley _ rosat _ tts samples . despite the evidence for a short duration of star - formation in us ( e.g. * ? ? ?
* ) , @xcite have recently suggested that the spread in li abundances among the low - mass members of us may be interpreted as evidence for a significant age spread .
a high - resolution spectroscopic study of low - mass sco - cen members may be able to address the following questions : is there a demonstrable spread in li abundances for stars of a given mass ? does the spread correlate with any fundamental parameters ( i.e. rotation , binarity , etc . ) ?
is there evidence for a spread in ages amongst the low - mass members ?
larger samples of low - mass sco - cen members are sorely needed to address these topics .
if the imf predictions are any indication , then the surface of the census of low - mass sco - cen members has barely been scratched . due to its intermediate age , between that of well - studied t associations and nearby zams clusters , as well as its close proximity , the nearest ob association promises to be an important region for studying star - formation up close , and understanding of the pms evolution of low - mass stars and their circumstellar disks .
the sco - cen region is a promising hunting ground for future ground - based , high - cadence , photometric and astrometric survey facilities ( e.g. , pan - starrs , skymapper ) as well as the gaia astrometric mission . by combining the photometric and astrometric capabilities of these facilities with a focused spectroscopic survey ,
it is possible that we could construct an accurate 7-dimensional picture ( position , velocity , age ) of the complete stellar census of our nearest `` starburst '' within the next decade or two .
we dedicate this review to adriaan blaauw , whose ph.d . thesis 60 years ago , and series of influential papers over the past six decades , laid the foundation for much of our understanding of sco - cen , and the other ob associations .
we are grateful to hans zinnecker and ronnie hoogerwerf for valuable comments on the manuscript , and tim de zeeuw for providing a figure .
eric mamajek is supported by a clay postdoctoral fellowship from the smithsonian astrophysical observatory .
thomas preibisch would like to thank hans zinnecker for many years of motivation and advice in studying upper scorpius .
this work made extensive use of nasa s astrophysics data system bibliographic services and the simbad database ( cds , strasbourg , france ) .
153557.8@xmath164232405 & k3 : & rc & 10.88 & 0.86&1.4 & 3.649 & @xmath165&0.90 & 3.0 + 154106.7@xmath164265626 & g7 & rc & 9.98 & 0.77&1.6 & 3.701 & @xmath166&1.42 & 3.0 + 154413.4@xmath164252258 & m1 & rc & 11.33 & 1.13&0.6 & 3.564 & @xmath167&0.38 & 1.0 + 154920.9@xmath164260005 & k0 & rc & 9.71 & 0.64&0.7 & 3.676 & @xmath168&1.10 & 1.8 + 155106.6@xmath164240218 & m2 & rc & 12.08 & 1.19&0.4 & 3.551 & @xmath169&0.35 & 2.5 + 155231.2@xmath164263351 & m0 & rc & 10.88 & 0.75&0.0 & 3.576 & @xmath170&0.43 & 1.0 + 155459.9@xmath164234718 & g2 & rnc & 8.14 & 0.41&0.2 & 3.738 & @xmath171&1.82 & 2.8 + 155506.2@xmath164252109 & m1 & rnc & 10.57 & 0.86&0.0 & 3.564 & @xmath172&0.33 & 0.6 + 155517.1@xmath164232216 & m2.5 & & 12.17 & 1.11&0.0 & 3.545 & @xmath173&0.30 & 2.9 + 155548.7@xmath164251223 & g3 & rc & 9.44 & 0.42&0.2 & 3.731 & @xmath174&1.34 & 9.0 + 155702.3@xmath164195042 & k7 : & rc & 10.24 & 0.77&0.2 & 3.609 & @xmath175&0.50 & 0.8 + 155716.6@xmath164252918 & m0 & & 10.94 & 0.85&0.0 & 3.576 & @xmath176&0.44 & 1.2 + 155720.0@xmath164233849 & m0 & & 10.95 & 0.92&0.0 & 3.576 & @xmath177&0.44 & 1.2 + 155725.8@xmath164235422 & m0.5 & & 10.94 & 1.16&0.9 & 3.570 & @xmath178&0.34 & 0.5 + 155734.4@xmath164232111 & m1 & rc & 11.37 & 1.18&0.8 & 3.564 & @xmath179&0.38 & 0.9 + 155750.0@xmath164230508 & m0 & rc & 11.46 & 0.95&0.1 & 3.576 & @xmath180&0.50 & 2.9 + 155812.7@xmath164232835 & g2 & rnc & 9.25 & 0.47&0.4 & 3.738 & @xmath181&1.47 & 8.0 + 155847.8@xmath164175800 & k3 & rc & 10.46 & 0.94&1.8 & 3.649 & @xmath182&0.78 & 1.0 + 155902.1@xmath164184414 & k6 & rc & 10.40 & 0.94&1.3 & 3.620 & @xmath183&0.53 & 0.7 + 155950.0@xmath164255557 & m2 & rc & 11.89 & 1.31&1.0 & 3.551 & @xmath184&0.35 & 1.5 + 160000.0@xmath164222037 & m1 & & 11.08 & 1.13&0.6 & 3.564 & @xmath185&0.36 & 0.8 + 160000.7@xmath164250941 & g0 & rnc & 9.69 & 0.36&0.0 & 3.750 & @xmath186&1.10&18.0 + 160013.3@xmath164241810 & m0 & & 11.68 & 1.04&0.6 & 3.576 & @xmath187&0.50 & 3.0 + 160031.3@xmath164202705 & m1 & & 11.24 & 1.22&1.0 & 3.564 & @xmath188&0.35 & 0.7 + 160040.6@xmath164220032 & g9 & rc & 9.95 & 0.56&0.5 & 3.685 & @xmath183&1.22 & 4.0 + 160042.8@xmath164212737 & k7 & & 10.97 & 0.88&0.8 & 3.609 & @xmath189&0.60 & 1.8 + 160105.2@xmath164222731 & m3 & & 11.27 & 1.33&0.5 & 3.538 & @xmath190&0.27 & 0.5 + 160108.0@xmath164211318 & m0 & rc & 10.86 & 0.86&0.0 & 3.576 & @xmath191&0.42 & 1.0 + 160125.7@xmath164224040 & k1 & rc & 10.16 & 0.63&0.6 & 3.667 & @xmath192&1.05 & 3.0 + 160147.4@xmath164204945 & m0 & rnc & 10.92 & 1.08&0.8 & 3.576 & @xmath193&0.38 & 0.8 + 160151.4@xmath164244524 & k7 & rc & 10.54 & 0.84&0.6 & 3.609 & @xmath194&0.53 & 1.0 + 160158.2@xmath164200811 & g5 & rc & 9.29 & 0.58&0.8 & 3.717 & @xmath195&1.60 & 3.3 + 160200.3@xmath164222123 & m1 & rc & 11.23 & 1.23&1.1 & 3.564 & @xmath196&0.35 & 0.8 + 160208.5@xmath164225457 & m1 & & 11.91 & 1.13&0.6 & 3.564 & @xmath197&0.43 & 2.8 + 160210.4@xmath164224128 & k5 & rc & 9.89 & 0.71&0.4 & 3.630 & @xmath198&0.60 & 0.7 + 160239.1@xmath164254208 & k7 & rc & 10.80 & 0.73&0.0 & 3.609 & @xmath199&0.65 & 2.7 + 160251.2@xmath164240156 & k4 & rc & 10.82 & 0.71&0.6 & 3.639 & @xmath200&0.90 & 4.0 + 160253.9@xmath164202248 & k7 & rc & 10.58 & 0.96&1.1 & 3.609 & @xmath201&0.50 & 0.8 + 160302.7@xmath164180605 & k4 & rc & 10.78 & 0.83&1.1 & 3.639 & @xmath202&0.80 & 2.2 + 160323.7@xmath164175142 & m2 & rc & 11.06 & 1.25&0.7 & 3.551 & @xmath203&0.30 & 0.4 + 160354.9@xmath164203137 & m0 & rc & 11.10 & 1.10&0.9 & 3.576 & @xmath204&0.40 & 0.9 + 160357.6@xmath164203105 & k5 & rc & 10.92 & 0.81&0.9 & 3.630 & @xmath205&0.78 & 2.9 + 160420.9@xmath164213042 & m2 & rc & 12.13 & 1.57&2.2 & 3.551 & @xmath167&0.33 & 0.9 + 160421.7@xmath164213028 & k2 & rc & 10.64 & 0.73&1.0 & 3.658 & @xmath165&1.00 & 3.7 + 160447.7@xmath164193023 & k2 & rc & 9.82 & 0.65&0.6 & 3.658 & @xmath206&0.90 & 1.2 + 160527.3@xmath164193846 & m1 & & 11.99 & 1.07&0.3 & 3.564 & @xmath207&0.44 & 3.5 + 160539.1@xmath164215230 & m1 & & 11.93 & 1.18&0.8 & 3.564 & @xmath208&0.43 & 2.5 + 160542.7@xmath164200415 & m2 & rc & 11.76 & 1.34&1.1 & 3.551 & @xmath209&0.34 & 1.0 + 160550.5@xmath164253313 & g7 & rc & 9.99 & 0.47&0.2 & 3.701 & @xmath210&1.20 & 8.5 + 160612.5@xmath164203647 & k5 & rc & 11.27 & 1.00&1.8 & 3.630 & @xmath211&0.75 & 2.5 + 160621.9@xmath164192844 & m0.5&rc & 11.33 & 1.09&0.6 & 3.570 & @xmath212&0.41 & 1.1 + 160637.4@xmath164210840 & m1 & rc & 11.67 & 1.33&1.5 & 3.564 & @xmath213&0.37 & 0.9 + 160639.9@xmath164200128 & m3 & & 12.79 & 1.43&1.0 & 3.538 & @xmath214&0.27 & 2.8 + 160654.4@xmath164241610 & m3 & rc & 10.87 & 1.08&0.0 & 3.538 & @xmath215&1.05 & 3.0 + 160703.5@xmath164203626 & m0 & rc & 10.22 & 0.90&0.0 & 3.576 & @xmath216&0.33 & 0.5 + 160703.6@xmath164204308 & m1 & & 12.15 & 1.22&1.0 & 3.564 & @xmath217&0.43 & 2.8 + 160703.9@xmath164191132 & m1 & rc & 11.81 & 1.23&1.1 & 3.564 & @xmath218&0.40 & 1.7 + 160814.7@xmath164190833 & k2 & rc & 10.28 & 0.87&1.6 & 3.658 & @xmath219&0.87 & 1.2 + 160831.4@xmath164180241 & m0 & rc & 11.26 & 0.93&0.1 & 3.576 & @xmath220&0.47 & 2.0 + 160843.4@xmath164260216 & g7 & rc & 9.30 & 0.71&1.3 & 3.701 & @xmath221&1.50 & 1.8 + 160856.7@xmath164203346 & k5 & rc & 10.86 & 0.92&1.4 & 3.630 & @xmath222&0.70 & 1.8 + 160930.3@xmath164210459 & m0 & rc & 11.14 & 0.81&0.0 & 3.576 & @xmath223&0.47 & 2.0 + 160941.0@xmath164221759 & m0 & rc & 11.06 & 0.84&0.0 & 3.576 & @xmath224&0.45 & 1.8 + 161019.1@xmath164250230 & m1 & rc & 9.76 & 1.01&0.0 & 3.564 & @xmath225&0.28 & 0.1 + 161021.7@xmath164190406 & m1 & rnc & 12.26 & 1.45&2.1 & 3.564 & @xmath226&0.40 & 1.8 + 161028.5@xmath164190446 & m3 & & 11.42 & 1.27&0.2 & 3.538 & @xmath227&0.28 & 0.9 + 161042.0@xmath164210132 & k5 & rc & 10.74 & 0.94&1.5 & 3.630 & @xmath228&0.66 & 1.1 + 161108.9@xmath164190446 & k2 & rc & 10.17 & 0.89&1.7 & 3.658 & @xmath229&0.84 & 1.0 + 161120.6@xmath164182054 & k5 & rnc & 10.73 & 0.83&1.0 & 3.630 & @xmath230&0.70 & 1.8 + 161156.3@xmath164230404 & m1 & rc & 10.91 & 1.18&0.8 & 3.564 & @xmath231&0.32 & 0.5 + 161159.2@xmath164190652 & k0 & rc & 10.18 & 0.75&1.3 & 3.676 & @xmath232&1.12 & 2.5 + 161220.9@xmath164190903 & m2.5 & & 12.68 & 1.36&0.9 & 3.545 & @xmath233&0.30 & 3.0 + 161240.5@xmath164185927 & k0 & rc & 9.38 & 0.72&1.1 & 3.676 & @xmath234&1.01 & 0.8 + 161302.7@xmath164225744 & k4 & rc & 10.38 & 1.08&2.3 & 3.639 & @xmath235&0.64 & 0.5 + 161318.6@xmath164221248 & g9 & rc & 9.25 & 0.66&0.9 & 3.685 & @xmath236&1.18 & 1.0 + 161329.3@xmath164231106 & k1 & rc & 10.29 & 0.92&2.0 & 3.667 & @xmath237&0.95 & 1.2 + 161402.1@xmath164230101 & g4 & rc & 10.06 & 0.81&2.0 & 3.724 & @xmath238&1.50 & 5.0 + 161411.0@xmath164230536 & k0 & rc & 9.19 & 0.99&2.4 & 3.676 & @xmath239&1.02 & 0.3 + 161459.2@xmath164275023 & g5 & rc & 10.11 & 0.61&1.0 & 3.717 & @xmath198&1.28 & 9.0 + 161534.6@xmath164224241 & m1 & rc & 10.38 & 1.06&0.3 & 3.564 & @xmath240&0.29 & 0.3 + 161618.0@xmath164233947 & g7 & rc & 9.55 & 0.67&1.1 & 3.701 & @xmath241&1.47 & 2.2 + 161731.4@xmath164230334 & g0 & rc & 9.14 & 0.47&0.5 & 3.750 & @xmath242&1.43 & 9.0 + 161933.9@xmath164222828 & k0 & rc & 10.09 & 0.75&1.3 & 3.676 & @xmath243&1.10 & 2.0 + 162046.0@xmath164234820 & k3 & rc & 10.70 & 1.03&2.2 & 3.649 & @xmath244&0.78 & 1.0 + 162307.8@xmath164230059 & k2 & rc & 10.18 & 0.69&0.8 & 3.658 & @xmath245&0.93 & 2.0 + 162948.6@xmath164215211 & k0 & rc & 9.77 & 0.71&1.1 & 3.676 & @xmath246&1.08 & 1.8 + [ pz99tab ] 155532.4 - 230817 & 16.8 & 0.15 & 0.5 & m1 & 5.7 & 3.569 & -0.75 + 155624.8 - 222555 & 15.5 & 0.78 & -5.4 & m4 & 1.7 & 3.516 & -1.12 + 155625.7 - 224027 & 14.6 & 0.53 & -4.2 & m3 & 1.8 & 3.532 & -0.85 + 155629.5 - 225657 & 14.3 & 0.54 & -3.1 & m3 & 0.9 & 3.533 & -0.98 + 155655.5 - 225839 & 13.1 & 0.54 & -1.9 & m0 & 0.7 & 3.578 & -0.69 + 155706.4 - 220606 & 16.2 & 0.52 & -3.6 & m4 & 2.0 & 3.513 & -1.31 + 155728.5 - 221904 & 16.7 & 0.40 & -6.2 & m5 & 2.3 & 3.505 & -1.37 + 155729.2 - 221523 & 17.1 & 0.29 & -4.0 & m5 & 2.1 & 3.503 & -1.58 + 155729.9 - 225843 & 15.8 & 0.55 & -7.0 & m4 & 1.4 & 3.511 & -1.30 + 155737.2 - 224524 & 16.7 & 0.15 & 0.0 & m2 & 1.6 & 3.544 & -1.76 + 155742.5 - 222605 & 15.7 & 0.15 & -2.9 & m3 & 1.8 & 3.526 & -1.22 + 155744.9 - 222351 & 14.3 & 0.13 & 0.9 & m2 & 0.7 & 3.558 & -1.11 + 155746.6 - 222919 & 14.7 & 0.69 & -2.1 & m3 & 1.3 & 3.533 & -1.01 + 155829.8 - 231007 & 15.4 & 0.41 & -250 & m3 & 0.0 & 3.532 & -1.63 + 155848.6 - 224657 & 14.5 & 0.19 & -0.5 & m0 & 0.7 & 3.579 & -1.24 + 155912.5 - 223650 & 16.4 & 0.89 & -10.5 & m5 & 0.3 & 3.493 & -1.70 + 155918.4 - 221042 & 14.7 & 0.98 & -1.0 & m4 & 1.3 & 3.515 & -0.90 + 155925.9 - 230508 & 17.0 & 0.65 & -19.5 & m6 & 1.5 & 3.485 & -1.57 + 155930.1 - 225125 & 17.6 & 0.70 & -5.0 & m4 & 2.5 & 3.511 & -1.71 + 160004.3 - 223014 & 14.9 & 0.20 & -0.4 & m3 & 0.2 & 3.528 & -1.40 + 160007.2 - 222406 & 16.8 & 0.27 & -4.4 & m4 & 0.6 & 3.509 & -1.91 + 160017.4 - 221810 & 16.5 & 0.50 & -3.9 & m6 & 0.0 & 3.481 & -1.79 + 160018.4 - 223011 & 14.7 & 0.30 & -150 & m3 & 1.5 & 3.539 & -0.97 + 160026.3 - 225941 & 16.8 & 1.00 & -10.0 & m5 & 2.5 & 3.495 & -1.25 + 160028.5 - 220922 & 16.8 & 0.65 & 0.0 & m6 & 0.0 & 3.476 & -1.93 + 160030.2 - 233445 & 17.2 & 0.50 & -15.8 & m6 & 1.0 & 3.476 & -1.71 + 160054.5 - 224908 & 14.7 & 0.10 & 0.0 & m3 & 1.0 & 3.530 & -1.09 + 160106.0 - 221524 & 15.9 & 0.80 & -8.0 & m5 & 1.0 & 3.501 & -1.38 + 160110.4 - 222227 & 14.2 & 0.80 & -9.4 & m4 & 1.0 & 3.513 & -0.78 + 160121.5 - 223726 & 14.2 & 0.80 & -8.4 & m4 & 0.4 & 3.507 & -0.92 + 160129.8 - 224838 & 14.9 & 0.80 & -4.4 & m4 & 1.6 & 3.515 & -0.90 + 160132.9 - 224231 & 14.6 & 0.20 & -1.4 & m0 & 0.9 & 3.580 & -1.22 + 160140.8 - 225810 & 14.0 & 0.45 & -120 & m3 & 0.0 & 3.528 & -1.16 + 160142.6 - 222923 & 13.5 & 0.33 & -0.6 & m0 & 0.7 & 3.580 & -0.87 + 160158.9 - 224036 & 14.4 & 0.80 & -7.6 & m4 & 0.7 & 3.515 & -0.97 + 160159.7 - 195219 & 16.3 & 0.47 & -2.7 & m5 & 2.0 & 3.497 & -1.22 + 160202.9 - 223613 & 14.7 & 0.18 & -1.5 & m0 & 1.6 & 3.575 & -1.08 + 160207.5 - 225746 & 13.7 & 0.30 & -3.2 & m1 & 0.9 & 3.565 & -0.81 + 160210.9 - 200749 & 16.4 & 0.65 & -3.5 & m5 & 1.6 & 3.501 & -1.40 + 160222.4 - 195653 & 14.7 & 0.40 & -6.6 & m3 & 1.6 & 3.533 & -0.95 + 160226.2 - 200241 & 15.2 & 0.47 & -4.9 & m5 & 0.0 & 3.495 & -1.39 + 160236.2 - 191732 & 17.3 & 0.40 & 0.7 & m3 & 2.5 & 3.539 & -1.74 + 160245.4 - 194604 & 15.9 & 0.30 & 0.7 & m2 & 1.4 & 3.550 & -1.54 + 160245.4 - 193037 & 16.4 & 0.51 & -1.1 & m5 & 1.8 & 3.495 & -1.31 + 160245.7 - 230450 & 16.7 & 1.00 & -100 & m6 & 1.5 & 3.485 & -1.46 + 160258.5 - 225649 & 13.9 & 1.00 & -9.4 & m2 & 0.8 & 3.546 & -0.88 + 160325.6 - 194438 & 15.4 & 0.20 & -1.4 & m2 & 1.6 & 3.544 & -1.22 + 160329.4 - 195503 & 15.7 & 0.40 & -4.9 & m5 & 1.0 & 3.503 & -1.30 + 160341.8 - 200557 & 13.5 & 0.30 & -2.2 & m2 & 0.9 & 3.555 & -0.70 + 160343.3 - 201531 & 13.6 & 0.49 & -3.5 & m2 & 0.9 & 3.558 & -0.76 + 160350.4 - 194121 & 15.0 & 0.59 & -4.9 & m5 & 0.6 & 3.505 & -1.17 + 160357.9 - 194210 & 14.5 & 0.35 & -3.0 & m2 & 1.7 & 3.546 & -0.87 + 160407.7 - 194857 & 16.8 & 0.22 & -4.0 & m5 & 0.7 & 3.499 & -1.80 + 160418.2 - 191055 & 15.4 & 0.76 & -0.5 & m4 & 2.3 & 3.524 & -0.99 + 160428.0 - 190434 & 15.8 & 0.55 & -3.4 & m4 & 2.4 & 3.516 & -1.05 + 160428.4 - 190441 & 13.5 & 0.68 & -5.0 & m3 & 1.0 & 3.532 & -0.63 + 160435.6 - 194830 & 16.1 & 0.63 & -7.2 & m5 & 1.6 & 3.499 & -1.29 + 160439.1 - 194245 & 15.1 & 0.61 & -3.8 & m4 & 0.9 & 3.518 & -1.18 + 160449.9 - 203835 & 16.0 & 0.65 & -7.8 & m5 & 1.0 & 3.497 & -1.37 + 160456.4 - 194045 & 15.0 & 0.54 & -5.2 & m4 & 0.9 & 3.507 & -1.10 + 160502.1 - 203507 & 13.6 & 0.57 & -5.9 & m2 & 1.8 & 3.551 & -0.53 + 160508.3 - 201531 & 13.3 & 0.58 & -4.3 & m4 & 0.0 & 3.524 & -0.80 + 160508.5 - 201532 & 14.0 & 0.46 & -5.5 & m4 & 0.4 & 3.518 & -0.90 + 160516.1 - 193830 & 15.6 & 0.50 & -2.4 & m4 & 1.1 & 3.520 & -1.35 + 160517.9 - 202420 & 12.9 & 0.41 & -5.2 & m3 & 0.6 & 3.542 & -0.56 + 160521.9 - 193602 & 13.9 & 0.50 & -1.9 & m1 & 1.2 & 3.563 & -0.83 + 160522.7 - 205111 & 14.8 & 0.10 & -4.5 & m4 & 1.1 & 3.518 & -1.03 + 160525.5 - 203539 & 16.1 & 0.48 & -6.1 & m5 & 1.5 & 3.499 & -1.30 + 160528.5 - 201037 & 14.1 & 0.55 & -1.9 & m1 & 1.4 & 3.569 & -0.85 + 160531.3 - 192623 & 16.7 & 0.58 & -8.8 & m5 & 1.8 & 3.497 & -1.48 + 160532.1 - 193315 & 16.5 & 0.61 & -26.0 & m5 & 0.0 & 3.499 & -1.88 + 160545.4 - 202308 & 14.4 & 0.31 & -35.0 & m2 & 1.4 & 3.560 & -0.98 + 160600.6 - 195711 & 14.9 & 0.80 & -7.5 & m5 & 1.7 & 3.505 & -0.82 + 160611.9 - 193532 & 16.0 & 0.63 & -8.2 & m5 & 0.9 & 3.495 & -1.39 + 160619.3 - 192332 & 16.4 & 0.52 & -5.5 & m5 & 1.9 & 3.505 & -1.37 + 160622.8 - 201124 & 15.4 & 0.53 & -6.0 & m5 & 0.0 & 3.499 & -1.47 + 160628.7 - 200357 & 15.0 & 0.58 & -30.0 & m5 & 0.6 & 3.499 & -1.12 + 160629.0 - 205216 & 15.1 & 0.68 & -6.2 & m5 & 1.2 & 3.497 & -0.99 + 160632.1 - 202053 & 17.3 & 0.69 & -5.1 & m5 & 1.5 & 3.495 & -1.78 + 160643.8 - 190805 & 12.7 & 0.65 & -3.8 & k6 & 1.9 & 3.623 & -0.31 + 160647.5 - 202232 & 13.9 & 0.41 & -3.2 & m2 & 1.6 & 3.557 & -0.70 + 160700.1 - 203309 & 14.1 & 0.65 & -0.6 & m2 & 1.8 & 3.551 & -0.69 + 160702.1 - 201938 & 16.3 & 0.39 & -30.0 & m5 & 1.7 & 3.501 & -1.37 + 160704.7 - 201555 & 16.2 & 0.61 & -4.2 & m4 & 1.7 & 3.511 & -1.39 + 160707.7 - 192715 & 14.0 & 0.57 & -5.0 & m2 & 2.2 & 3.548 & -0.52 + 160708.7 - 192733 & 15.8 & 0.52 & -4.0 & m4 & 1.7 & 3.516 & -1.26 + 160710.0 - 191703 & 16.6 & 0.90 & -9.0 & m2 & 3.6 & 3.551 & -1.23 + 160716.0 - 204443 & 14.7 & 0.68 & -3.6 & m4 & 1.2 & 3.524 & -0.97 + 160719.7 - 202055 & 15.3 & 0.57 & -2.3 & m3 & 2.7 & 3.542 & -0.89 + 160722.4 - 201158 & 17.2 & 0.50 & -14.0 & m5 & 2.3 & 3.491 & -1.47 + 160726.8 - 185521 & 14.1 & 0.33 & -1.9 & m1 & 1.1 & 3.569 & -0.97 + 160727.5 - 201834 & 17.3 & 0.90 & -13.0 & m5 & 2.4 & 3.487 & -1.45 + 160735.5 - 202713 & 17.2 & 0.75 & -4.9 & m5 & 2.3 & 3.505 & -1.57 + 160739.4 - 191747 & 14.0 & 0.52 & -2.3 & m2 & 1.6 & 3.560 & -0.72 + 160744.5 - 203602 & 13.6 & 0.50 & -4.8 & m4 & 0.8 & 3.516 & -0.61 + 160745.8 - 203055 & 17.3 & 0.30 & -2.0 & m3 & 1.9 & 3.530 & -1.87 + 160800.5 - 204028 & 16.3 & 0.34 & -5.2 & m5 & 1.9 & 3.501 & -1.31 + 160801.4 - 202741 & 12.8 & 0.83 & -2.3 & k8 & 1.5 & 3.601 & -0.45 + 160801.5 - 192757 & 14.0 & 0.52 & -3.6 & m4 & 1.0 & 3.524 & -0.76 + 160802.4 - 202233 & 15.3 & 0.60 & -6.1 & m5 & 0.5 & 3.499 & -1.27 + 160803.6 - 181237 & 17.3 & 0.35 & -7.7 & m4 & 3.9 & 3.515 & -1.23 + 160804.3 - 194712 & 14.0 & 0.91 & -5.2 & m4 & 0.0 & 3.516 & -1.17 + 160815.3 - 203811 & 15.1 & 0.57 & -1.9 & m3 & 2.1 & 3.542 & -0.98 + 160818.4 - 190059 & 14.1 & 0.72 & -3.7 & m3 & 0.3 & 3.528 & -1.03 + 160822.4 - 193004 & 12.9 & 0.29 & -3.0 & m1 & 1.2 & 3.571 & -0.43 + 160823.2 - 193001 & 13.1 & 0.49 & -6.0 & k9 & 1.5 & 3.586 & -0.49 + 160823.5 - 191131 & 14.1 & 0.59 & -4.1 & m2 & 1.5 & 3.544 & -0.79 + 160823.8 - 193551 & 13.2 & 0.72 & -2.1 & m1 & 1.5 & 3.569 & -0.47 + 160825.1 - 201224 & 13.7 & 0.41 & -2.0 & m1 & 1.3 & 3.574 & -0.77 + 160827.5 - 194904 & 15.6 & 0.57 & -12.3 & m5 & 1.4 & 3.495 & -1.11 + 160841.7 - 185610 & 16.9 & 0.70 & -11.0 & m6 & 1.1 & 3.483 & -1.61 + 160843.1 - 190051 & 14.9 & 0.57 & -5.8 & m4 & 1.3 & 3.515 & -1.00 + 160845.6 - 182443 & 13.7 & 0.70 & 0.5 & m3 & 0.2 & 3.533 & -0.93 + 160854.0 - 203417 & 14.8 & 0.57 & -3.4 & m4 & 1.5 & 3.518 & -0.90 + 160900.0 - 190836 & 15.9 & 0.62 & -15.4 & m5 & 0.7 & 3.503 & -1.45 + 160900.7 - 190852 & 12.8 & 0.53 & -12.7 & k9 & 0.8 & 3.592 & -0.60 + 160903.9 - 193944 & 14.6 & 0.68 & -7.2 & m4 & 0.6 & 3.520 & -1.10 + 160904.0 - 193359 & 15.5 & 0.80 & -2.2 & m4 & 2.5 & 3.524 & -0.96 + 160908.4 - 200928 & 13.4 & 0.66 & -4.0 & m4 & 0.7 & 3.518 & -0.60 + 160913.4 - 194328 & 14.9 & 0.54 & -1.6 & m3 & 1.8 & 3.537 & -0.97 + 160915.8 - 193706 & 16.1 & 0.78 & -4.4 & m5 & 0.7 & 3.499 & -1.53 + 160916.8 - 183522 & 13.8 & 0.55 & -3.0 & m2 & 1.0 & 3.546 & -0.80 + 160926.7 - 192502 & 14.9 & 0.40 & 0.0 & m3 & 1.0 & 3.528 & -1.16 + 160933.8 - 190456 & 14.1 & 0.57 & -3.5 & m2 & 1.4 & 3.557 & -0.83 + 160935.6 - 182822 & 15.5 & 0.75 & -4.2 & m3 & 1.8 & 3.528 & -1.17 + 160936.5 - 184800 & 15.0 & 0.57 & -18.0 & m3 & 1.2 & 3.532 & -1.14 + 160943.8 - 182302 & 15.0 & 0.48 & -5.8 & m4 & 1.9 & 3.516 & -0.86 + 160946.4 - 193735 & 13.5 & 0.49 & -1.6 & m1 & 1.6 & 3.569 & -0.57 + 160953.6 - 175446 & 16.6 & 1.30 & -22.0 & m3 & 4.1 & 3.539 & -1.03 + 160954.4 - 190654 & 13.5 & 0.70 & -3.1 & m1 & 0.9 & 3.569 & -0.78 + 160959.4 - 180009 & 14.7 & 0.58 & -4.0 & m4 & 0.2 & 3.518 & -1.26 + 161007.5 - 181056 & 17.8 & 0.44 & -13.3 & m6 & 1.9 & 3.474 & -1.70 + 161010.4 - 194539 & 14.7 & 0.49 & -5.6 & m3 & 1.4 & 3.533 & -1.01 + 161011.0 - 194603 & 16.2 & 0.70 & -4.4 & m5 & 0.3 & 3.487 & -1.58 + 161014.7 - 191909 & 14.4 & 0.62 & -2.3 & m3 & 1.0 & 3.535 & -0.97 + 161021.5 - 194132 & 14.0 & 0.40 & -4.3 & m3 & 2.0 & 3.541 & -0.59 + 161024.7 - 191407 & 14.9 & 0.55 & -3.7 & m3 & 1.5 & 3.528 & -0.99 + 161026.4 - 193950 & 15.0 & 0.62 & -4.4 & m4 & 1.9 & 3.516 & -0.87 + 161028.1 - 191043 & 17.3 & 0.65 & -11.4 & m4 & 2.1 & 3.511 & -1.72 + 161030.0 - 183906 & 16.1 & 0.54 & -6.5 & m4 & 2.0 & 3.507 & -1.23 + 161030.9 - 182422 & 15.1 & 0.55 & -3.4 & m3 & 1.8 & 3.535 & -1.04 + 161031.9 - 191305 & 12.6 & 0.50 & -2.3 & k7 & 1.1 & 3.607 & -0.48 + 161039.5 - 191652 & 14.5 & 0.53 & -4.3 & m2 & 1.5 & 3.551 & -0.96 + 161043.9 - 192225 & 14.0 & 0.67 & -2.3 & m3 & 1.1 & 3.539 & -0.83 + 161046.3 - 184059 & 16.8 & 0.51 & -7.2 & m4 & 3.1 & 3.516 & -1.28 + 161052.4 - 193734 & 15.4 & 0.93 & -3.9 & m3 & 2.3 & 3.526 & -1.00 + 161110.9 - 193331 & 16.2 & 0.63 & -6.3 & m5 & 1.1 & 3.501 & -1.46 + 161112.3 - 192737 & 17.3 & 0.30 & -50.0 & m5 & 1.4 & 3.501 & -1.83 + 161115.3 - 175721 & 13.1 & 0.55 & -2.4 & m1 & 1.6 & 3.574 & -0.42 + 161116.6 - 193910 & 13.9 & 0.47 & -3.4 & m4 & 0.6 & 3.513 & -0.77 + 161118.1 - 175728 & 13.9 & 0.55 & -4.8 & m4 & 0.9 & 3.515 & -0.69 + 161118.2 - 180358 & 17.5 & 0.90 & -20.0 & m6 & 1.6 & 3.476 & -1.65 + 161120.4 - 191937 & 14.1 & 0.51 & -5.5 & m2 & 0.6 & 3.544 & -1.00 + 161123.0 - 190522 & 14.4 & 0.40 & -6.9 & m3 & 1.7 & 3.528 & -0.74 + 161129.4 - 194224 & 16.4 & 0.80 & -13.0 & m6 & 0.8 & 3.483 & -1.49 + 161133.6 - 191400 & 14.4 & 0.56 & -3.7 & m3 & 1.8 & 3.537 & -0.79 + 161146.1 - 190742 & 16.2 & 0.47 & -6.3 & m5 & 1.6 & 3.503 & -1.37 + 161156.2 - 194323 & 14.1 & 0.60 & -6.3 & m3 & 0.7 & 3.542 & -0.97 + 161247.2 - 190353 & 17.4 & 0.88 & -10.6 & m6 & 1.3 & 3.485 & -1.78 + 161248.9 - 180052 & 14.6 & 0.52 & -3.8 & m3 & 1.4 & 3.530 & -0.92 + 161328.0 - 192452 & 17.8 & 0.40 & -17.0 & m5 & 2.2 & 3.491 & -1.73 + 161347.5 - 183459 & 14.6 & 0.52 & -3.5 & m2 & 1.5 & 3.544 & -0.98 + 161358.1 - 184828 & 13.7 & 0.57 & -1.9 & m2 & 1.0 & 3.553 & -0.76 + 161420.2 - 190648 & 13.2 & 0.37 & -52.0 & k5 & 1.8 & 3.630 & -0.59 + 161433.6 - 190013 & 14.2 & 0.64 & -26.0 & m2 & 2.2 & 3.551 & -0.60 + 161437.5 - 185823 & 15.7 & 0.13 & -3.6 & m3 & 1.4 & 3.533 & -1.42 + [ 2dfpmstab ] + name & 2mass j & spt & @xmath251 & a@xmath252 & dist & log@xmath253 & log@xmath254 & log@xmath255 & log@xmath256 & age & @xmath257 & x - ray + & & & [ mag ] & [ mag ] & [ pc ] & [ erg / s ] & & [ k ] & & [ myr ] & [ @xmath146 ] & counterpart + + name & 2mass j & spt & @xmath251 & a@xmath252 & dist & log@xmath253 & log@xmath254 & log@xmath255 & log@xmath256 & age & @xmath257 & x - ray + & & & [ mag ] & [ mag ] & [ pc ] & [ erg / s ] & & [ k ] & & [ myr ] & [ @xmath146 ] & counterpart + mml 36 & 13375730 - 4134419 & k0iv & 10.08 & -0.07 & 98 & 30.4 & -3.2 & 3.72 & 0.00 & 19 & 1.1 & 1rxs j133758.0 - 413448 + mml 38 & 13475054 - 4902056 & g8ive & 10.82 & 0.45 & 148 & 30.5 & -3.2 & 3.74 & 0.05 & 22 & 1.1 & 1rxs j134748.0 - 490158 + hip 67522 & 13500627 - 4050090 & g0.5iv & 9.79 & 0.02 & 126 & 30.4 & -3.5 & 3.77 & 0.25 & 21 & 1.2 & 1rxs j135005.7 - 405001 + mml 39 & 13524780 - 4644092 & g0iv & 9.62 & 0.34 & 145 & 30.8 & -3.3 & 3.78 & 0.44 & 16 & 1.3 & 1rxs j135247.0 - 464412 + mml 40 & 14022072 - 4144509 & g9iv & 10.71 & 0.90 & 130 & 30.3 & -3.3 & 3.73 & 0.07 & 18 & 1.1 & 1rxs j140220.9 - 414435 + mml 41 & 14090357 - 4438442 & f9iv & 9.39 & 0.38 & 142 & 30.7 & -3.5 & 3.78 & 0.50 & 15 & 1.3 & 1rxs j140902.6 - 443838 + hip 70350 & 14233787 - 4357426 & f7v & 8.13 & 0.12 & 107 & 30.8 & -3.6 & 3.81 & 0.64 & 15 & 1.4 & 1rxs j142338.0 - 435814 + hip 70376 & 14235639 - 5029585 & f7v & 9.20 & 0.27 & 122 & 30.4 & -3.6 & 3.81 & 0.39 & ... & ... & 1rxs j142356.2 - 503006 + mml 43 & 14270556 - 4714217 & g7iv & 10.59 & 0.17 & 132 & 30.4 & -3.3 & 3.75 & 0.06 & 23 & 1.1 & 1rxs j142705.3 - 471420 + hip 70689 & 14273044 - 5231304 & f2v & 8.53 & 0.00 & 105 & 29.6 & -4.4 & 3.85 & 0.40 & ... & ... & 1rxs j142729.5 - 523141 + mml 44 & 14280929 - 4414175 & g5.5iv & 9.78 & 0.33 & 161 & 30.6 & -3.5 & 3.75 & 0.54 & 9 & 1.4 & 1rxs j142809.6 - 441438 + mml 45 & 14281937 - 4219341 & g3.5iv & 10.47 & 0.37 & 159 & 30.8 & -3.0 & 3.76 & 0.33 & 16 & 1.3 & 1rxs j142817.6 - 421958 + hip 70919 & 14301035 - 4332490 & g8iii & 8.88 & 0.00 & 193 & 30.8 & -3.9 & 3.68 & 0.84 & @xmath1031 & 2.2 & 1rxs j143008.7 - 433313 + hip 71023 & 14313339 - 4445019 & f0v & 8.94 & 0.15 & 160 & 30.6 & -3.9 & 3.86 & 0.65 & ... & ... & 1rxs j143135.2 - 444526 + hip 71178 & 14332578 - 3432376 & g8ive & 10.18 & 0.35 & 115 & @xmath10329.8 & @xmath103 - 3.9 & 3.74 & 0.16 & 16 & 1.2 & ... + mml 46 & 14370422 - 4145028 & g0.5iv & 9.70 & 0.18 & 156 & 30.7 & -3.4 & 3.78 & 0.59 & 11 & 1.4 & 1rxs j143704.6 - 414504 + mml 47 & 14375022 - 5457411 & k0+iv & 10.72 & 0.30 & 132 & 30.6 & -3.0 & 3.72 & 0.08 & 13 & 1.2 & 1rxs j143750.9 - 545708 + hip 71767 & 14404593 - 4247063 & f3v & 9.02 & 0.20 & 168 & 30.6 & -3.9 & 3.84 & 0.70 & 15 & 1.5 & 1rxs j144044.6 - 424720 + mml 48 & 14413499 - 4700288 & g4iv & 10.00 & 0.02 & 110 & 30.9 & -2.8 & 3.76 & 0.20 & 20 & 1.2 & 1rxs j144135.3 - 470039 + hip 72033 & 14440435 - 4059223 & f7iv / v & 9.17 & 0.23 & 156 & 30.3 & -3.9 & 3.81 & 0.60 & 15 & 1.4 & 1rxs j144405.2 - 405940 + hd 129791b & 14455620 - 4452346 & k5ve & 12.93 & 0.26 & 132 & 30.6 & -2.9 & 3.63 & -0.52 & 18 & 0.9 & 1rxs j144556.0 - 445202 + mml 49 & 14473176 - 4800056 & g2.5iv & 10.72 & 0.43 & 130 & 30.2 & -3.4 & 3.77 & -0.05 & ... & ... & 1rxs j144732.2 - 480019 + mml 50 & 14502581 - 3506486 & k0iv & 10.73 & 0.21 & 190 & 30.6 & -3.3 & 3.72 & 0.52 & 5 & 1.6 & 1rxs j145025.4 - 350645 + mml 51 & 14524198 - 4141552 & k1ive & 10.89 & 0.61 & 145 & 30.5 & -3.2 & 3.70 & 0.10 & 9 & 1.3 & 1rxs j145240.7 - 414206 + mml 52 & 14571962 - 3612274 & g6iv & 10.27 & 0.15 & 132 & 30.3 & -3.4 & 3.75 & 0.18 & 18 & 1.1 & 1rxs j145720.4 - 361242 + mml 53 & 14583769 - 3540302 & k2-ive & 10.75 & 0.23 & 136 & 30.3 & -3.4 & 3.69 & 0.21 & 6 & 1.4 & 1rxs j145837.6 - 354036 + mml 54 & 14592275 - 4013120 & g3iv & 9.71 & 0.40 & 122 & 30.5 & -3.4 & 3.76 & 0.32 & 17 & 1.2 & 1rxs j145923.0 - 401319 + mml 55 & 15005189 - 4331212 & g9iv & 11.15 & 0.30 & 163 & 30.4 & -3.2 & 3.73 & 0.16 & 14 & 1.2 & 1rxs j150052.5 - 433107 + mml 56 & 15011155 - 4120406 & g0.5iv & 10.01 & 0.48 & 214 & 31.0 & -3.3 & 3.77 & 0.64 & 10 & 1.4 & 1rxs j150112.0 - 412040 + mml 57 & 15015882 - 4755464 & g1.5iv & 10.15 & 0.22 & 156 & 30.2 & -3.8 & 3.77 & 0.29 & 19 & 1.2 & 1rxs j150158.5 - 475559 + mml 58 & 15071481 - 3504595 & g9.5iv & 10.49 & 0.44 & 101 & 30.0 & -3.4 & 3.73 & -0.12 & 28 & 1.0 & 1rxs j150714.5 - 350500 + mml 59 & 15083773 - 4423170 & g1.5ive & 10.83 & 0.13 & 180 & 30.9 & -2.9 & 3.77 & 0.29 & 19 & 1.2 & 1rxs j150836.0 - 442325 + mml 60 & 15083849 - 4400519 & g1.5iv & 10.54 & 0.11 & 134 & 30.4 & -3.3 & 3.77 & 0.20 & 23 & 1.1 & 1rxs j150838.5 - 440048 + mml 61 & 15125018 - 4508044 & g5iv & 10.71 & 0.27 & 151 & 30.6 & -3.1 & 3.76 & 0.19 & 19 & 1.2 & 1rxs j151250.0 - 450822 + hip 74501 & 15132923 - 5543545 & g2iv & 7.47 & 0.00 & 153 & 30.0 & -5.2 & 3.74 & 1.19 & 1 & 2.4 & 1rxs j151330.3 - 554341 + mml 62 & 15180174 - 5317287 & g7iv & 10.12 & 0.77 & 103 & 30.4 & -3.3 & 3.75 & 0.08 & 22 & 1.1 & 1rxs j151802.0 - 531719 + mml 63 & 15182692 - 3738021 & g9iv & 11.02 & 0.72 & 127 & 30.4 & -3.0 & 3.73 & 0.05 & 18 & 1.1 & 1rxs j151827.3 - 373808 + mml 64 & 15255964 - 4501157 & g8iv & 10.90 & 0.19 & 149 & 30.1 & -3.5 & 3.74 & -0.02 & 26 & 1.0 & 1rxs j152600.9 - 450113 + mml 65 * & 15293858 - 3546513 & k0+iv & 10.43 & 0.16 & 135 & 30.3 & -3.4 & 3.72 & 0.17 & 10 & 1.3 & 1rxs j152937.7 - 354656 + hip 75924 & 15302626 - 3218122 & g2.5iv & 8.80 & 0.37 & 102 & 30.9 & -2.9 & 3.77 & 0.69 & 5 & 1.7 & 1rxs j153026.1 - 321815 + hip 76084 & 15322013 - 3108337 & f2v & 8.62 & 0.16 & 149 & 29.8 & -4.6 & 3.85 & 0.73 & 15 & 1.5 & 1rxs j153218.2 - 310828 + mml 66 & 15370214 - 3136398 & g6iv & 9.99 & 0.43 & 135 & 30.9 & -3.0 & 3.75 & 0.42 & 11 & 1.3 & 1rxs j153701.9 - 313647 + hip 76472 & 15370466 - 4009221 & g1iv & 9.39 & 0.29 & 138 & 30.8 & -3.3 & 3.77 & 0.60 & 11 & 1.4 & 1rxs j153706.0 - 400929 + mml 67 & 15371129 - 4015566 & g8.5ive & 10.43 & 0.40 & 164 & 30.7 & -3.2 & 3.74 & 0.25 & 11 & 1.3 & 1rxs j153711.6 - 401608 + mml 68 & 15384306 - 4411474 & g8.5iv & 10.28 & 0.62 & 124 & 30.3 & -3.4 & 3.74 & 0.08 & 19 & 1.1 & 1rxs j153843.1 - 441149 + mml 69 & 15392440 - 2710218 & g5iv & 9.57 & 0.28 & 127 & 30.7 & -3.3 & 3.76 & 0.46 & 11 & 1.3 & 1rxs j153924.0 - 271035 + mml 70 * & 15440376 - 3311110 & k0ive & 10.87 & 0.47 & 124 & 30.2 & -3.3 & 3.72 & 0.01 & 18 & 1.1 & 1rxs j154404.1 - 331120 + hip 77081 * & 15442105 - 3318549 & g7.5iv & 9.69 & 0.15 & 131 & @xmath10329.9 & @xmath103 - 4.1 & 3.74 & 0.35 & 11 & 1.3 & ... + hip 77135 * & 15445769 - 3411535 & g4iv & 9.88 & 0.08 & 139 & 30.5 & -3.4 & 3.76 & 0.42 & 12 & 1.3 & 1rxs j154458.0 - 341143 + hip 77144 & 15450184 - 4050310 & g0iv & 9.46 & 0.08 & 125 & 30.6 & -3.4 & 3.78 & 0.36 & 20 & 1.2 & 1rxs j154502.0 - 405043 + mml 71 & 15455225 - 4222163 & k2-ive & 10.50 & 0.80 & 130 & 30.9 & -2.9 & 3.69 & 0.12 & 8 & 1.3 & 1rxs j154552.7 - 422227 + mml 72 & 15465179 - 4919048 & g7.5iv & 10.18 & 0.29 & 132 & 30.5 & -3.3 & 3.74 & 0.21 & 15 & 1.2 & 1rxs j154651.5 - 491922 + hip 77524 * & 15494499 - 3925089 & k1-ive & 10.64 & -0.03 & 151 & 30.5 & -3.3 & 3.71 & 0.23 & 9 & 1.3 & 1rxs j154944.7 - 392509 + hip 77656 & 15511373 - 4218513 & g5iv & 9.58 & 0.52 & 130 & 30.3 & -3.8 & 3.76 & 0.41 & 12 & 1.3 & 1rxs j155113.5 - 421858 + mml 73 * & 15565905 - 3933430 & g9.5iv & 10.81 & 0.35 & 174 & 30.5 & -3.4 & 3.73 & 0.21 & 11 & 1.3 & 1rxs j155659.0 - 393400 + hd 143099 * & 15595826 - 3824317 & g0v & 9.33 & 0.00 & 142 & 30.4 & -3.7 & 3.78 & 0.48 & 15 & 1.3 & 1rxs j155958.3 - 382352 + mml 74 & 16010792 - 3254526 & g0iv & 9.50 & 0.15 & 132 & 30.6 & -3.4 & 3.78 & 0.33 & 20 & 1.2 & 1rxs j160108.0 - 325455 + mml 75 & 16010896 - 3320141 & g5iv & 10.88 & 0.16 & 176 & 30.6 & -3.1 & 3.76 & 0.41 & 11 & 1.4 & 1rxs j160108.9 - 332021 + mml 76 & 16034536 - 4355492 & g9.5iv & 9.64 & 0.51 & 178 & 31.2 & -3.1 & 3.73 & 0.75 & 4 & 1.8 & 1rxs j160345.8 - 435544 + mml 77 * & 16035250 - 3939013 & k2-ive & 11.01 & 0.18 & 140 & 30.4 & -3.2 & 3.69 & 0.09 & 7 & 1.4 & 1rxs j160352.0 - 393901 + hd 143939b * & 16044404 - 3926117 & k3ve & 11.80 & 0.00 & 149 & 30.5 & -3.0 & 3.67 & -0.37 & 23 & 0.9 & 1rxs j160444.6 - 392602 + mml 78 * & 16054499 - 3906065 & g6.5iv & 10.53 & 0.19 & 130 & 30.5 & -3.1 & 3.75 & 0.13 & 20 & 1.1 & 1rxs j160545.8 - 390559 + hip 78881 * & 16060937 - 3802180 & f3v & 8.03 & 0.26 & 128 & 30.7 & -3.8 & 3.84 & 0.88 & 11 & 1.6 & 1rxs j160610.3 - 380215 + hip 79516 & 16133433 - 4549035 & f5v & 8.91 & 0.02 & 120 & 30.0 & -4.1 & 3.82 & 0.38 & 20 & 1.3 & 1rxs j161335.9 - 454901 + mml 79 & 16135801 - 3618133 & g9ive & 11.14 & 0.31 & 123 & 30.4 & -2.9 & 3.73 & -0.13 & 31 & 1.0 & 1rxs j161357.9 - 361813 + mml 80 & 16145207 - 5026187 & g9.5iv & 10.41 & 0.39 & 129 & 30.6 & -3.1 & 3.73 & 0.34 & 10 & 1.3 & 1rxs j161451.3 - 502621 + mml 81 & 16183856 - 3839117 & f9iv & 9.02 & 0.32 & 108 & 30.3 & -3.8 & 3.78 & 0.33 & 20 & 1.2 & 1rxs j161839.0 - 383927 + mml 82 & 16211219 - 4030204 & g8iv & 10.57 & 0.69 & 156 & 30.5 & -3.3 & 3.74 & 0.23 & 13 & 1.3 & 1rxs j162112.0 - 403032 + mml 83 & 16232955 - 3958008 & g3iv & 10.64 & 0.95 & 169 & 30.4 & -3.4 & 3.76 & 0.11 & 26 & 1.1 & 1rxs j162330.1 - 395806 + mml 84 & 16273054 - 3749215 & g9.5iv & 10.96 & 0.50 & 180 & 30.5 & -3.3 & 3.73 & 0.21 & 11 & 1.3 & 1rxs j162730.0 - 374929 + hip 80636 & 16275233 - 3547003 & g0.5iv & 9.37 & 0.39 & 152 & 30.9 & -3.3 & 3.77 & 0.65 & 10 & 1.4 & 1rxs j162752.8 - 354702 + mml 85 & 16314204 - 3505171 & g7.5iv & 10.64 & 0.33 & 143 & 30.4 & -3.3 & 3.74 & 0.05 & 22 & 1.1 & 1rxs j163143.7 - 350521 + mml 86 & 16353598 - 3326347 & k2-iv & 11.00 & 0.34 & 192 & 30.9 & -3.0 & 3.69 & 0.35 & 4 & 1.5 & 1rxs j163533.9 - 332631 + hip 81380 & 16371286 - 3900381 & g0iv & 9.82 & 0.33 & 200 & 30.1 & -3.8 & 3.78 & 0.74 & 5 & 1.7 & 1rxs j163713.7 - 390104 + hip 81447 & 16380553 - 3401106 & g0.5iv & 9.08 & 0.03 & 172 & 30.0 & -4.3 & 3.78 & 0.78 & 5 & 1.7 & 1rxs j163805.3 - 340110 + mml 87 & 16395929 - 3924592 & g4iv & 10.59 & 0.27 & 216 & 30.7 & -3.3 & 3.76 & 0.74 & 5 & 1.7 & 1rxs j163958.7 - 392457 + mml 88 & 16422399 - 4003296 & g1iv & 9.62 & 0.74 & 199 & 30.7 & -3.7 & 3.77 & 0.92 & 4 & 1.9 & 1rxs j164224.5 - 400329 + hd 151868 & 16514560 - 3803088 & f6v & 9.37 & 0.00 & 192 & @xmath10330.3 & @xmath103 - 4.0 & 3.80 & 0.60 & 15 & 1.4 & ... + hip 82569 & 16524171 - 3845372 & f3v & 8.85 & 0.13 & 185 & 30.5 & -3.9 & 3.84 & 0.82 & 12 & 1.5 & 1rxs j165242.7 - 384534 + hip 82747 & 16544485 - 3653185 & f5ive & 9.21 & 0.52 & 144 & @xmath10330.0 & @xmath103 - 4.3 & 3.81 & 0.76 & 15 & 1.4 & ... + hip 83159 & 16594248 - 3726168 & f5v & 9.02 & 0.00 & 152 & 29.9 & -4.3 & 3.82 & 0.54 & 19 & 1.4 & 1rxs j165943.1 - 372614 + [ table_ucl_list ] + name & 2mass j & spt & @xmath251 & a@xmath252 & dist & log@xmath253 & log@xmath254 & log@xmath255 & log@xmath256 & age & @xmath257 & x - ray + & & & [ mag ] & [ mag ] & [ pc ] & [ erg / s ] & & [ k ] & & [ myr ] & [ @xmath146 ] & counterpart + + name & 2mass j & spt & @xmath251 & a@xmath252 & dist & log@xmath253 & log@xmath254 & log@xmath255 & log@xmath256 & age & @xmath257 & x - ray + & & & [ mag ] & [ mag ] & [ pc ] & [ erg / s ] & & [ k ] & & [ myr ] & [ @xmath146 ] & counterpart + sacy 606 & 10065573 - 6352086 & k0v(e ) & 11.05 & 0.71 & 173 & 30.5 & -3.4 & 3.720 & 0.31 & 6 & 1.6 & 1rxs j100659.0 - 635212 + sacy 653 & 10494839 - 6446284 & g9ve & 11.73 & 0.21 & 195 & 30.5 & -3.2 & 3.733 & 0.09 & 17 & 1.2 & 1rxs j104949.5 - 644613 + mml 1 & 10574936 - 6913599 & k1+iv & 10.39 & 0.35 & 102 & 30.3 & -3.3 & 3.699 & 0.02 & 11 & 1.2 & 1rxs j105751.2 - 691402 + sacy 671 & 11080791 - 6341469 & m0ve & 11.86 & 0.00 & 113 & 30.0 & -3.2 & 3.585 & -0.34 & 3 & 0.7 & 1rxs j110808.9 - 634125 + hip 55334 & 11195276 - 7037065 & f2v & 8.14 & 0.24 & 87 & 29.8 & -4.3 & 3.838 & 0.60 & 27 & 1.4 & 1rxs j111948.2 - 703711 + twa 12 & 11210549 - 3845163 & m2e & 13.6 : & 0.45 & 109 & 30.1 & -3.4 & 3.550 & -0.39 & 2 & 0.6 & 1rxs j112105.2 - 384529 + sacy 681 & 11275535 - 6626046 & k1v(e ) & 10.82 & 0.57 & 109 & 30.3 & -3.2 & 3.706 & -0.06 & 16 & 1.1 & 1rxs j112755.0 - 662558 + mml 2 & 11320835 - 5803199 & g7iv & 9.92 & 0.68 & 93 & 29.9 & -3.7 & 3.747 & 0.19 & 17 & 1.2 & 1rxs j113209.3 - 580319 + sacy 684 & 11350376 - 4850219 & g7v & 10.28 & 0.00 & 153 & 30.5 & -3.2 & 3.751 & 0.06 & 25 & 1.1 & 1rxs j113501.5 - 485011 + hip 56673 & 11371464 - 6940272 & f5iv & 6.62 & 0.22 & 106 & 30.8 & -4.2 & 3.809 & 1.40 & 2 & 2.4 & 1rxs j113714.2 - 694025 + twa 19b & 11472064 - 4953042 & k7e & 11.6 : & 0.77 & 113 & @xmath10329.8 & @xmath103 - 3.8 & 3.605 & -0.20 & 2 & 0.8 & ... + hip 57524 & 11472454 - 4953029 & f9iv & 9.07 & 0.30 & 113 & 30.8 & -3.3 & 3.783 & 0.44 & 17 & 1.3 & 1rxs j114724.3 - 495250 + sacy 695 & 11515049 - 6407278 & k1v & 11.99 & 0.33 & 174 & 30.3 & -3.1 & 3.706 & -0.19 & 24 & 1.0 & 1rxs j115149.1 - 640705 + hip 57950 & 11530799 - 5643381 & f2iv / v & 8.26 & 0.17 & 99 & 29.9 & -4.3 & 3.838 & 0.64 & 18 & 1.4 & 1rxs j115308.5 - 564317 + sacy 699 & 11554295 - 5637314 & m0ve & 11.69 & 0.00 & 87 & 30.3 & -2.8 & 3.585 & -0.55 & 7 & 0.7 & 1rxs j115544.5 - 563739 + hip 58167 & 11554354 - 5410506 & f3iv & 8.30 & 0.06 & 103 & 29.7 & -4.5 & 3.829 & 0.63 & 15 & 1.4 & 1rxs j115543.4 - 541049 + sacy 700 & 11555771 - 5254008 & k4v & 11.00 & 0.01 & 109 & 29.9 & -3.5 & 3.662 & -0.26 & 13 & 1.1 & 1rxs j115554.5 - 525332 + sacy 706 & 11594608 - 6101132 & k4v(e ) & 11.36 & 0.01 & 133 & 30.1 & -3.3 & 3.662 & -0.23 & 12 & 1.1 & 1rxs j115946.5 - 610111 + hip 58528 & 12000940 - 5707021 & f5v & 8.54 & 0.05 & 100 & 29.5 & -4.6 & 3.809 & 0.51 & 17 & 1.3 & 1rxs j120009.5 - 570646 + sacy 708 & 12041439 - 6418516 & g8v & 9.93 & 0.00 & 135 & 30.8 & -3.0 & 3.742 & 0.16 & 17 & 1.2 & 1rxs j120413.3 - 641837 + mml 3 & 12044888 - 6409555 & g1iv & 9.41 & 0.60 & 120 & 30.8 & -3.2 & 3.773 & 0.44 & 15 & 1.3 & 1rxs j120448.2 - 640942 + hip 58996 & 12054748 - 5100121 & g1iv & 8.89 & 0.14 & 102 & 30.5 & -3.6 & 3.772 & 0.43 & 15 & 1.3 & 1rxs j120547.8 - 510007 + mml 4 & 12061352 - 5702168 & g4iv & 10.69 & 1.45 & 154 & 30.6 & -3.1 & 3.760 & 0.15 & 23 & 1.1 & 1rxs j120613.9 - 570215 + sacy 713 & 12063292 - 4247508 & k0v & 10.66 & 0.31 & 86 & 29.9 & -3.4 & 3.720 & -0.30 & 42 & 0.9 & 1rxs j120632.7 - 424750 + sacy 715 & 12074236 - 6227282 & k3ve & 10.90 & 0.41 & 120 & 30.3 & -3.4 & 3.675 & 0.06 & 6 & 1.3 & 1rxs j120741.4 - 622720 + mml 5 & 12094184 - 5854450 & k0ive & 10.09 & 0.28 & 111 & 30.5 & -3.2 & 3.720 & 0.20 & 12 & 1.2 & 1rxs j120941.5 - 585440 + mml 6 & 12113142 - 5816533 & g9iv & 10.19 & 0.28 & 108 & 30.5 & -3.2 & 3.729 & 0.05 & 18 & 1.1 & 1rxs j121131.9 - 581651 + mml 7 & 12113815 - 7110360 & g3.5iv & 9.15 & 0.48 & 99 & 30.5 & -3.5 & 3.762 & 0.39 & 14 & 1.3 & 1rxs j121137.3 - 711032 + sacy 724 & 12120804 - 6554549 & k3ve & 11.25 & 0.53 & 102 & 30.0 & -3.4 & 3.675 & -0.23 & 15 & 1.1 & 1rxs j121206.3 - 655456 + sacy 725 & 12121119 - 4950081 & k2ve & 11.37 & 0.58 & 113 & 30.1 & -3.3 & 3.690 & -0.24 & 21 & 1.0 & 1rxs j121210.7 - 494955 + mml 8 & 12123577 - 5520273 & k0+iv & 10.48 & 0.48 & 108 & 30.3 & -3.2 & 3.717 & 0.00 & 17 & 1.1 & 1rxs j121236.4 - 552037 + sacy 727 & 12124890 - 6230317 & k7ve & 11.47 & 0.43 & 101 & 30.7 & -2.6 & 3.609 & -0.31 & 4 & 0.8 & 1rxs j121248.7 - 623027 + hip 59603 & 12132235 - 5653356 & f2v & 8.56 & 0.31 & 115 & 29.8 & -4.5 & 3.838 & 0.71 & 14 & 1.5 & 1rxs j121321.6 - 565323 + sacy 728 & 12135700 - 6255129 & k4ve & 11.58 & 0.38 & 122 & 30.4 & -3.0 & 3.662 & -0.17 & 9 & 1.1 & 1rxs j121356.3 - 625508 + mml 9 & 12143410 - 5110124 & g9iv & 10.28 & 0.0 : & 106 & 30.4 & -3.2 & 3.733 & 0.01 & 21 & 1.0 & 1rxs j121434.2 - 511004 + hip 59716 & 12145071 - 5547235 & f5v & 8.45 & 0.11 & 104 & 30.9 & -3.3 & 3.809 & 0.60 & 15 & 1.4 & 1rxs j121452.4 - 554704 + mml 10 & 12145229 - 5547037 & g6iv & 9.64 & 0.87 & 103 & 30.9 & -2.9 & 3.750 & 0.28 & 14 & 1.2 & 1rxs j121452.4 - 554704 + hip 59764 & 12151855 - 6325301 & f8/g0v : & 8.43 & 0.25 & 109 & 30.8 & -3.5 & 3.786 & 0.72 & 10 & 1.5 & 1rxs j121518.6 - 632517 + hip 59781 & 12152822 - 6232207 & f8/g0v & 9.12 & 0.30 & 102 & 29.9 & -4.1 & 3.786 & 0.40 & 19 & 1.2 & 1rxs j121529.1 - 623209 + sacy 738 & 12160114 - 5614068 & k5ve & 11.22 & 0.64 & 91 & 30.3 & -3.0 & 3.638 & -0.29 & 8 & 1.0 & 1rxs j121601.8 - 561405 + hip 59854 & 12162783 - 5008356 & g1iv & 9.34 & 0.37 & 130 & 30.9 & -3.2 & 3.772 & 0.49 & 13 & 1.4 & 1rxs j121627.9 - 500829 + sacy 740 & 12163007 - 6711477 & k4ive & 11.56 & 0.58 & 107 & 30.3 & -3.2 & 3.662 & -0.08 & 7 & 1.2 & 1rxs j121630.6 - 671146 + mml 11 & 12182762 - 5943128 & k1.5ive & 10.73 & 0.12 & 96 & 30.3 & -3.1 & 3.695 & 0.03 & 10 & 1.2 & 1rxs j121828.6 - 594307 + mml 12 & 12185802 - 5737191 & g9.5iv & 9.87 & 0.49 & 106 & 30.6 & -3.2 & 3.724 & 0.32 & 10 & 1.3 & 1rxs j121858.2 - 573713 + mml 13 & 12192161 - 6454101 & k1-iv & 10.11 & 0.25 & 103 & 30.5 & -3.2 & 3.707 & 0.26 & 8 & 1.3 & 1rxs j121919.4 - 645406 + hip 60205 & 12204420 - 5215249 & f5 & 10.06 & 0.19 & 147 & 29.8 & -4.0 & 3.809 & 0.28 & ... & 1.4 : & 1rxs j122047.8 - 521509 + sacy 750 & 12205449 - 6457242 & k4ve & 11.00 & 0.00 & 94 & 29.9 & -3.4 & 3.662 & -0.22 & 11 & 1.1 & 1rxs j122050.8 - 645724 + sacy 751 & 12210808 - 5212226 & k4ve & 11.85 & 0.40 & 106 & 30.0 & -3.2 & 3.662 & -0.41 & 22 & 0.9 & 1rxs j122108.0 - 521217 + mml 14 & 12211648 - 5317450 & g1.5iv & 9.34 & 0.44 & 107 & 30.7 & -3.3 & 3.770 & 0.31 & 18 & 1.2 & 1rxs j122116.7 - 531747 + mml 15 & 12215566 - 4946125 & g5.5iv & 10.02 & 0.14 & 102 & 30.2 & -3.4 & 3.754 & 0.08 & 25 & 1.1 & 1rxs j122155.9 - 494609 + mml 16 & 12220430 - 4841248 & k0ive & 10.50 & 0.21 & 126 & 30.4 & -3.2 & 3.721 & 0.10 & 13 & 1.2 & 1rxs j122204.0 - 484118 + hip 60348 & 12222484 - 5101343 & f5v & 8.80 & 0.06 & 104 & 29.8 & -4.2 & 3.809 & 0.44 & 19 & 1.3 & 1rxs j122226.1 - 510120 + mml 17 & 12223322 - 5333489 & g0iv & 9.42 & 0.41 & 124 & 30.3 & -3.7 & 3.778 & 0.41 & 17 & 1.3 & 1rxs j122233.4 - 533347 + mml 18 & 12234012 - 5616325 & k0+iv & 10.85 & 0.44 & 112 & 30.0 & -3.4 & 3.715 & -0.09 & 21 & 1.1 & 1rxs j122339.9 - 561628 + sacy 759 & 12242065 - 5443540 & g5v & 10.37 & 1.14 & 145 & 30.7 & -3.5 & 3.761 & 0.56 & 7 & 1.8 & 1rxs j122421.0 - 544343 + hip 60567 & 12245491 - 5200157 & f6/f7v & 9.77 & 0.19 & 136 & 30.1 & -3.8 & 3.801 & 0.34 & 23 & 1.3 & 1rxs j122452.5 - 520014 + sacy 762 & 12264842 - 5215070 & k5ve & 11.66 & 0.47 & 92 & 29.9 & -3.2 & 3.638 & -0.42 & 14 & 1.0 & 1rxs j122648.5 - 521453 + sacy 764 & 12282540 - 6320589 & g7v & 9.25 & 0.32 & 109 & 30.5 & -3.5 & 3.751 & 0.41 & 9 & 1.5 & 1rxs j122823.4 - 632100 + hip 60885 & 12284005 - 5527193 & g0iv & 8.89 & 0.16 & 105 & 30.3 & -3.8 & 3.778 & 0.46 & 15 & 1.3 & 1rxs j122840.3 - 552707 + hip 60913 & 12290224 - 6455006 & g4.5iv & 9.04 & 0.21 & 102 & 29.6 & -4.4 & 3.758 & 0.43 & 12 & 1.3 & 1rxs j122858.3 - 645448 + sacy 766 & 12302957 - 5222269 & k3v(e ) & 12.04 & 0.85 & 103 & 29.9 & -3.3 & 3.675 & -0.41 & 28 & 0.9 & 1rxs j123031.1 - 522221 + sacy 768 & 12333381 - 5714066 & k1v(e ) & 10.92 & 0.69 & 92 & 29.9 & -3.5 & 3.706 & -0.16 & 22 & 1.0 & 1rxs j123332.4 - 571345 + sacy 769 & 12361767 - 5042421 & k4ve & 11.39 & 0.00 & 105 & 29.9 & -3.3 & 3.662 & -0.45 & 25 & 0.9 & 1rxs j123620.3 - 504238 + mml 19 & 12363895 - 6344436 & k0+iii & 9.87 & 0.39 & 103 & 30.4 & -3.4 & 3.717 & 0.25 & 10 & 1.3 & 1rxs j123637.5 - 634446 + mml 20 & 12365895 - 5412178 & k0iv & 10.40 & 0.11 & 116 & 30.3 & -3.4 & 3.721 & 0.04 & 16 & 1.2 & 1rxs j123657.4 - 541217 + sacy 773 & 12383556 - 5916438 & k3ve & 11.62 & 0.55 & 87 & 29.9 & -3.2 & 3.675 & -0.50 & 38 & 0.8 & 1rxs j123834.9 - 591645 + mml 21 & 12393796 - 5731406 & g7.5iv & 10.12 & 0.30 & 98 & 30.4 & -3.1 & 3.742 & -0.06 & 30 & 1.0 & 1rxs j123938.4 - 573141 + sacy 779 & 12404664 - 5211046 & k2v(e ) & 11.91 & 0.34 & 135 & 30.1 & -3.2 & 3.690 & -0.26 & 23 & 1.0 & 1rxs j124046.9 - 521108 + mml 22 & 12411820 - 5825558 & g1.5iv & 9.93 & 0.67 & 97 & 30.2 & -3.4 & 3.771 & 0.14 & 26 & 1.1 & 1rxs j124118.5 - 582556 +
sacy 782 & 12442412 - 5855216 & k2ve & 10.26 & 0.48 & 106 & 30.3 & -3.4 & 3.690 & 0.15 & 6 & 1.4 & 1rxs j124423.9 - 585428 + hip 62171 & 12442659 - 5420480 & f3v & 8.90 & 0.25 & 113 & 29.6 & -4.6 & 3.829 & 0.54 & 26 & 1.4 & 1rxs j124427.5 - 542043 + mml 23 & 12443482 - 6331463 & k1ive & 10.78 & 0.27 & 125 & 30.6 & -3.0 & 3.704 & 0.25 & 8 & 1.3 & 1rxs j124432.6 - 633139 + mml 24 & 12450674 - 4742580 & g8.5iv & 10.40 & 0.38 & 120 & 30.6 & -3.1 & 3.738 & 0.13 & 17 & 1.2 & 1rxs j124506.9 - 474254 + sacy 787 & 12454884 - 5410583 & k2v(e ) & 11.40 & 0.60 & 129 & 30.2 & -3.4 & 3.690 & -0.04 & 11 & 1.2 & 1rxs j124547.1 - 541104 + pf96 1 & 12464097 - 5931429 & m3e & 13.9 : & 0.13 & 110 : & 30.0 & -2.8 & 3.532 & -0.85 & 4 & 0.4 & 1wga j1246.6 - 5931 + hip 62427 & 12473870 - 5824567 & f8 & 9.28 & 0.00 & 134 & 30.5 & -3.5 & 3.792 & 0.39 & 19 & 1.3 & 1rxs j124742.1 - 582544 + hip 62431 & 12474180 - 5825558 & f0 & 8.02 & 0.34 & 140 & 30.5 & -4.2 & 3.857 & 1.10 & 5 & 1.9 & 1rxs j124742.1 - 582544 + sacy 789 & 12474824 - 5431308 & m0ve & 11.78 & 0.03 & 96 & 30.2 & -3.1 & 3.585 & -0.26 & 2 & 0.7 & 1rxs j124747.8 - 543141 + hip 62445 & 12475186 - 5126382 & g4.5ive & 9.52 & 0.66 & 130 & 30.7 & -3.4 & 3.758 & 0.62 & 5 & 1.7 & 1rxs j124751.7 - 512638 + mml 25 & 12480778 - 4439167 & g6ive & 9.73 & 0.18 & 91 & 30.6 & -3.1 & 3.751 & 0.14 & 20 & 1.1 & 1rxs j124807.6 - 443913 + pf96 3a & 12483152 - 5944493 & k5e & 11.5 : & 0.63 & 110 : & 29.9 : & -3.5 : & 3.644 & -0.10 & 5 & 1.1 & 1wga j1248.5 - 5944 + pf96 3b & 12483152 - 5944493 & k5e & 11.5 : & 0.63 & 110 : & 29.9 : & -3.5 : & 3.644 & -0.10 & 5 & 1.1 & 1wga j1248.5 - 5944 + mml 26 & 12484818 - 5635378 & g5iv & 10.22 & 0.53 & 132 & 30.4 & -3.4 & 3.755 & 0.14 & 22 & 1.1 & 1rxs j124847.4 - 563525 + pf96 4 & 12485496 - 5949476 & m4e & 12.7 : & 0.07 & 110 : & 29.8 & -2.9 & 3.517 & -0.92 & 3 & 0.3 & 1wga j1248.9 - 5949 + pf96 6 & 12492437 - 5913112 & m1e & 12.8 : & 0.10 & 110 : & 29.8 & -3.2 & 3.564 & -0.60 & 5 & 0.6 & 1wga j1249.3 - 5913 + hip 62657 & 12501971 - 4951488 & f5/f6v & 8.91 & 0.09 & 107 & 29.7 & -4.3 & 3.806 & 0.43 & 19 & 1.3 & 1rxs j125020.5 - 495144 + sacy 797 & 12505143 - 5156353 & k5ve & 11.67 & 0.33 & 109 & 29.9 & -3.4 & 3.638 & -0.33 & 10 & 1.0 & 1rxs j125051.3 - 515655 + sacy 799 & 12560830 - 6926539 & k7ve & 11.80 & 0.62 & 118 & 30.8 & -2.7 & 3.609 & -0.05 & 2 & 0.8 & 1rxs j125608.8 - 692652 + sacy 800 & 12560940 - 6127256 & k0ve & 9.62 & 0.27 & 75 & 30.7 & -3.0 & 3.720 & 0.05 & 15 & 1.2 & 1rxs j125609.4 - 612724 + hip 63272 & 12575777 - 5236546 &
f3iv / v & 8.40 & 0.00 & 113 & 29.6 & -4.6 & 3.829 & 0.63 & 15 & 1.4 & 1rxs j125757.2 - 523659 + mml 27 & 12582559 - 7028490 & k0+iv & 9.91 & 0.46 & 85 & 30.3 & -3.3 & 3.714 & -0.01 & 16 & 1.1 & 1rxs j125824.6 - 702848 + mml 28 & 13015069 - 5304581 & k2-iv & 11.08 & 0.55 & 108 & 30.0 & -3.3 & 3.690 & -0.37 & 32 & 0.9 & 1rxs j130153.7 - 530446 + mml 29 & 13023752 - 5459370 & g1iv & 10.28 & 0.38 & 156 & 30.8 & -3.1 & 3.772 & 0.35 & 18 & 1.2 & 1rxs j130237.2 - 545933 + hip 63847 & 13050530 - 6413552 & g3iv & 9.18 & 0.33 & 98 & 30.4 & -3.5 & 3.764 & 0.36 & 15 & 1.3 & 2rxp j130506.7 - 641346 + hd 113703b & 13061785 - 4827456 & k0ve & 10.8 : & 0.00 & 118 & 30.5 & -2.9 & 3.717 & -0.05 & 19 & 1.1 & 1rxs j130618.4 - 482744 + hip 63975 & 13063577 - 4602018 & f3/f5v & 8.08 & 0.00 & 111 & 29.4 & -4.8 & 3.819 & 0.70 & 12 & 1.4 & 1rxs j130630.7 - 460215 + mml 30 & 13064012 - 5159386 & k0ive & 10.53 & 0.17 & 112 & 30.2 & -3.4 & 3.722 & -0.05 & 21 & 1.1 & 1rxs j130638.5 - 515948 + sacy 808 & 13065439 - 4541313 & k5ve & 12.03 & 0.53 & 124 & 30.2 & -3.1 & 3.638 & -0.28 & 8 & 1.0 & 1rxs j130655.0 - 454125 + hd 113791b & 13065720 - 4954265 & f7v & 10.90 & 0.03 & 145 & 29.9 & -3.5 & 3.798 & 0.49 & 17 & 1.3 & 1rxs j130659.7 - 495405 + hip 64044 & 13073350 - 5254198 & f5v & 8.83 & 0.27 & 111 & 30.3 & -3.8 & 3.809 & 0.57 & 16 & 1.4 & 1rxs j130733.2 - 525421 + sacy 814 & 13130714 - 4537438 & k5ve & 11.63 & 0.52 & 101 & 29.9 & -3.4 & 3.638 & -0.23 & 6 & 1.1 & 1rxs j131306.7 - 453740 + sacy 815 & 13132810 - 6000445 & g3v & 10.01 & 0.20 & 157 & 30.6 & -3.4 & 3.766 & 0.40 & 14 & 1.3 & 1rxs j131327.3 - 600032 + mml 31 & 13142382 - 5054018 & g5.5iv & 10.39 & 0.27 & 130 & 30.8 & -2.9 & 3.754 & 0.25 & 16 & 1.2 & 1rxs j131424.3 - 505402 + mml 32 & 13175694 - 5317562 & g1iv & 10.41 & 0.22 & 167 & 30.7 & -3.2 & 3.774 & 0.43 & 15 & 1.3 & 1rxs j131754.9 - 531758 + twa 17 & 13204539 - 4611377 & k7e & 13.1 : & 0.67 & 142 & 30.2 & -3.0 & 3.635 & -0.38 & 11 & 1.0 & 1rxs j132046.5 - 461139 + twa 18 & 13213722 - 4421518 & m0.5e & 13.7 : & 0.16 & 119 & 30.0 & -2.9 & 3.574 & -0.63 & 6 & 0.6 & 1rxs j132137.0 - 442133 + mml 33 & 13220446 - 4503231 & g0iv & 10.02 & 0.22 & 140 & 30.3 & -3.6 & 3.779 & 0.18 & 25 & 1.1 & 1rxs j132204.7 - 450312 + mml 34 & 13220753 - 6938121 & k1ive & 10.44 & 0.17 & 86 & 30.2 & -3.2 & 3.702 & 0.00 & 13 & 1.2 & 1rxs j132207.2 - 693812 + sacy 822 & 13233587 - 4718467 & k3ve & 11.19 & 1.23 & 111 & 30.0 & -3.6 & 3.733 & 0.06 & 19 & 1.1 & 1rxs j132336.3 - 471844 + hip 65517 & 13254783 - 4814577 & g1.5iv & 9.76 & 0.14 & 95 & 30.4 & -3.3 & 3.770 & 0.02 & 35 & 1.1 & 1rxs j132548.2 - 481451 + sacy 828 & 13270594 - 4856180 & k3ve & 10.68 & 0.51 & 104 & 30.4 & -3.2 & 3.675 & 0.01 & 7 & 1.3 & 1rxs j132706.3 - 485617 + hip 66001 & 13315360 - 5113330 & g2.5iv & 9.84 & 0.11 & 127 & 30.6 & -3.2 & 3.766 & 0.33 & 17 & 1.2 & 1rxs j133152.6 - 511335 + mml 35 & 13342026 - 5240360 & g1iv & 9.29 & 0.45 & 105 & 30.5 & -3.4 & 3.774 & 0.32 & 19 & 1.2 & 1rxs j133420.0 - 524032 + sacy 835 & 13343188 - 4209305 & k2ive & 9.64 & 0.49 & 91 & 30.2 & -3.2 & 3.690 & -0.15 & 16 & 1.1 & 1rxs j133432.2 - 420929 + sacy 841 & 13402554 - 4633514 & k3ve & 11.37 & 0.68 & 143 & 30.8 & -2.9 & 3.675 & 0.08 & 5 & 1.4 & 1rxs j134025.6 - 463323 + hip 66941 & 13430870 - 6907393 & g0.5iv & 7.57 & 0.38 & 107 & 31.1 & -3.5 & 3.775 & 1.09 & 4 & 2.0 & 1rxs j134306.8 - 690754 + mml 37 & 13432853 - 5436434 & g2iv & 9.32 & 0.28 & 85 & 30.6 & -3.2 & 3.769 & 0.13 & 26 & 1.1 & 1rxs j134332.7 - 543638 + sacy 848 & 13444279 - 6347495 & k4ve & 11.04 & 0.50 & 84 & 30.2 & -3.2 & 3.662 & -0.24 & 12 & 1.1 & 1rxs j134442.5 - 634758 + sacy 849 & 13455599 - 5222255 & k3ve & 11.34 & 0.49 & 121 & 30.4 & -3.1 & 3.675 & -0.13 & 11 & 1.2 & 1rxs j134555.2 - 522215 + sacy 857 & 13540743 - 6733449 & g6v & 10.93 & 0.56 & 142 & 30.2 & -3.4 & 3.756 & 0.05 & 27 & 1.1 & 1rxs j135404.0 - 673334 + mml 42 & 14160567 - 6917359 & g1iv & 10.14 & 0.62 & 123 & 30.1 & -3.7 & 3.773 & 0.19 & 24 & 1.1 & 1rxs j141605.3 - 691756 + [ table_lcc_list ] briceno c. , preibisch , th . , sherry , w. , mamajek , e. , mathieu , r. , walter , f. , zinnecker , h. , 2007 , in _ protostars & planets v _ ,
b. reipurth , d. jewitt , & k. keil , university of arizona press , tucson , p. 345 | we summarize observational results on the stellar population and star formation history of the scorpius - centaurus ob association ( sco ob2 ) , the nearest region of recent massive star formation .
it consists of three subgroups , upper scorpius ( us ) , upper centaurus - lupus ( ucl ) , and lower centaurus - crux ( lcc ) which have ages of about 5 , 17 , and @xmath0 myr .
while the high- and intermediate mass association members have been studied for several decades , the low - mass population remained mainly unexplored until rather recently . in upper scorpius
, numerous studies , in particular large multi - object spectroscopic surveys , have recently revealed hundreds of low - mass association members , including dozens of brown dwarfs .
the investigation of a large representative sample of association members provided detailed information about the stellar population and the star formation history . the empirical mass function
could be established over the full stellar mass range from @xmath1 up to @xmath2 , and was found to be consistent with recent determinations of the field initial mass function .
a narrow range of ages around 5 myr was found for the low - mass stars , the same age as had previously ( and independently ) been derived for the high - mass members .
this supports earlier indications that the star formation process in us was triggered , and agrees with previous conjectures that the triggering event was a supernova- and wind - driven shock - wave originating from the nearby ucl group . in the older ucl and lcc regions , large numbers of low - mass members
have recently been identified among x - ray and proper - motion selected candidates . in both subgroups , low - mass members have also been serendipitously discovered through investigations of x - ray sources in the vicinity of better known regions ( primarily the lupus and tw hya associations ) .
while both subgroups appear to have mean ages of @xmath316myr , they both show signs of having substructure
. their star - formation histories may be more complex than that of the younger , more compact us group .
sco - cen is an important `` astrophysics laboratory '' for detailed studies of recently formed stars .
for example , the ages of the sub - groups of 5 myr and @xmath4 myr are ideal for studying how circumstellar disks evolve .
while no more than a few percent of the sco - cen members appear to be accreting from a circumstellar disk , recent _ spitzer _ results suggest that at least @xmath335% still have cold , dusty , debris disks . |
recently the branching ratios of the @xmath0 decays have been measured by the babar@xcite , belle@xcite and cleo@xcite collaborations .
these decay modes are important to understand penguin dynamics , because the @xmath0 modes are pure penguin processes .
there is a following problem related a penguin contribution for amplitude@xcite .
a naive estimate of the loop diagram leads to @xmath2 where @xmath3 is a penguin amplitude and @xmath4 is a tree amplitude . but
experimental data for @xmath5 and @xmath6 leads to @xmath7 .
the penguin contribution is enhanced dynamically .
this enhancement can not be explained by factorization assumption for nonleptonic two - body @xmath8 meson decays@xcite .
this problem is explained by keum , li and sanda using perturbative qcd ( pqcd ) approach@xcite .
pqcd method for inclusive decays was developed by many researcher over many years , and this formalism has been successful .
recently , pqcd has been applied to the exclusive b meson decays , @xmath9@xcite , @xmath10@xcite , @xmath11 , @xmath12@xcite , @xmath13@xcite and @xmath14@xcite .
we calculate the branching ratios for the @xmath0 modes using this approach to leading order in @xmath15 and leading power in @xmath16 where @xmath17 is the @xmath8 meson mass .
the @xmath0 modes depend only on penguin amplitudes , and the branching ratio for this modes is proportional to @xmath18 .
thus the @xmath0 modes are useful to seeing the size of the penguin amplitudes . in this paper
we compute the branching ratio for the @xmath0 modes using pqcd approach . in sec . [
sec : pf ] , we introduce the pqcd formalism for exclusive b meson decays . in sec .
[ sec : amp ] , we calculate analytic formula of the factorizable amplitudes and the nonfactorizable amplitudes of the decays . in sec .
[ sec : num ] , we predict the branching ratios of the decays and the direct cp asymmetry of the charged mode . our predicted branching ratios agree with the current experimental data and are larger than the values obtained from the naive factorization assumption and the qcd - improved factorization@xcite@xcite . in the appendix ,
we review the meson wave functions with up to twist-3 terms which are proportional to @xmath19 or @xmath20 .
the effective hamiltonian for @xmath21 transitions is @xmath22\ ; , \label{hbk}\ ] ] where @xmath23 and @xmath24 are the cabibbo - kobayashi - maskawa matrix elements@xcite and @xmath25 are local four - fermi operators .
the local operators are given by @xmath26 where @xmath27 is the color indices and @xmath28 is taken @xmath29 and @xmath30 .
the wilson coefficients @xmath31 are calculated up to next - to - leading order by buras _
et al_.@xcite .
pqcd approach is based on the three scale factorization theorem@xcite@xcite in which the matrix element can be written as the convolution of the hard part @xmath32 with the hard scale @xmath33 , the meson wave function @xmath34 with the soft scale @xmath35 and the wilson coefficient @xmath36 whose evolves from the @xmath37 boson mass @xmath38 down to the scale @xmath39 .
@xmath40 is the conjugate variable of the parton transverse momenta @xmath1 .
for instance , the @xmath41 transition form factor is written as @xmath42[db ] c(t ) \phi_k(x_2,b_2 ) h(t)\phi_b(x_1,b_1 ) \exp\left[-\sum_{i=1,2}\int_{1/b_i}^t\frac{d{\bar\mu } } { \bar\mu}\gamma_\phi(\alpha_s({\bar\mu}))\right]\;,\ ] ] where @xmath43 is the momentum fraction of parton and @xmath44 is the anomalous dimension of meson .
the hard part @xmath32 can be calculated perturbatively and the feynman diagrams are described as fig .
[ fig : lowest ] in leading order of @xmath15 . transition form factor in leading order of @xmath15 ] it is important that the scale of the wilson coefficient is determined dynamically , since the scale depends on @xmath43 and @xmath40 . then in pqcd , the scale of the wilson coefficient can reach lower scale than @xmath46 which are usually taken in the naive factorization assumption case .
the wilson coefficients for the penguin operator are enhanced as the scale evolves to lower scale .
therefore , the penguin amplitudes are larger than those in the naive factorization assumption . in this method , we can calculate not only factorizable amplitudes but also nonfactorizable and annihilation amplitudes which can not be calculated in the factorization assumption .
this @xmath41 transition form factor has two types of double logarithmic corrections in infrared divergences .
the one is @xmath47 from the overlap of collinear and soft divergences which are generated by the correction of the meson wave function as fig .
[ fig : resum](a ) and fig .
[ fig : resum](b ) .
the resummation of this double logarithms leads to the factor @xmath48$]@xcite@xcite .
the explicit form of this factor is shown , for example , in ref .
this resummation is called @xmath1 resummation .
the other double logarithm is @xmath49 from the end - point region of the momentum fraction @xmath43@xcite .
this double logarithm is generated by the corrections of the hard part as fig .
[ fig : resum](c ) and fig .
[ fig : resum](d ) .
the resummation of this double logarithms is called the threshold resummation .
we use the approximate form of this resummation factor that is @xmath50 where @xmath51 and @xmath52@xcite . the sudakov factor from @xmath1 and
threshold resummations suppress large @xmath40 region and end - point of @xmath43 where the hard part @xmath32 is singular , the hard part @xmath32 can then be calculated perturbatively .
( a ) and fig . [ fig : resum](b ) are the correction of the meson wave function .
[ fig : resum](c ) and fig .
[ fig : resum](c ) are the correction of the hard part . ] in pqcd approach , we can calculate hard part perturbatively , and we use the meson wave functions which are formed in the light - cone qcd sum rules .
the hard part depends on processes , but the wave functions are independent of those
. we can determine the parameters in the wave functions , since there are many modes in the @xmath8 meson two - body decays .
so , we can predict quantitative values for the other decays . in this study
, we calculate the hard part in leading order of @xmath15 .
the meson wave functions are used with leading twist and twist-3 terms which are proportional to @xmath19 or @xmath20 .
we review the meson wave functions in the appendix .
in the light - cone coordinate , the @xmath8 meson momentum @xmath53 , the @xmath54 meson momentum @xmath55 and the @xmath56 meson momentum @xmath57 are taken to be @xmath58 where @xmath59 , and the @xmath54 meson mass is neglected . in this situation ,
the @xmath8 meson is at rest .
the quark and anti - quark momenta in the mesons are displayed in fig .
[ fig : mom ] .
their components are @xmath60 we can integrate over @xmath61 and choose @xmath62 , since the hard part is independent of @xmath61 .
the @xmath56 meson longitudinal polarization vector @xmath63 and two transverse polarization vector @xmath64 are given by @xmath65 the @xmath8 meson wave function for incoming state and the @xmath54 and @xmath56 meson wave functions for outgoing state with up to twist-3 terms are written as @xmath66_{\alpha\beta}\;,\\ \phi^{(out)}_{k,\alpha\beta , ij } & = & \frac{-i\delta_{ij}}{\sqrt{2n_c}}\int^1_0dx_2 e^{ix_2p_2\cdot z_2 } \gamma_5\left [ \not p_2\phi^a_k(x_2)+m_{0k}\phi_k^p(x_2 ) + m_{0k } ( \not v\not
n -1)\phi_k^t(x_2 ) \right]_{\alpha\beta}\;,\\ \phi^{(out)}_{\phi,\alpha\beta , ij } & = & \frac{\delta_{ij}}{\sqrt{2n_c}}\int^1_0dx_3 e^{ix_3p_3\cdot z_3 } \left [ m_\phi\not\epsilon_\phi \phi_\phi(x_3)+ \not\epsilon_\phi \not p_3\phi_\phi^t(x_3 ) + m_\phi\phi_\phi^s(x_3 ) \right]_{\alpha\beta}\;,\end{aligned}\ ] ] where @xmath67 and @xmath68 is the color indices , @xmath69 and @xmath70 are the dirac indices .
@xmath71 is related to chiral symmetry breaking scale , @xmath72 .
@xmath73 and @xmath74 are defined as @xmath75 and @xmath76 .
we neglect the wave functions which are proportional to the transverse polarization vector @xmath77 , because these terms do not appear in our calculations .
it must be noted that the expression of @xmath78 depends on how to define the sign of @xmath79 .
the explicit form of these wave functions will be shown in sec .
[ sec : num ] and the detail is shown in appendix .
the decay rates for the @xmath0 have the expressions @xmath80 the decay amplitudes @xmath81 and @xmath82 corresponding to @xmath83 and @xmath84 respectively , are written as @xmath85 @xmath86 is the amplitude for the factorizable diagrams which are considered in the factorization assumption . @xmath87 and @xmath88 are the annihilation factorizable and the nonfactorizable diagrams which are neglected in the factorization assumption .
the indices @xmath89 and @xmath90 denote the tree topology as fig .
[ fig : mom](a ) and annihilation topology as fig .
[ fig : mom](b ) respectively .
the index @xmath3 denotes the contribution from the diagram with a penguin operator .
the decay amplitudes @xmath91 and @xmath92 corresponding to @xmath93 and @xmath94 respectively , are written as @xmath95 where there are very small contribution from tree diagrams , the index @xmath4 denotes these contribution .
the factorizable amplitudes which come from fig .
[ fig : fact ] are written as @xmath96 \nonumber\\ & & \;\;\;\;\;\;\times e_{e}(t^{(1)}_e ) n_t \{x_2(1-x_2)\}^c h_e(x_1,x_2,b_1,b_2 ) \nonumber\\ & & \;\;\;\;\;\ ; + 2r_k \phi_k^p(x_2)e_{e}(t^{(2)}_e ) n_t \{x_1(1-x_1)\}^c h_e(x_2,x_1,b_2,b_1 ) \bigg\}\;,\\ f^p_a&=&f^p_{a4}+f^p_{a6}\ ; , \nonumber\\ f^p_{a4}&= & 8\pi c_f m_b^4 \int_0 ^ 1 dx_2 dx_3 \int_0^{\infty } b_2db_2 b_3db_3 \nonumber \\ & & \times \bigg\ { \left [ x_3\phi_k^a(x_2)\phi_{\phi}(x_3 ) -2r_kr_\phi \phi_k^p(x_2 ) \left ( \left ( \phi_{\phi}^t(x_3 ) - \phi_{\phi}^s(x_3 ) \right ) - x_3\left ( \phi_{\phi}^t(x_3 ) + \phi_{\phi}^s(x_3 )
\right ) \right ) \right ] \nonumber \\ & & \;\;\;\;\;\;\times e_{a4}(t^{(1)}_a ) n_t \{x_3(1-x_3)\}^c
h_a(1-x_2,x_3,b_2,b_3 ) \nonumber\\ & & \;\;\;\;\;\ ; - \left [ ( 1-x_2)\phi_k^a(x_2)\phi_{\phi}(x_3 ) + 2r_kr_\phi \left ( 2\phi_k^p(x_2 ) - x_2\left ( \phi_k^p(x_2)-\phi_k^t(x_2 )
\right ) \right ) \phi_{\phi}^s(x_3 ) \right ] \nonumber \\ & & \;\;\;\;\;\;\times e_{a4}(t^{(2)}_a ) n_t \{x_2(1-x_2)\}^c h_a(x_3,1-x_2,b_3,b_2 ) \bigg\}\;,\\ f^p_{a6}&= & 16\pi c_f m_b^4 \int_0 ^ 1 dx_2 dx_3 \int_0^{\infty } b_2db_2 b_3db_3 \nonumber \\ & & \times \bigg\ { \left [ 2 r_k \phi_k^p(x_2)\phi_{\phi}(x_3 ) - r_{\phi}x_3 \phi_k^a(x_2 ) \left ( \phi_{\phi}^t(x_3 ) - \phi_{\phi}^s(x_3 ) \right ) \right ] \nonumber\\ & & \;\;\;\;\;\;\times
e_{a6}(t^{(1)}_a ) n_t \{x_3(1-x_3)\}^c h_a(1-x_2,x_3,b_2,b_3 ) \nonumber\\ & & \;\;\;\;\;\ ; + \left [ r_k ( 1-x_2)\left ( \phi_k^p(x_2)+\phi_k^t(x_2 ) \right)\phi_{\phi}(x_3 ) + 2 r_{\phi } \phi_k^a(x_2 ) \phi_{\phi}^s(x_3 ) \right ] \nonumber\\ & & \;\;\;\;\;\;\times e_{a6}(t^{(2)}_a ) n_t \{x_2(1-x_2)\}^c h_a(x_3,1-x_2,b_3,b_2 ) \bigg\}\;.\end{aligned}\ ] ] the tree amplitude @xmath97 is the same as @xmath98 , but the wilson coefficient is different .
@xmath99 is the factor for the threshold resummation .
@xmath100 is a normalization factor and @xmath30 is a constant , we use @xmath51 and @xmath52 .
the evolution factors are defined by @xmath101\;,\;\ ; e_{ai}(t)=\alpha_s(t)a_{ai}(t)\exp[-s_k(t)-s_{\phi}(t)]\;.\end{aligned}\ ] ] the explicit forms of the factor @xmath102 are given , for example , in ref .
the hard scales @xmath39 , which are the typical scales in hard process , are given by @xmath103 the wilson coefficients are given by @xmath104 where @xmath105 is the wilson coefficient for the amplitude @xmath97 .
the hard functions , which are the fourier transformation of the virtual quark propagator and the hard gluon propagator , are given by @xmath106\;,\\ h_a(x_2,x_3,b_2,b_3)&=&\left(\frac{i\pi}{2}\right)^2 h_0^{(1)}\left(\sqrt{x_2x_3}m_bb_2\right ) \left[\theta(b_2-b_3 ) h_0^{(1)}\left(\sqrt{x_3}m_bb_2\right ) j_0\left(\sqrt{x_3}m_bb_3\right)\right .
\nonumber \\ & & \hspace{41 mm } \left.+\theta(b_3-b_2)h_0^{(1 ) } \left(\sqrt{x_3}m_bb_3\right ) j_0\left(\sqrt{x_3}m_bb_2\right)\right]\;.\end{aligned}\ ] ] the nonfactorizable amplitudes which come from fig .
[ fig : nonfact ] are written as @xmath107\int_0^{\infty } b_1 db_1 b_3 db_3\phi_b(x_1,b_1)\phi_{\phi}(x_3 ) \nonumber \\ & & \times \bigg\ { \left[(1-x_1-x_3)\phi_k^a(x_2 ) -r_k x_2 \left ( \phi^p_k(x_2 ) - \phi^t_k(x_2 ) \right ) \right ] e'_{e4}(t^{(1)}_d ) h^{(1)}_d(x_1,x_2,x_3,b_1,b_1,b_3 ) \nonumber \\ & & \;\;\;\;\;\ ; + \left[(x_1-x_2-x_3)\phi_k^a(x_2 ) + r_k x_2 \left ( \phi_k^p(x_2)+\phi_k^t(x_2 ) \right ) \right ]
e'_{e4}(t^{(2)}_d ) h^{(2)}_d(x_1,x_2,x_3,b_1,b_1,b_3 ) \bigg\}\ ; , \\
{ \cal m}^p_{e5}&= & 16\pi c_f \frac{\sqrt{2n_c}}{n_c}m_b^4\int_0 ^ 1 [ dx]\int_0^{\infty } b_1 db_1 b_3 db_3\phi_b(x_1,b_1)\phi_{\phi}(x_3 ) \nonumber \\ & & \times \bigg\ { \left[(1-x_1+x_2-x_3)\phi_k^a(x_2 ) -r_k x_2 \left ( \phi^p_k(x_2 ) + \phi^t_k(x_2 ) \right ) \right ] e'_{e5}(t^{(1)}_d ) h^{(1)}_d(x_1,x_2,x_3,b_1,b_1,b_3 ) \nonumber\\ & & \;\;\;\;\;\ ; + \left[(x_1-x_3)\phi_k^a(x_2 ) + r_k x_2 \left ( \phi_k^p(x_2)-\phi_k^t(x_2 ) \right ) \right ] e'_{e5}(t^{(2)}_d ) h^{(2)}_d(x_1,x_2,x_3,b_1,b_1,b_3 ) \bigg\}\ ; , \\ { \cal m}^p_{e6}&= & 16\pi c_f \frac{\sqrt{2n_c}}{n_c}m_b^4 r_{\phi } \int_0 ^ 1 [ dx]\int_0^{\infty } b_1 db_1 b_3 db_3\phi_b(x_1,b_1 ) \nonumber \\ & & \times \bigg\ { \left [ ( 1-x_1-x_3)\phi_k^a(x_2 ) \left ( \phi_{\phi}^t(x_3 ) + \phi_{\phi}^s(x_3 ) \right ) \right.\nonumber \\ & & \;\;\;\;\;\;\left .
+ r_k(1-x_1-x_3)\left ( \phi_k^p(x_2)-\phi_k^t(x_2 ) \right ) \left ( \phi_{\phi}^t(x_3 ) + \phi_{\phi}^s(x_3 ) \right ) \right.\nonumber \\ & & \;\;\;\;\;\;\left . -r_k x_2 \left ( \phi_k^p(x_2)+\phi_k^t(x_2 ) \right ) \left ( \phi_{\phi}^t(x_3 ) - \phi_{\phi}^s(x_3 ) \right ) \right ]
e'_{e6}(t^{(1)}_d ) h^{(1)}_d(x_1,x_2,x_3,b_1,b_1,b_3 ) \nonumber \\ & & - \left [ ( x_1-x_3)\phi_k^a(x_2 ) \left ( \phi_{\phi}^t(x_3 ) - \phi_{\phi}^s(x_3 ) \right ) + r_k(x_1-x_3 ) \left ( \phi_k^p(x_2)-\phi_k^t(x_2 ) \right ) \left ( \phi_{\phi}^t(x_3 ) - \phi_{\phi}^s(x_3 )
\right ) \right.\nonumber \\ & & \;\;\;\;\;\;\left .
+ r_k x_2 \left ( \phi_k^p(x_2)+\phi_k^t(x_2 ) \right ) \left ( \phi_{\phi}^t(x_3 ) + \phi_{\phi}^s(x_3 ) \right ) \right ]
e'_{e6}(t^{(2)}_d ) h^{(2)}_d(x_1,x_2,x_3,b_1,b_1,b_3 ) \bigg\}\;.\end{aligned}\ ] ] @xmath108 is added to @xmath109 term , because @xmath110 is the same as @xmath109 without the wilson coefficient . @xmath111\int_0^{\infty } b_1 db_1 b_3 db_3\phi_b(x_1,b_1 ) \nonumber \\ & & \times \bigg\ { \left [ ( 1-x_2)\phi_k^a(x_2)\phi_{\phi}(x_3 ) + r_kr_\phi \left ( ( 1-x_2)\left ( \phi_k^p(x_2)-\phi_k^t(x_2 ) \right ) \left ( \phi_{\phi}^t(x_3 ) + \phi_{\phi}^s(x_3 ) \right ) \right.\right.\nonumber \\ & & \;\;\;\;\;\;\left.\left .
+ ( x_1-x_3)\left ( \phi_k^p(x_2)+\phi_k^t(x_2 ) \right ) \left ( \phi_{\phi}^t(x_3 ) - \phi_{\phi}^s(x_3 ) \right ) \right ) \right ] e'_{a4}(t^{(1)}_f ) h^{(1)}_f(x_1,x_2,x_3,b_1,b_3,b_3 ) \nonumber \\ & & - \left [ ( x_1+x_3)\phi_k^a(x_2)\phi_{\phi}(x_3 ) \right.\nonumber \\ & & \;\;\;\;\;\;\left .
+ r_kr_\phi \left ( 4\phi_k^p(x_2)\phi_{\phi}^s(x_3)-(1-x_1-x_3 ) \left ( \phi_k^p(x_2)-\phi_k^t(x_2 )
\right ) \left ( \phi_{\phi}^t(x_3 ) + \phi_{\phi}^s(x_3 ) \right ) \right.\right.\nonumber \\ & & \;\;\;\;\;\;\left.\left .
+ x_2\left ( \phi_k^p(x_2)+\phi_k^t(x_2 ) \right ) \left ( \phi_{\phi}^t(x_3 ) - \phi_{\phi}^s(x_3 ) \right ) \right ) \right ] e'_{a4}(t^{(2)}_f ) h^{(2)}_f(x_1,x_2,x_3,b_1,b_3,b_3 ) \bigg\}\;,\\ { \cal m}^p_{a6}&= & 16\pi c_f \frac{\sqrt{2n_c}}{n_c}m_b^4\int_0 ^ 1 [ dx]\int_0^{\infty } b_1 db_1 b_3 db_3\phi_b(x_1,b_1 ) \nonumber \\ & & \times \bigg\ { \big [ -r_k ( 1-x_2 ) \left ( \phi^p_k(x_2 ) - \phi^t_k(x_2 ) \right ) \phi_{\phi}(x_3 ) \nonumber\\ & & \;\;\;\;\;\ ; -r_{\phi } ( x_1-x_3 ) \phi^a_k(x_2 ) \left ( \phi_{\phi}^t(x_3 ) + \phi_{\phi}^s(x_3 ) \right ) \big ] e'_{a6}(t^{(1)}_f ) h^{(1)}_f(x_1,x_2,x_3,b_1,b_3,b_3 ) \nonumber \\ & & + \big [ -r_k ( 1+x_2 ) \left ( \phi^p_k(x_2 ) - \phi^t_k(x_2 ) \right ) \phi_{\phi}(x_3 ) \nonumber\\ & & \;\;\;\;\;\ ; -r_{\phi } ( -2+x_1+x_3 ) \phi^a_k(x_2 ) \left ( \phi_{\phi}^t(x_3 ) + \phi_{\phi}^s(x_3 ) \right ) \big ] e'_{a6}(t^{(2)}_f ) h^{(2)}_f(x_1,x_2,x_3,b_1,b_3,b_3 ) \bigg\}\;.\end{aligned}\ ] ] the tree amplitude @xmath112 is the same as @xmath113 , but the wilson coefficient is different .
the evolution factors are defined by @xmath114\ ; , \;\ ; e'_{ai}(t)=\alpha_s(t)a'_{ai}(t)\exp[-s(t)|_{b_2=b_3}]\;.\end{aligned}\ ] ] the hard scales @xmath39 are defined by @xmath115 the wilson coefficients are given by @xmath116 the hard functions are given by @xmath117 \nonumber \\ & & \hspace{10 mm } \times k_{0}(d_{j}m_bb_{3 } ) \;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\ ; \mbox{for $ d^2_{j } \geq 0 $ } \nonumber \\ & & \hspace{10 mm } \times \frac{i\pi}{2 } h_{0}^{(1 ) } ( \sqrt{|d_{j}^2|}m_bb_{3})\;\;\;\ ; \mbox{for $ d^2_{j } \leq 0$}\ ; , \\ h^{(j)}_f&= & \frac{i\pi}{2 } \left[\theta(b_1-b_3)h_0^{(1)}\left(fm_b b_1\right)j_0\left(fm_bb_3\right ) + \theta(b_3-b_1)h_0^{(1)}\left(fm_b b_3\right ) j_0\left(fm_b b_1\right)\right ] \nonumber \\ & & \hspace{10 mm } \times k_{0}(f_{j}m_bb_{1})\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\ ; \mbox{for $ f^2_{j } \geq 0 $ } \nonumber\\ & & \hspace{10 mm } \times \frac{i\pi}{2 } h_{0}^{(1 ) } ( \sqrt{|f_{j}^2|}m_bb_{1 } ) \;\;\;\ ; \mbox{for $ f^2_{j } \leq 0$}\;,\end{aligned}\ ] ] with the variables @xmath118
we summarize the parameters which is used in the numerical analysis of our calculation .
we use the model of the @xmath8 meson wave function written as @xmath119 \;,\end{aligned}\ ] ] where @xmath120 gev@xcite .
@xmath121 is determined from normalization condition eq .
( [ eq : bnor ] ) .
the @xmath54 meson wave functions are given as @xmath122 \ ; , \\
\phi^p_{k}(x ) & = & \frac{f_k}{2\sqrt{2n_c } } \bigg [ 1 + \frac{1}{2}\left(30\eta_3 -\frac{5}{2}\ , \rho_k^2\right ) \left\{3(1 - 2x)^2 - 1\right\ } \nonumber\\ & & -\frac{1}{8}\left ( 3\eta_3\omega_3+\frac{27}{20}\ , \rho_k^2 + \frac{81}{10}\ , \rho_k^2 a_2\right ) \left\{3 - 30(1 - 2x)^2 + 35(1 - 2x)^4\right\ } \bigg]\ ; , \\
\phi^t_{k}(x ) & = & \frac{f_k}{2\sqrt{2n_c}}(1 - 2x ) \bigg [ 1 + 6\left(5\eta_3 -\frac{1}{2}\,\eta_3\omega_3 - \frac{7}{20}\ , \rho_k^2 - \frac{3}{5}\,\rho_k^2 a_2 \right ) ( 1 - 10x+10x^2 ) \bigg]\;.\end{aligned}\ ] ] where @xmath123@xcite@xcite .
the parameters of these wave functions are given as @xmath124 where renormalization scale is 1 gev .
we fix the parameters as above values .
the @xmath56 meson wave functions are given as @xmath125\ ; , \\
\phi^s_{\phi}(x ) & = & \frac{f^t_{\phi}}{4\sqrt{2n_c } } \bigg [ ( 1 - 2x)\left ( 6 + 9\delta_{+}+140\zeta_3^t-1400\zeta_3^tx+1400\zeta_3^tx^2 \right)+3\delta_{+}\log\frac{x}{1-x } \bigg]\;,\end{aligned}\ ] ] where @xmath126 @xcite .
we fix the parameters as above values , since the numerical results is insensitive to these parameters .
we use the wolfenstein parameters for the ckm matrix elements @xmath127 @xcite , and choose the angle @xmath128 @xcite .
in addition , we use the following parameters .
@xmath129 , + @xmath130 , + @xmath131 @xmath132 .
+ we show the numerical results of each amplitude for the @xmath133 and @xmath134 decays in table [ table1 ] . the factorizable penguin amplitude @xmath135 is dominant contribution to the @xmath136 decays .
the factorizable annihilation penguin amplitude @xmath137 generates large strong phase . in the @xmath134 modes ,
there are the contribution from @xmath138 and @xmath139 .
these tree amplitudes contribute a few percent to the whole amplitude , since the ckm matrix elements related to the tree amplitudes are very small .
we predict the branching ratios for the @xmath136 decays as @xmath140 the current experimental values are summarized in table [ table2 ] .
the values which are predicted in pqcd are consistent with the current experimental data . however , our branching ratios have the theoretical error from the @xmath141 corrections , the higher twist corrections , and the error of input parameters .
large uncertainties come from the meson decay constants , the shape parameter @xmath142 , and @xmath71 .
these parameters are fixed from the other modes(@xmath143 , etc . ) .
we try to vary @xmath142 from 0.36 gev to 0.44 gev , then we obtain br(@xmath144 .
next , we set @xmath145 and try to vary @xmath71 from 1.40 gev to 1.80 gev , then we obtain br(@xmath146 . in the naive factorization assumption ,
the branching ratio is very sensitive to the effective number of colors @xmath147 . if we set @xmath147=3 , then the branching ratio is about @xmath148 where the scale of the wilson coefficient is taken to @xmath46 and @xmath149 is 0.38 from the bsw model .
our predicted values are larger than those obtains from the naive factorization assumption .
this is due to the enhancement of the wilson coefficient . in pqcd approach
, the scale of the wilson coefficients , which is equal to the hard scale @xmath39 , can reach lower values than @xmath46 in case of the naive factorization assumption , and the wilson coefficients are then enhanced@xcite .
therefore , our branching ratios are enhanced over those obtained from the naive factorization assumption . in the qcd - improved factorization ,
the branching ratios for the @xmath150 decays are predicted as br@xmath151 and br@xmath152 with annihilation effect@xcite .
our predicted values are larger than these values .
we also compute the direct cp asymmetry defined as @xmath153 if the amplitude of the @xmath134 decay can be written as @xmath154 where the indices @xmath4 and @xmath3 denote the tree and penguin respectively , @xmath155 and @xmath156 are the strong phase , and the amplitude @xmath157 and @xmath158 are real , then the direct cp asymmetry is given by @xmath159 our results of the direct cp asymmetry is about few percent .
the reason why the direct cp asymmetry is very small is that the ckm matrix element for the tree amplitude is much smaller than for the penguin amplitude .
in this paper , we calculate the @xmath83 and @xmath160 decays in pqcd approach .
our predicted branching ratios agree with the current experimental data and are larger than the values obtained by the naive factorization assumption and the qcd - improved factorization .
this is because the wilson coefficients related to penguin operators are enhanced dynamically in pqcd . in the pqcd approach
, we can calculate the strong phase from nonfactorizable amplitudes and annihilation amplitudes explicitly .
we can then calculate the direct cp asymmetry of this modes , however this asymmetry is very small since the ckm matrix element for the tree amplitude is much smaller than for the penguin amplitude .
* note added : * after this work has been completed , we become aware of a similar calculation by chen , keum and li@xcite .
our results are in agreement .
* acknowledgment * the topic of this research was suggested by professor a.i .
the author thanks our pqcd group members : y.y .
keum , e. kou , t. kurimoto , h - n .
li , t. morozumi , r. sinha , k. ukai for useful discussions .
the author thanks jsps for partial support .
the work was supported in part by grant - in aid for special project research ( physics of cp violation ) ; grant - in aid for scientific research from the ministry of education , science and culture of japan .
the wilson coefficients for the @xmath161 gev are @xmath162
we consider the @xmath8 meson wave function @xmath163 where the indices @xmath69 and @xmath70 are the spin indices , @xmath67 and @xmath68 are the color indices .
this wave function can be decomposed into some components with different spin structure .
the @xmath8 meson wave function is then decomposed to be @xmath164_{\alpha\beta}\;,\end{aligned}\ ] ] where @xmath165 is color s degree of freedom .
the axial - vector wave function @xmath166 is defined by @xmath167 and the pseudo - scalar wave function @xmath168 is defined by @xmath169 using hqet , we can denote @xmath170 . the @xmath8 meson wave function
is then written by @xmath171_{\alpha\beta}\;.\end{aligned}\ ] ] the normalization of this wave function is given by @xmath172 where @xmath173 is the conjugate space of @xmath174 , and @xmath175 is the decay constant of the @xmath8 meson .
the @xmath54 meson wave function is defined that @xmath176 , @xmath177 is the coordinate of the @xmath178 quark .
@xmath179 can be decomposed into some components with different spin structure @xmath180 , @xmath181 , @xmath182 , @xmath183 .
these matrix elements are calculated using the twist expansion in the light - cone qcd sum rules .
the twist expansion is the expansion in terms of @xmath71 where @xmath72 .
the leading twist terms are independent on @xmath71 , the twist-3 and twist-4 terms are proportional to @xmath71 and @xmath184 respectively . up to twist-3
, the matrix elements are given as@xcite@xcite @xmath185 where @xmath186 and @xmath187 .
@xmath188 is the leading twist wave function , @xmath189 and @xmath190 are the twist-3 wave functions .
these wave functions are normalized as @xmath191 we define the following new wave functions , @xmath192 the @xmath54 meson wave functions are then written as @xmath193_{\alpha\beta}\;.\end{aligned}\ ] ]
the @xmath56 meson wave function has much terms than the @xmath8 meson and @xmath54 meson wave functions , since the @xmath56 meson is a vector meson .
the wave function @xmath194 can be decomposed into some components with different spin structure @xmath195 , @xmath196 , @xmath197 , @xmath198 , @xmath183 .
the components are calculated using the twist expansion in the light - cone qcd sum rules . in this case , the twist expansion is the expansion in terms of @xmath199 .
up to twist-3 , the components are given as@xcite @xmath200 \nonumber\\ & = & f_{\phi } m_{\phi } \int_0 ^ 1 dx_3 e^{-ix_3p_3\cdot z } \left [ \epsilon_{\phi\mu}\phi_{//}(x_3 ) + \epsilon^t_{\phi\mu}g^{(v)}_{\perp}(x_3 ) \right]\ ; , \\
\langle 0|{\bar s}(0)\gamma_\mu\gamma_5 s(z)|\phi(p_3)\rangle & = & -\frac{1}{4}\left(f_\phi - f^t_\phi\frac{2m_s}{m_\phi}\right ) m_\phi\epsilon_\mu^{\;\;\nu\alpha\beta}\epsilon^t_{\phi\nu } p_{3\alpha}z_\beta \int_0 ^ 1 dx_3 e^{-ix_3p_3\cdot z}g^{(a)}_{\perp}(x_3 ) \nonumber\\ & = & -\frac{1}{4}f_\phi m_\phi\epsilon_\mu^{\;\;\nu\alpha\beta}\epsilon^t_{\phi\nu } p_{3\alpha}z_\beta \int_0 ^ 1 dx_3 e^{-ix_3p_3\cdot z}g^{(a)}_{\perp}(x_3)\ ; , \\ \langle 0|{\bar s}(0)\sigma_{\mu\nu}s(z)|\phi(p_3)\rangle & = & i f^t_{\phi}\int_0 ^ 1 dx_3 e^{-ix_3p_3\cdot z}\bigg [ \left ( \epsilon^t_{\phi\mu}p_{3\nu}-\epsilon^t_{\phi\nu}p_{3\mu } \right ) \phi_{\perp}(x_3)\nonumber\\ & & \;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\ ; + ( p_{3\mu } z_\nu - p_{3\nu } z_\mu)\frac{\epsilon_\phi\cdot z } { ( p_3\cdot z)^2}m_\phi^2 h^{(t)}_{//}(x_3 ) \bigg]\nonumber\\ & = & i f^t_{\phi}\int_0 ^ 1 dx_3 e^{-ix_3p_3\cdot z}\bigg [ \left ( \epsilon^t_{\phi\mu}p_{3\nu}-\epsilon^t_{\phi\nu}p_{3\mu } \right ) \phi_{\perp}(x_3)\nonumber\\ & & \;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\ ; + ( \epsilon_{\phi\mu } p_{3\nu}-\epsilon_{\phi\nu } p_{3\mu } ) h^{(t)}_{//}(x_3 ) \bigg]\;,\\ \langle 0|{\bar s}(0)i s(z)|\phi(p_3)\rangle & = & \frac{i}{2 } \left(f^t_\phi - f_\phi\frac{2m_s}{m_\phi}\right ) \epsilon_\phi\cdot z m^2_{\phi } \int_0 ^ 1 dx_3 e^{-ix_3p_3\cdot z } h^{(s)}_{//}(x_3 ) \nonumber\\ & = & - f_\phi^tm_{\phi } \int_0 ^ 1 dx_3 e^{-ix_3p_3\cdot z } \frac{1}{2}\frac{d}{dx_3}h^{(s)}_{//}(x_3)\;,\end{aligned}\ ] ] where @xmath201,@xmath202 .
@xmath203 and @xmath204 are the leading twist wave functions , @xmath205 , @xmath206 , @xmath207 and @xmath208 are the twist-3 wave functions .
these wave functions satisfy the normalization conditions @xmath209 we can neglect the wave functions which are proportional to the transverse polarization vector @xmath77 , because these terms do not appear in our calculations .
we define the following new wave functions , @xmath210 the @xmath56 meson wave functions are then written as @xmath211_{\alpha\beta}\;.\end{aligned}\ ] ] it must be noted that this expression is dependent on how to define the sign of @xmath79 .
we considered the wave functions for an initial state . in this study , we need the @xmath54 meson and @xmath56 meson wave functions for a final state . the wave function for a final state is given as @xmath212 .
g. cavoto on behalf of the babar collaboration , hep - ex/0105018 .
a. bozek for the belle collaboration , talk presented at the international europhysics conference on high energy physics , budapest , hungary , july 12 - 18 , 2001 .
cleo collaboration , phys .
lett . * 86 * , 3718 ( 2001 ) .
for example , see i.i .
bigi and a.i .
sanda , _ cp violation _
( 2000 , cambridge ) .
m. bauer , b. stech , m. wirbel , z. phys .
c * 34 * , 103 ( 1987 ) ; z. phys .
c * 29 * , 637 ( 1985 ) .
keum , h - n .
li and a.i .
sanda , phys .
b * 504 * , 6(2001 ) ; phys .
d * 63 * , 054008 ( 2001 ) .
l , k. ukai , and m.z .
yang , phys .
d * 63 * , 074009 ( 2001 ) .
l and m.z .
yang , hep - ph/0011238
. c.h . chang and h - n .
li , phys . rev .
d * 63 * , 014003 ( 2001 ) .
e. kou and a.i .
sanda , hep - ph/0106159 .
ma and c.y .
wu , phys .
d * 63 * , 094004 ( 2001 ) .
cheng and k.c .
yang , hep - ph/0012152 .
m. kobayashi and t. maskawa , prog .
phys . * 49 * , 652 ( 1973 ) .
g. buchalla , a.j .
buras and m.e .
lautenbacher , review of modern physics , * 68 * , 1125 ( 1996 ) .
c.h . chang and h - n .
li , phys .
d * 55 * , 5577 ( 1997 ) .
t.w . yeh and h - n .
li , phys .
d * 56 * , 1615 ( 1997 ) .
collins and d.e .
soper , nucl .
b*193 * , 381 ( 1981 ) .
j. botts and g. sterman , nucl .
* b325 * , 62 ( 1989 ) .
h - n . li , phys .
d * 52 * , 3958 ( 1995 ) .
li , hep - ph/0102013 .
t. kurimoto , h - n .
li and a.i .
sanda , hep - ph/0105003 .
m. bauer and m. wirbel , z. phys .
c*42 * , 671 ( 1989 ) .
p. ball , jhep 09 , 005 ( 1998 ) .
p. ball , jhep 01 , 010 ( 1999 ) .
p. ball , v.m .
braun , y. koike , and k. tanaka , nucl .
b*529 * , 323 ( 1998 ) .
particle data group , eur .
j. c*15 * 1 ( 2000 ) .
chen , y.y .
keum and h - n .
li , hep - ph/0107165 . | we calculate the branching ratios and cp asymmetries of the @xmath0 decays using perturbative qcd approach , which includes @xmath1 and threshold resummations
. our results of branching ratios are consistent with the experimental data and larger than those obtained from the naive factorization assumption and the qcd - improved factorization approach .
= -10 mm = -6 mm = -6 mm # 1#2#3 |
in a real projective space @xmath0 , a pair of distinct points determines a unique projective line passing through them . however , there exist two line segments joining them because a real projective line is homeomorphic to a circle .
a natural question arises how we can introduce a notion of convex sets in @xmath0 .
the first definition probably dates back to steinitz @xcite and veblen - young @xcite according to danzer , grnbaum , and klee @xcite and deumlich - elster @xcite .
the following condition is equivalent to that of steinitz ( see subsection [ subsect : conv ] ) : a subset @xmath1 of @xmath0 is a _
( projective ) convex set _ if , for any two points of @xmath1 , exactly one of the two line segments joining them is contained in @xmath1 .
a convex set @xmath1 in @xmath0 is contained in some affine subspace of @xmath0 according to de groot and de vries ( * ? ? ?
* theorem 4 ) , which implies that @xmath1 is a convex set in the usual sense in this affine subspace .
however , for a collection of convex sets @xmath2 in @xmath0 , there may not exist any affine subspace of @xmath0 containing all of them simultaneously .
also , the intersection of these convex sets may split into multiple connected components each of which is a convex set .
hence , we call an intersection of a nonempty family of convex sets a _
multi - convex set_. our main interest concerns saturated multi - convex sets , which will be defined in section [ sect : satmultconv ] .
the notion of multi - convex sets appears in the following problem in computational geometry : given a collection of @xmath3 objects , say polygons , @xmath4 in @xmath5 , construct a representation of all lines avoiding every object @xmath6 . by extending @xmath5 to the projective plane @xmath0
, such a set of lines can be represented as a set of points in the dual projective plane @xmath7 through the well - known one - to - one correspondence between lines in @xmath0 and points in @xmath7 .
hence , it is useful to introduce the following function @xmath8 . for a subset @xmath9 of @xmath0
, we define @xmath10 where @xmath11 denotes the point in @xmath7 corresponding to the line @xmath12 in @xmath0 .
notably @xmath13 sends a convex set in @xmath0 to a convex set in @xmath7 , which we will prove in general ( see proposition [ prop : convtoconv ] ) .
thus , the problem above is transformed into the problem computing the intersection of @xmath3 convex sets @xmath14 in @xmath7 , which turns out to be a multi - convex set . moreover ,
this intersection is saturated as we will see in section [ sect : satmultconv ] .
this problem will be pursued in a later paper .
we remark that our problem is in contrast to problems concerning , for the case of @xmath5 , common line transversals , that is , lines meeting every object @xmath6 .
many authors have investigated these problems from computational or combinatorial aspects ( see @xcite , @xcite , @xcite , @xcite , @xcite , @xcite ) . in this paper
, we study fundamental properties of multi - convex sets and derive a duality for saturated multi - convex sets : there exists an order anti - isomorphism between nonempty saturated multi - convex sets in @xmath0 and those in @xmath7 .
moreover , we show that the following two antithetical notions are reversed through this duality : components and co - components of a saturated multi - convex set .
the notions and results studied in this paper will serve as useful tools in studying the problem presented above . in a complex projective space ,
the notion of linearly convex sets has been studied in @xcite and @xcite .
we shall show in remark [ rem : linconv ] that the notion of saturated multi - convex sets is almost identical to that of linearly convex sets defined in a real projective space .
in section [ sect : conv ] , we study the order structure of the family of convex sets in @xmath0 .
we focus on the notion of irreducible convex sets , and show that irreducible convex sets are the complements of projective hyperplanes .
in section [ sect : multconv ] , we study multi - convex sets and show two separation properties for multi - convex sets . in section [
sect : satmultconv ] , we study saturated multi - convex sets and derive the duality for saturated multi - convex sets . in this paper , we only consider finite - dimensional real projective spaces .
let @xmath15 be an @xmath16-dimensional real vector space .
we denote by @xmath17 , @xmath0 for short , the @xmath18-dimensional real projective space associated with @xmath15 , and by @xmath19 the projection of @xmath20 onto @xmath17 , which sends each nonzero vector in @xmath15 to the linear subspace spanned by the vector .
the dual projective space of @xmath0 is denoted by @xmath7 .
let @xmath21 be distinct two points in @xmath0 , and let @xmath22 be nonzero vectors in @xmath15 such that @xmath23 and @xmath24 .
( projective ) line segments _ joining @xmath25 and @xmath26 are defined as follows ( see figure [ fig : segment ] ) : @xmath27 , the gray area shows the set of all vectors @xmath28 such that @xmath29 , @xmath30 , and both of them are not zero.,width=151 ] in an @xmath18-dimensional affine or projective space , an @xmath31-dimensional subspace is called a _
hyperplane_. in order to avoid confusing geometric objects in an affine space with those in a projective space , they are sometimes prefixed with the word `` affine '' or `` projective '' such as affine hyperplanes or projective hyperplanes .
let @xmath32 and @xmath33 be partially ordered sets .
a subset @xmath9 of @xmath34 is _ directed _ if it is nonempty and every finite subset of @xmath9 has an upper bound in @xmath9 .
a partially ordered set is a _ meet - semilattice _ if any two elements have a greatest lower bound , that is , an infimum .
a function @xmath35 is called to be _ order preserving _ or _ monotone _ if @xmath36 always implies @xmath37 , and _ order reversing _ or _ antitone _ if @xmath36 always implies @xmath38 .
a function @xmath35 is called a _
( order ) anti - isomorphism _ if it is a bijection and both @xmath39 and @xmath40 are antitone .
a pair @xmath41 of functions @xmath42 and @xmath43 is a _ galois connection _ between @xmath34 and @xmath44 if it satisfies the following two conditions : 1 . both @xmath45 and @xmath46
are monotone ; 2 . the relations @xmath47 and @xmath48 are equivalent for all pairs of elements @xmath49 .
[ theo : galois ] for every pair of order preserving functions between posets , @xmath50 and @xmath51 , the following conditions are equivalent : 1 .
@xmath41 is a galois connection ; 2 .
@xmath52 and @xmath53 , where @xmath54 and @xmath55 denote the identities on @xmath34 and on @xmath44 .
moreover , these conditions imply 1 .
@xmath56 and @xmath57 , 2 .
@xmath58 and @xmath59 are idempotent .
in this section , we study the order structure of the family of convex sets in @xmath0 .
we focus on the notion of irreducible convex sets , which is introduced in terms of order , and show that irreducible convex sets are the complements of projective hyperplanes .
this result leads to a one - to - one correspondence between irreducible convex sets in @xmath0 and points in @xmath7 .
in section [ sect : satmultconv ] , we will extend this correspondence to a duality for saturated multi - convex sets .
a subset @xmath9 of @xmath0 is _ semiconvex _ if any two points in @xmath9 can be joined by a line segment which is contained in @xmath9 .
a subset @xmath9 of @xmath0 is a _
( projective ) convex set _ if @xmath9 is semiconvex and there exists a projective hyperplane avoiding @xmath9 .
since the complement of a projective hyperplane is an affine subspace of @xmath0 , a projective convex set in @xmath0 turns out to be a convex set in the usual sense in some affine subspace of @xmath0 . the definition of projective convex sets is equivalently described as follows .
a subset @xmath9 of @xmath0 is a projective convex set if and only if , for any two points in @xmath9 , exactly one of the two line segments joining them is contained in @xmath9 .
according to this theorem , the notion of projective convex sets can be characterized by only the notion of line segments : we need not refer to hyperplanes in order to define convex sets .
we could adopt it as the definition of projective convex sets .
let @xmath60 be a set , and let @xmath61 be a family of subsets of @xmath60 .
the family @xmath61 is called an _ algebraic closure system _ on @xmath60 if it has the following three properties : 1 .
the set @xmath60 is in @xmath61 ; 2 .
the intersection of any nonempty subfamily of @xmath61 is in @xmath61 ; 3 .
the union of any directed subfamily of @xmath61 is in @xmath61 .
an algebraic closure system is referred to as an _ alignment _ , a _ topped algebraic @xmath62-structure _ , or a _ convexity _ by some authors ( see coppel @xcite , davey - priestley @xcite , and van de vel @xcite ) .
the family of convex sets in @xmath63 is a typical example of an algebraic closure system .
an algebraic closure system is known to provide various notions such as extreme point , independent set , basis and face .
we have to take into account that projective convex sets do not form an algebraic closure system . hence , it is worth considering what kind of set - system they form . [
exam : notalgsys ] the family of convex sets in @xmath0 is not an algebraic closure system on @xmath0 . 1 . clearly the whole space @xmath0 is not convex .
2 . for two distinct points in @xmath0 ,
the intersection of the two line segments joining them is the two - point set .
thus , the intersection of two convex sets need not be convex .
3 . in @xmath64 ,
the union of convex sets @xmath65 for @xmath66 covers the whole space @xmath67 , where @xmath68 is an open interval in @xmath69 .
thus , the union of a directed family of convex sets need not be convex .
the following proposition states that the two conditions ( c2 ) and ( c3 ) above can be satisfied if a collection @xmath70 of convex sets in @xmath0 is consistent , where @xmath70 is called to be _ consistent _ if there exists an affine subspace of @xmath0 containing all members of @xmath70 simultaneously .
[ prop : localgsys ] let @xmath61 be the family of all convex sets in @xmath0
. then we have the following two properties : 1 .
the intersection of any nonempty consistent subfamily of @xmath61 is in @xmath61 ; 2 .
the union of any consistent directed subfamily of @xmath61 is in @xmath61 .
one can say that the family of convex sets in @xmath0 is a local algebraic closure system on @xmath0 : the whole space @xmath0 is covered by open affine subspaces , and `` locally '' , that is , on each of these open affine subspaces , the convex sets form an algebraic closure system .
this observation leads to the following notion of local algebraic closure systems : let @xmath60 be a set , and let @xmath61 be a family of subsets of @xmath60 .
then @xmath61 is called a _ local algebraic closure system _ on @xmath60 if it satisfies the two conditions in proposition [ prop : localgsys ] , where a collection @xmath70 of @xmath61 is called to be _ consistent _ if there exists a member of @xmath61 containing all members of @xmath70 simultaneously .
clearly an algebraic closure system is a local algebraic closure system , however the converse is not true in general as example [ exam : notalgsys ] shows .
we remark that the hypothesis of consistency in proposition [ prop : localgsys](ii ) can be dropped for open convex sets : [ prop : directclosed ] the union of any directed family of open convex sets is an open convex set .
let @xmath70 be a directed family of open convex sets . for two points @xmath71 ,
let @xmath72 and @xmath73 be members of @xmath70 containing @xmath25 and @xmath26 , respectively .
since @xmath70 is directed , there exists a member @xmath15 of @xmath70 containing both @xmath72 and @xmath73 .
then @xmath15 contains one of the two line segments joining @xmath25 and @xmath26 .
this line segment is contained in @xmath74 . to prove the uniqueness , assume that @xmath74 contains both of the two line segments joining @xmath25 and @xmath26 .
then @xmath74 contains the whole line spanned by @xmath25 and @xmath26 . by compactness
there exists a member of @xmath70 containing this whole line .
however , this contradicts that every member of @xmath70 is convex .
we have proved the proposition .
the notion of irreducible convex sets we introduce now appears in studying multi - convex sets : when we have a multi - convex set @xmath1 , we want to associate it with a family @xmath61 of convex sets satisfying @xmath75 ; this family @xmath61 of @xmath1 is reduced to another one if some element @xmath76 in @xmath61 can separate into a pair of strictly greater convex sets @xmath77 and
@xmath78 satisfying @xmath79 ; if this reduction terminates , then we have such convex sets that can not be the intersection of any pair of strictly greater convex sets .
one can say that convex sets of this kind are irreducible .
a convex set @xmath80 is _ irreducible _ if , for any convex sets @xmath81 and @xmath1 , @xmath82 implies @xmath83 or @xmath84 . in a meet - semilattice
, an element @xmath85 is called to be _ irreducible _ if @xmath86 always implies @xmath87 or @xmath88 ( see @xcite and @xcite ) .
[ theo : irred ] for a subset @xmath1 of @xmath0 , the following conditions are equivalent : 1 .
@xmath1 is an irreducible convex set ; 2 .
@xmath1 is a maximal convex set ; 3 .
@xmath1 is the complement of a projective hyperplane .
that ( iii ) @xmath89 ( ii ) @xmath89 ( i ) is trivial .
we prove that ( i ) implies ( iii ) .
assume that @xmath1 is an irreducible convex set , and not the complement of any projective hyperplane .
let @xmath80 be a maximal affine subspace , which is the complement of a hyperplane , of @xmath0 which contains @xmath1 . since @xmath1 is a proper subset of @xmath80 , clearly the closure of @xmath1 relative to @xmath80 is not @xmath80 .
let @xmath25 be a point in @xmath80 which is not adherent to @xmath1 .
it is known that a closed affine convex set is the intersection of the closed half - spaces which contain it ( see ( * ? ? ? * theorem 11.5 ) ) .
thus we can easily derive that there exists an affine hyperplane in @xmath80 passing through @xmath25 and avoiding @xmath1 . by extending such an affine hyperplane to the projective hyperplane , we have another affine subspace @xmath90 of @xmath0 containing @xmath1 , which avoids @xmath25 by construction .
let us choose an arbitrary point @xmath91 in @xmath1 , and let @xmath92 be an arbitrary nonzero vector in @xmath90 emanating from @xmath91 to @xmath25 .
we define the following two cones in @xmath90 : @xmath93 trivially we have @xmath94 .
we prove @xmath95 .
any point @xmath96 in @xmath97 has the following two forms : @xmath98 and @xmath99 for @xmath100 , @xmath101 and @xmath102 .
hence , we have @xmath103 .
this implies that the point @xmath96 lies in the line segment joining @xmath104 and @xmath105 which is contained in @xmath1 .
thus we have @xmath106 .
since @xmath25 is not adherent to @xmath1 , both @xmath107 and @xmath108 are strictly greater than @xmath1 . however , these results contradict that @xmath1 is irreducible . therefore ,
if @xmath1 is irreducible , then it is the complement of a projective hyperplane .
[ rem : irred ] the irreducibility of convex sets can be reduced to the irreducibility of open convex sets , where an open convex set @xmath80 is called to be _
irreducible _ if , for any open convex set @xmath81 and @xmath1 , @xmath82 implies @xmath83 or @xmath84 .
clearly any irreducible convex set is an irreducible open convex set by definition .
the converse can be easily proved in a similar way to the proof of theorem [ theo : irred ] .
thus it follows that a subset of @xmath0 is an irreducible open convex set if and only if it is an irreducible convex set .
therefore , in order to prove that a given convex set @xmath9 is irreducible , it suffices to prove that @xmath9 is an irreducible open convex set . the projective duality between points and hyperplanes states that there exists a natural one - to - one correspondence between points in @xmath0 and hyperplanes in @xmath7 ( and equivalently between hyperplanes in @xmath0 and points in @xmath7 ) which reverses the incidence relation : a point @xmath25 in @xmath0 lies in a hyperplane @xmath109 in @xmath0 if and only if the point in @xmath7 corresponding to @xmath109 lies in the hyperplane in @xmath7 corresponding to @xmath25 . from this correspondence and theorem [ theo : irred ]
, we immediately obtain a one - to - one correspondence between points in @xmath0 and irreducible convex sets in @xmath7 ( and between irreducible convex sets in @xmath0 and points in @xmath7 ) . when there is no danger of confusion , we denote by @xmath110 and @xmath111 the irreducible convex set and the point in @xmath7 corresponding to a point @xmath25 and an irreducible convex set @xmath1 in @xmath0 , respectively .
similarly we denote by @xmath112 and @xmath113 the irreducible convex set and the point in @xmath0 corresponding to a point @xmath114 and an irreducible convex set @xmath115 in @xmath7 , respectively .
[ prop : projduality ] for a point @xmath25 in @xmath0 and for an irreducible convex set @xmath1 in @xmath0 , we have the following two properties : 1 .
@xmath116 and @xmath117 ; 2 .
the relations @xmath118 and @xmath119 are equivalent . this result will be extended to a duality for saturated multi - convex sets in section [ sect : satmultconv ] .
a subset @xmath9 of an affine space always has a least convex set containing @xmath9 , which is called the convex hull of @xmath9 , while a subset @xmath9 of @xmath0 need not : for example , distinct two points are connected by distinct two line segments .
however , `` locally '' , that is , on each convex set containing @xmath9 , we have the following notion of convex hull of @xmath9 .
let @xmath1 be a convex set in @xmath0 .
for a subset @xmath9 of @xmath1 , the _ convex hull _
@xmath120_{c}}$ ] of @xmath9 _ relative to _ @xmath1 is a least convex subset of @xmath1 containing @xmath9 : @xmath121_{c}}=\bigcap{\left\{n\ , |\ , \text{$n$ is a convex set satisfying $ s\subseteq n\subseteq c$}\right\}}.\ ] ] in particular , for two points @xmath25 and @xmath26 in a convex set @xmath1 , we denote by @xmath122_{c}}$ ] the line segment joining @xmath25 and @xmath26 which is contained in @xmath1 .
let @xmath1 be a convex set in @xmath0 . for subsets @xmath9 and @xmath123 of @xmath1 , we have the following four properties : 1 .
@xmath124_{c}}$ ] ; 2 .
@xmath125 implies @xmath120_{c}}\subseteq { \left [ t \right]_{c}}$ ] ; 3 .
@xmath126_{c } } \right]_{c}}={\left [ s \right]_{c}}$ ] ; 4 . @xmath120_{c}}=\bigcup{\left\{{\left [ f \right]_{c}}\ , |\ , \text{$f\subseteq s$ and
$ f$ is finite}\right\}}$ ] . by proposition [ prop : localgsys ]
, the family @xmath61 of all convex subsets of @xmath1 is an algebraic closure system on @xmath1 .
the first two properties ( i ) and ( ii ) are trivial .
\(iii ) it follows from ( c1 ) and ( c2 ) that @xmath120_{c}}$ ] is in @xmath61 .
hence we obtain @xmath126_{c } } \right]_{c}}={\left [ s \right]_{c}}$ ] .
\(iv ) let @xmath127 be the family of all convex sets @xmath128_{c}}$ ] such that @xmath129 is a finite subset of @xmath9 .
clearly we have @xmath120_{c}}\supseteq\bigcup\mathcal{t}$ ] .
since @xmath130 , we derive @xmath120_{c}}\subseteq { \left [ \bigcup\mathcal{t } \right]_{c}}$ ] from the property ( ii ) .
since @xmath127 is directed , its union is in @xmath61 by ( c3 ) , which implies @xmath131_{c}}=\bigcup\mathcal{t}$ ] .
thus we have @xmath120_{c}}=\bigcup\mathcal{t}$ ] .
we remark that the proposition above is derived from only the fact that convex sets in @xmath0 form a local algebraic closure system . in an affine space ,
the convex hull of an open set is open , and that of a compact set is compact , both of which are essentially due to carathodory s theorem ( see @xcite ) , although the convex hull of a closed set need not be closed : for example , the convex hull of @xmath9 is not closed when @xmath9 is the union of a closed half - space and a single point not on the half - space .
since @xmath0 is a compact hausdorff space , a closed set in @xmath0 is compact , and vice versa .
thus we have : [ prop : convhull ] let @xmath1 be a convex set in @xmath0 .
for a subset @xmath9 of @xmath1 , we have the following two properties : 1 .
if @xmath9 is closed in @xmath0 , then the convex hull of @xmath9 relative to @xmath1 is closed in @xmath0 ; 2 .
if @xmath9 is open in @xmath0 , then the convex hull of @xmath9 relative to @xmath1 is open in @xmath0 .
in this section , we introduce two antithetical notions : components and co - components of a multi - convex set .
we show two separation properties for multi - convex sets , which motivate the notion of saturated multi - convex sets in section [ sect : satmultconv ] .
a subset of @xmath0 is a _ multi - convex set _ if it is the intersection of a nonempty family of projective convex sets .
figure [ fig : convcomp - c ] shows an example of a disconnected multi - convex set in @xmath132 .
let @xmath1 and @xmath133 be two convex sets in @xmath132 such that there exists no convex set containing both of them as in figure [ fig : convcomp - a ] and [ fig : convcomp - b ] . by definition ,
there exist two hyperplanes which avoid @xmath1 and @xmath133 , respectively .
the intersection of @xmath1 and @xmath133 consists of two connected regions which are separated by two hyperplanes . [
prop : comp ] every connected component of a multi - convex set is a convex set .
let @xmath1 be a multi - convex set , and let @xmath134 be a family of convex sets such that @xmath135 .
for any two points @xmath136 , if there exist two members @xmath137 and @xmath138 such that @xmath122_{c_{i}}}\not={\left [ p , q \right]_{c_{j}}}$ ] , then @xmath25 and @xmath26 lie in different connected components of @xmath139 , hence @xmath25 and @xmath26 line in different connected components of @xmath1 .
it follows that if @xmath25 and @xmath26 lie in the same connected component @xmath76 of @xmath1 , all the segments @xmath122_{c_{i}}}\ ( i\in i)$ ] coincide .
this implies that the line segment is contained in @xmath76 , joining @xmath25 and @xmath26 .
clearly @xmath76 does not contain both of the two line segments joining @xmath25 and @xmath26 .
we have proved the proposition .
note that every connected component of a multi - convex set @xmath1 is a maximal convex subset of @xmath1 , and vice versa .
we introduce the following notion of components in terms of order rather than topology : this view allows to dualize the notion of components as we shall see later .
let @xmath1 be a multi - convex set .
we call every maximal convex subset of @xmath1 a _ component _ of @xmath1 and denote by @xmath140 the family of all components of @xmath1 .
we call the cardinality of @xmath140 the _ degree _ of @xmath1 . from proposition [ prop : comp ]
, we immediately derive the following result .
let @xmath1 be a multi - convex set .
then @xmath140 is a mutually disjoint family of convex sets such that @xmath141 .
we have to take into account that a disjoint union of convex sets need not be a multi - convex set ( see figure [ fig : conv-4deg ] ) .
let us consider three lines in general position in @xmath132 .
the complement of the union of these three lines is a multi - convex set of degree @xmath142 , however the disjoint union @xmath9 of all except for the @xmath142th component is not a multi - convex set . because any convex set containing @xmath9 is the complement of one of those three lines , and the least multi - convex set containing @xmath9 must contain the @xmath142th component .
let @xmath1 be a multi - convex set .
we call every minimal convex set containing @xmath1 a _ co - component _ of @xmath1 and denote by @xmath143 the family of all co - components of @xmath1 .
we call the cardinality of @xmath143 the _ co - degree _ of @xmath1
. figure [ fig : convcomp ] shows that @xmath1 and @xmath133 are the co - components of the multi - convex set @xmath144 .
a collection of convex sets is called to be _ mutually inconsistent _ if there is no convex set in @xmath0 containing any pair of convex sets in this collection .
[ prop : cocomp ] let @xmath1 be a multi - convex set . then @xmath143 is a mutually inconsistent family of convex sets such that @xmath145 . for two members
@xmath146 , if @xmath76 and @xmath147 are contained in some convex set @xmath44 , then the convex hull @xmath148_{m}}$ ] of @xmath1 relative to @xmath1 coincides with both @xmath76 and @xmath147 , hence we obtain @xmath149 .
therefore every pair of distinct members in @xmath143 is inconsistent .
let @xmath61 be a family of convex sets such that @xmath75 .
clearly we have @xmath150 .
conversely we have @xmath151 by the definition of co - components .
we have proved @xmath145 .
[ prop : separation1 ] let @xmath1 be a multi - convex set , and let @xmath25 be a point not in the topological interior @xmath152 of @xmath1
. then there exists an open convex set which contains @xmath152 and avoids @xmath25 .
the topology of @xmath0 has a countable base of open convex sets .
hence , there exists a countable family @xmath153 of open convex sets such that @xmath154 .
we introduce the following notation : @xmath155 denotes @xmath156 , and @xmath157}$ ] denotes the least multi - convex set containing @xmath155 , that is , the intersection of all convex sets containing @xmath155 .
we first prove that there exists a co - component of @xmath157}$ ] which avoids @xmath25 . since @xmath155 is open , every minimal convex set containing @xmath155
is open by proposition [ prop : convhull ] .
since @xmath155 consists of finitely many connected components , the number of all minimal convex sets containing @xmath155 is finite .
it follows that @xmath157}$ ] is an open multi - convex set of finite co - degree .
since @xmath157}$ ] is an open subset of @xmath1 , we obtain @xmath157}\subseteq{\mathop{\rm int}\nolimits c}$ ] .
since @xmath157}$ ] is the intersection of its co - components , there exists a co - components of @xmath157}$ ] which avoids @xmath25 .
we show that , for each @xmath158 , we can choose a co - component @xmath159 of @xmath157}$ ] avoiding @xmath25 in such a way that @xmath160 .
let us consider the following directed graph @xmath123 : the vertices are the pairs @xmath161 , where @xmath158 and @xmath76 is a co - component of @xmath157}$ ] avoiding @xmath25 ; there exists an edge from @xmath161 to @xmath162 if @xmath163 and @xmath164 .
then @xmath123 is a finitely branching , infinite , rooted tree . in deed ,
as we have proved above , for each @xmath165 , there exists a co - component @xmath76 of @xmath157}$ ] avoiding @xmath25 , hence @xmath161 is a vertex of @xmath123 . for each @xmath166 , the convex hull @xmath167_{n}}$ ] of @xmath137
relative to @xmath76 is a co - component of @xmath167}$ ] avoiding @xmath25 , hence @xmath168_{n } } \right\rangle}$ ] is also a vertex of @xmath123 .
thus we obtain the following decreasing sequence of vertices starting at @xmath161 : @xmath169_{n } } \right\rangle}\leftarrow\cdots\leftarrow { \left\langle 1 , { \left [ c_{1 } \right]_{n } } \right\rangle}.\ ] ] since the co - component @xmath170_{n}}$ ] of @xmath171 coincides with @xmath172 by definition , this sequence ends with the root of @xmath123 .
it follows that @xmath123 is a connected , infinite , rooted graph .
since @xmath157}$ ] is of finite co - degree for all @xmath158 , the graph @xmath123 is finitely branching . to prove that @xmath123 is a tree ,
assume that there exists a cycle @xmath173 in @xmath123 .
let @xmath161 be a vertex in @xmath173 whose index @xmath3 is maximum among those of all vertices in @xmath173 .
then we have distinct two vertices @xmath174 and @xmath175 in @xmath173 adjacent to @xmath161 such that @xmath176 .
since @xmath76 contains both @xmath44 and @xmath177 , we obtain @xmath178 , which is a contradiction .
we have proved that @xmath123 is a finitely branching , infinite , rooted tree .
knig s lemma states that if a finitely branching rooted tree is infinite , then there exists an infinitely long rooted path ( see @xcite ) .
thus we have one .
clearly such a path is increasing : @xmath179 the family of all the co - components @xmath180 is totally ordered and consists of open convex sets avoiding @xmath25 .
hence the union @xmath181 is an open convex set avoiding @xmath25 .
moreover , since @xmath182 for all @xmath183 , we obtain @xmath184 .
[ prop : separation2 ] let @xmath185 and @xmath34 be closed multi - convex sets
. then the following conditions are equivalent : 1 .
there exist a co - component of @xmath185 and a co - component of @xmath34 which are disjoint ; 2 . there exist disjoint open convex sets @xmath72 and @xmath15 such that @xmath186 and @xmath187 .
assume the condition ( i ) .
let @xmath188 be a disjoint pair of a co - component of @xmath185 and that of @xmath34 .
then there exist disjoint two open sets @xmath9 and @xmath123 such that @xmath189 and @xmath190 .
let @xmath80 and @xmath81 be two affine subspaces of @xmath0 containing @xmath76 and @xmath44 , respectively . without loss of generality , we can assume @xmath191 and @xmath192 .
let us denote by @xmath193 the vector space associating to @xmath80 . from the compactness of @xmath76
, we can derive that there exists an open convex subset @xmath194 of @xmath193 such that @xmath195 , where @xmath196 .
then the open convex set @xmath197 satisfies @xmath198 .
similarly , we have an open convex set @xmath15 in @xmath0 such that @xmath199 .
clearly @xmath72 and @xmath15 are disjoint .
conversely , assume the condition ( ii ) .
then @xmath200_{u}}$ ] and @xmath201_{v}}$ ] are co - components of @xmath185 and @xmath34 , respectively . since @xmath72 and @xmath15 are disjoint , @xmath200_{u}}$ ] and @xmath201_{v}}$ ] are disjoint .
in this section , we introduce the notion of saturated multi - convex sets , and show that a subset @xmath9 of @xmath0 is a saturated multi - convex set if and only if @xmath9 is the intersection of a nonempty family of irreducible convex sets containing @xmath9 .
we derive a duality for saturated multi - convex sets : there exists an order anti - isomorphism between nonempty saturated multi - convex sets in @xmath0 and those in @xmath7 .
moreover , we show that the notion of components and that of co - components are reversed through this duality .
we introduce a class of multi - convex sets with the following separation property : a multi - convex set @xmath9 of @xmath0 is _ saturated _ if , for each point @xmath25 not in @xmath9 , there exists an open convex set containing @xmath9 and avoiding @xmath25 .
note that a subset of @xmath0 is a saturated multi - convex set if and only if it is the intersection of a nonempty family of open convex sets . from the two separation properties in the previous section
, we can immediately deduce the following two corollaries .
the interior of a multi - convex set is a saturated multi - convex set .
note that the closure of a multi - convex set need not even be a multi - convex set : for example , the closure of the intersection of finitely many irreducible convex sets is the whole space .
an open or closed multi - convex set is saturated . in subsection
[ subsect : comp - cocomp ] , we will see examples of multi - convex sets which are not saturated ( see figure [ fig : antiexam - c ] and [ fig : antiexam - d ] ) . in subsection [ subsect : irred ] , we mentioned the reduction of families @xmath61 of convex sets satisfying @xmath75 for a multi - convex set @xmath1 .
we now want to know when a multi - convex set is the intersection of such a family of convex sets that can not be reduced any more , that is , the intersection of irreducible convex sets . in deed
, it is saturated one as we show now : [ lem : separation ] let @xmath1 be a saturated multi - convex set , and let @xmath202 be a point not in @xmath1 .
then there exists an irreducible convex set which contains @xmath1 and avoids @xmath25 .
let @xmath70 be the family of all open convex sets that contains @xmath1 and avoids @xmath25 . by proposition [ prop : directclosed ] , the union of any nonempty totally ordered subfamily of @xmath70 is an open convex set .
clearly this open convex set avoids @xmath25 , hence it belongs to @xmath70 .
by zorn s lemma , there exists a maximal member @xmath80 in @xmath70 .
we prove that @xmath80 is an irreducible convex set . by remark [ rem : irred ]
, it suffices to prove that @xmath80 is an irreducible open convex set .
assume the opposite .
then there exist strictly greater open convex sets @xmath81 and @xmath1 such that @xmath82 . by the maximality of @xmath80 , both @xmath81 and @xmath1 contain @xmath25 , hence @xmath80 also contains @xmath25 .
this is a contradiction .
we have proved the lemma .
[ theo : sat - irred ] a subset @xmath9 of @xmath0 is a saturated multi - convex set if and only if @xmath9 is the intersection of a nonempty family of irreducible convex sets containing @xmath9 . since an irreducible convex set is open , the sufficiency of the condition is satisfied .
the necessity immediately follows from lemma [ lem : separation ] .
recall that there exists a natural one - to - one correspondence between irreducible convex sets in @xmath0 and points in @xmath7 ( see proposition [ prop : projduality ] ) .
given a saturated multi - convex set @xmath1 in @xmath0 , this correspondence allows to transform the family of all irreducible convex sets containing @xmath1 into a set of points in @xmath7 .
we define the following function @xmath203 , @xmath204 where @xmath205 denotes the one - to - one correspondence between irreducible convex sets in @xmath0 and points in @xmath206 . in the same way
, we have a function from @xmath207 to @xmath208 and denote it by the same symbol @xmath13 when there is no danger of confusion . the following proposition states that @xmath13 is a galois connection between the poset @xmath208 ordered by the inclusion and the poset @xmath207 ordered by the reserve inclusion ( see subsection [ subsect : pre ] ) . 1
. the functions @xmath209 and @xmath210 are antitone .
2 . the relations @xmath211 and @xmath212 are equivalent for all pairs @xmath213 .
the part ( i ) is trivial .
the part ( ii ) is proved as follows : @xmath214 note that @xmath110 and @xmath215 denote the irreducible convex sets corresponding to @xmath25 and @xmath26 , respectively .
[ prop : infsup ] for a family @xmath216 of subsets of @xmath0 , we have the following two properties : 1 .
@xmath217 ; 2 .
@xmath218 .
\(i ) @xmath219 .
\(ii ) since @xmath13 is antitone , we obtain @xmath220 for all @xmath221 .
hence we obtain @xmath222 . in the part ( ii ) above , both sides of the relation do not coincide in general , however if the family of convex sets consists of the co - components of some multi - convex set , then both sides of it coincide .
[ prop : infsup2 ] for a multi - convex set @xmath1 in @xmath0 , we have @xmath223 by proposition[prop : infsup](ii ) , it suffices to prove @xmath224 .
let @xmath25 be a point in @xmath225 . recall that @xmath110 is an irreducible convex set in @xmath0 containing @xmath226 .
since @xmath1 is the intersection of its co - components , we obtain @xmath227 .
the convex hull @xmath148_{\delta(p)}}$ ] of @xmath1 relative to @xmath110 is a co - component of @xmath1 , which implies that the point @xmath25 belongs to @xmath228_{\delta(p)}})$ ] .
hence the point
@xmath25 belongs to @xmath229 .
let @xmath230 be an abbreviation for @xmath231 .
we call the set @xmath232 the _ saturation _ of a subset @xmath9 of @xmath0 ( or @xmath7 ) .
since @xmath13 establishes a galois connection , we immediately obtain the following result ( see theorem [ theo : galois ] ) .
[ prop : linhull ] for subsets @xmath9 and @xmath123 of @xmath0 , we have the following four properties : 1 .
@xmath233 ; 2 .
@xmath125 implies @xmath234 ; 3 .
@xmath235 ; 4 .
@xmath236 .
[ prop : sat - linhull ] for a subset @xmath9 of @xmath0 , the following conditions are equivalent : 1 .
@xmath9 is a saturated multi - convex set ; 2 .
@xmath9 is a proper subset of @xmath0 and @xmath237 .
we first prove @xmath238 : @xmath239 note that @xmath110 and @xmath215 denote the irreducible convex sets corresponding to @xmath25 and @xmath26 , respectively .
assume the condition ( i ) .
then @xmath9 is clearly a proper subset of @xmath0 . by theorem [ theo : sat - irred ]
, the set @xmath9 is the intersection of all irreducible convex sets containing @xmath9 . recall that @xmath240 is the set of points in @xmath7 corresponding to the family of all irreducible convex sets containing @xmath9 in @xmath0 .
thus we derive @xmath241 . conversely assume the condition ( ii ) .
then we obtain @xmath242 .
hence @xmath9 is the intersection of a nonempty family of irreducible convex sets . by theorem [ theo : sat - irred ]
, the set @xmath9 is a saturated multi - convex set .
note that we have the corresponding results in @xmath7 of the preceding four propositions in @xmath0 , although we omit the statements .
the function @xmath13 is an order anti - isomorphism between nonempty saturated multi - convex sets in @xmath0 and those in @xmath206 .
let @xmath9 be a nonempty saturated multi - convex set in @xmath0 .
we prove that @xmath240 is a nonempty saturated multi - convex set in @xmath7 . by proposition [ prop : linhull](iv ) , we have @xmath236 . clearly @xmath240 is a proper subset of @xmath7 . by the corresponding result in @xmath7 of proposition [ prop : sat - linhull ] , we deduce that @xmath240 is a saturated multi - convex set in @xmath7 .
clearly @xmath240 is nonempty .
similarly we can prove that @xmath210 sends a nonempty saturated multi - convex set in @xmath7 to a nonempty saturated multi - convex set in @xmath0 . by proposition
[ prop : sat - linhull ] , we have @xmath243 for all saturated multi - convex sets
@xmath9 in @xmath0 ( or @xmath7 ) .
this implies that @xmath13 is a bijection between nonempty saturated multi - convex sets in @xmath0 and those in @xmath7 .
[ rem : linconv ] in a complex projective space , the notion of linearly convex sets has been studied in @xcite and @xcite .
let us introduce the notion of linearly convex sets in a real projective space @xmath0 .
we first define the following function @xmath39 from @xmath208 to @xmath244 : @xmath245 where @xmath11 denotes the point in @xmath7 corresponding to the hyperplane @xmath12 in @xmath0 through the projective duality . in the same way
, we have a function from @xmath244 to @xmath208 and denote it by the same symbol @xmath39 .
then a subset @xmath9 of @xmath0 is _ linearly convex _ if @xmath246 . by theorem [ theo : irred ]
, the function @xmath39 turns out to coincide with @xmath13 , and by proposition [ prop : sat - linhull ] , a subset of @xmath0 is a saturated multi - convex set if and only if it is a linearly convex , proper subset of @xmath0 .
as expected , the function @xmath13 preserves the convexity , and furthermore @xmath13 interchanges the openness and the closedness of convex sets .
[ prop : convtoconv ] for a nonempty subset @xmath9 of @xmath0 , we have the following three properties : 1 .
if @xmath9 is a convex set , then @xmath240 is a convex set ; 2 .
if @xmath9 is an open convex set , then @xmath240 is a closed convex set ; 3 . if @xmath9 is a closed convex set , then @xmath240 is an open convex set .
let @xmath0 be the @xmath18-dimensional projective space associated with an @xmath16-dimensional vector space @xmath15 , and let @xmath7 be the dual projective space of @xmath0 .
let @xmath9 be a convex set in @xmath0 .
we can take an affine hyperplane @xmath80 in @xmath15 such that @xmath80 avoids the origin of @xmath15 and @xmath247 , where @xmath19 denotes the projection of @xmath20 onto @xmath0 ( see subsection [ subsect : pre ] ) .
then @xmath9 is identified with a convex set in the affine space @xmath80 .
\(i ) for distinct two points @xmath248 , let @xmath185 and @xmath34 be @xmath249-dimensional linear subspaces of @xmath250 such that @xmath251 and @xmath252 . note that the orthogonal complements @xmath253 and @xmath254 are distinct @xmath18-dimensional linear subspaces of @xmath15 .
the intersection of @xmath253 and @xmath254 is an @xmath31-dimensional linear subspace of @xmath15 , hence the set of all @xmath18-dimensional linear subspaces of @xmath15 containing @xmath255 forms a projective line in @xmath7 . among those @xmath18-dimensional linear subspaces of @xmath15 ,
the set of all @xmath18-dimensional linear subspaces of @xmath15 meeting @xmath9 is mapped to an interval in @xmath7 . since both of @xmath25 and @xmath26 lie in the complement of the interval , one of the two line segments joining @xmath25 and @xmath26
are contained in @xmath240 , but the other is not .
we have proved the property ( i ) .
\(ii ) let @xmath256 be the set of all orthogonal complements @xmath257 such that @xmath194 is an @xmath18-dimensional linear subspace of @xmath15 meeting @xmath9 .
then @xmath256 is a subset of @xmath7 .
since @xmath9 is open in the affine space @xmath80 , we can derive that @xmath256 is open .
the set @xmath256 is the complement of @xmath240 , hence @xmath240 is closed . by the property ( i )
, the set @xmath240 is convex .
\(iii ) note that @xmath240 coincides with the set of all orthogonal complements @xmath257 such that @xmath194 is an @xmath18-dimensional linear subspace of @xmath15 avoiding @xmath9 .
since @xmath9 is compact in the affine space @xmath80 , we can derive that @xmath240 is open . by the property ( i ) , the set @xmath240 is convex .
we call a convex set to be _ saturated _ if it is saturated as a multi - convex set .
[ cor : duality3 ] the function @xmath13 is an order anti - isomorphism between nonempty saturated convex sets in @xmath0 and those in @xmath206 which interchanges the openness and the closedness of convex sets .
moreover , @xmath13 interchanges the openness and the closedness of multi - convex sets .
[ prop : fliptop ] the function @xmath13 sends a nonempty open multi - convex set to a closed multi - convex set , and a nonempty closed multi - convex set to an open multi - convex set .
let @xmath1 be a nonempty open multi - convex set in @xmath0 .
we obtain @xmath258 from proposition [ prop : infsup](i ) .
since every component @xmath76 of @xmath1 is an open convex set , @xmath259 is a closed convex set .
thus @xmath260 is a closed multi - convex set .
let @xmath1 be a nonempty closed multi - convex set in @xmath0 .
we obtain @xmath261 from proposition [ prop : infsup2 ] .
then we can similarly prove that @xmath260 is an open multi - convex set in @xmath7 .
for a saturated multi - convex set @xmath1 in @xmath0 , let us consider the following two statements : 1 .
every component of @xmath1 is saturated ; 2 .
every co - component of @xmath1 is saturated .
the statement ( 1 ) is clearly true , however the statement ( 2 ) is not in general : for example , the multi - convex set illustrated in figure [ fig : antiexam - b ] is saturated , while its co - components illustrated in figure [ fig : antiexam - c ] and [ fig : antiexam - d ] are not saturated .
thus the function @xmath13 need not establish a bijection between @xmath143 and @xmath262 ( and similarly between @xmath140 and @xmath263 ) . however , if @xmath1 is open or closed , then every co - component of @xmath1 is open or closed , hence every co - component of @xmath1 is saturated .
thus we have : let @xmath1 be a nonempty open or closed multi - convex set in @xmath0 .
then @xmath13 is a bijection between @xmath140 and @xmath263 , and between @xmath143 and @xmath262 . clearly
every component of @xmath1 is saturated . by corollary
[ cor : duality3 ] , the function @xmath13 is a bijection between components of @xmath1 and minimal saturated convex sets containing @xmath260 . by proposition [ prop : fliptop ] ,
the set @xmath260 is an open or closed multi - convex set in @xmath7 .
thus every co - component of @xmath260 is saturated .
hence minimal saturated convex sets containing @xmath260 are the co - components of @xmath260 .
thus @xmath13 is a bijection between @xmath140 and @xmath263 .
similarly we can prove that @xmath13 is a bijection between @xmath143 and @xmath262 .
for a nonempty saturated multi - convex set @xmath1 , even if @xmath13 is not a bijection between @xmath140 and @xmath263 , the pair of the following functions derived from @xmath13 establishes a bijection between them .
let @xmath1 be a nonempty saturated multi - convex set in @xmath0 .
then the pair of the following functions is a bijection between @xmath140 and @xmath263 : @xmath264_{\phi(n)}};\\ & \cdot^{\triangleleft}\colon { { \rm cocmp}\left(\phi(c)\right)}\to{{\rm cmp}\left(c\right)},\ m\mapsto\phi({\left\llbracket m \right\rrbracket}).\end{aligned}\ ] ] we first prove @xmath265 for all components @xmath76 of @xmath1 . by corollary [ cor : duality3 ] ,
the set @xmath259 is a minimal saturated convex set containing @xmath260 .
clearly it is a saturation of the convex hull @xmath266_{\phi(n)}}$ ] of @xmath260 relative to @xmath259 , and we obtain @xmath267_{\phi(n ) } } \right\rrbracket}={\left\llbracket n^{\triangleright } \right\rrbracket}.\ ] ] since @xmath76 is saturated , we derive @xmath268 . all that remains is to prove @xmath269 for all co - components @xmath44 of @xmath260 .
we have @xmath270 .
thus we derive @xmath271_{{\left\llbracket m \right\rrbracket}}}$ ] . both @xmath266_{{\left\llbracket m \right\rrbracket}}}$ ] and @xmath44 are co - components of @xmath260 and
are contained in @xmath272 .
this implies @xmath266_{{\left\llbracket m \right\rrbracket}}}=m$ ] .
hence we have @xmath269 . in the same way
, we have a bijection between between @xmath143 and @xmath262 .
the author is grateful to professor hideki tsuiki and professor klaus keimel for their encouragement , advices , and comments during the preparation of this paper .
the author also thanks professor masahiro hachimori for providing some important references . 99 w.a .
coppel , _ foundations of convex geometry _ , australian mathematical society lecture series 12 ( cambridge university press , cambridge , 1998 ) .
l. danzer , b. grnbaum v. klee , ` helly s theorem and its relatives ' , convexity ( ed .
klee ) , proceedings of symposia in pure mathematics 7 ( american mathematical society , providence , ri , 1963 ) 101180 .
davey h.a .
priestley , _ introduction to lattices and order _ , ( cambridge university press , cambridge , 2002 ) .
r. deumlich k.h .
elster , ` konvexe mengen in projektiven rumen ' , _ optimization _ vol
. 8 no . 3 ( 1977 ) 343365
. j. eckhoff , helly , radon , and carathodory type theorems , _ handbook of convex geometry _ ,
vol . a ( eds p.m. gruber and j.m .
wills ; north - holland , amsterdam , 1993 ) 389448 .
h. edelsbrunner , _ algorithms in combinatorial geometry _
( springer - verlag berlin , heidelberg , 1987 ) . g. gierz k.h .
hofmann k. keimel j.d .
lawson m.w .
mislove d.s .
scott , _ continuous lattices and domains _ ( cambridge university press , cambridge , 2003 ) . j.e .
goodman r. pollack , ` foundations of a theory of convexity of affine grassmann manifolds ' , _ mathematika _ 42 ( 1995 ) 305328 .
j. de groot h. de vries , ` convex sets in projective space ' , _ compositio mathematica _ 13 ( 1958 ) 113118 .
b. grnbaum , ` common transversals for families of sets ' , _ j. london math .
_ 35 ( 1960 ) 408416 .
hansen , ` a study of knig s lemma ' , preprint , department of philosophy , uppsala university , 2002 , + http://www.filosofi.uu.se/personal/kajborgeha.htm .
a. martineau , ` sur la notion densemble fortement linellement convexe ' , _ an .
acad . brasil .
cinc _ 40 ( 1968 ) 427435 .
rockafellar , _ convex analysis _
( princeton university press , princeton , 1970 ) .
e. steinitz , ` bedingt konvergente reihen und konvexe systeme .
teil iii ' , _ j. reine angrew . math .
_ 146 ( 1916 ) 152 .
van de vel , _ theory of convex structure _ ( elsevier science publishers b.v . ,
netherlands , 1993 ) . o. veblen j.w .
young , _ projective geometry _
( ginn and company , boston , 1918 ) .
r. wenger , helly - type theorem and geometric transversals , _ handbook of discrete and computational geometry _ , ( eds j.e .
goodman and j. orourke ; crc press , florida , 1997 ) 6382 .
r. wenger , progress in geometric transversal theory , _ advances in discrete and computational geometry _
( 2001 ) 375393 .
znamenskii l.n .
znamenskaya , ` projective convexity in @xmath273 ' , _ siberian mathematical journal _ 38 ( 1997 ) 685698 . | we study intersections of projective convex sets in the sense of steinitz . in a projective space ,
an intersection of a nonempty family of convex sets splits into multiple connected components each of which is a convex set .
hence , such an intersection is called a multi - convex set .
we derive a duality for saturated multi - convex sets : there exists an order anti - isomorphism between nonempty saturated multi - convex sets in a real projective space and those in the dual projective space . in discrete geometry and computational geometry , these results allow to transform a given problem into a dual problem which sometimes is easier to solve .
this will be pursued in a later paper . |
the eliashberg equation is usually solved in the limit of an infinite electronic bandwidth , @xmath7 , and fermi energy , @xmath8 @xcite .
the @xmath9 limit may be justified in the conventional superconductors where the fermi energy is much larger than the characteristic phonon frequency , @xmath10 .
the superconducting pairing occurs mainly within a region of width @xmath10 around the fermi surface , and for @xmath11 , it makes no difference whether we take @xmath8 finite or infinite .
there are , however , superconducting materials where @xmath8 is comparable with @xmath10 : for fullerene superconductors , @xmath12 ev , @xmath13 ev , and @xmath14 @xcite .
it is , therefore , of great interest to study how the superconducting properties are modified as the bandwidth and fermi energy are reduced .
this is an important as well as difficult problem which concomitantly calls for a reinvestigation of the migdal theorem and coulomb pseudopotential @xmath15 @xcite .
the migdal theorem ensures that the phonon vertex corrections are smaller than those terms included in the theory of superconductivity by the factor of @xmath16 , hence may be neglected for conventional low temperature superconductors with @xmath11 .
the coulomb repulsion , @xmath17 , is reduced to @xmath18 \approx 0.1 \-- 0.2 $ ] , owing to the retardation effect . if @xmath14 , however , then the phonon vertex correction should be important , and the coulomb repulsion will not be reduced , @xmath19 , which will almost always kill superconductivity .
we will not , however , try to investigate these complications in this paper .
we will instead assume that superconductivity results within the framework of eliashberg theory , and will investigate how the strong - coupling eliashberg theory is modified due to a finite bandwidth and what the consequences are of the modification .
this problem was first treated by zheng and bennemann @xcite in the context of fullerene superconductors , who calculated the pressure dependence of the transition temperature , @xmath0 , for doped fullerenes within the strong - coupling eliashberg formalism .
their calculations with the finite bandwidth explicitly included agree well with the experimental observations . in the different context of a @xmath20-@xmath21 superconductivity ,
marsiglio @xcite investigated the dependence of @xmath0 on nonmagnetic and magnetic impurities including the effects of a finite bandwidth .
he found that the nonmagnetic impurities reduce @xmath0 for finite @xmath7 .
this is interesting in view of the recent debate over the abrikosov and gorkov ( ag ) theory of impure superconductors @xcite .
kim and overhauser @xcite pointed out that the ag theory , if evaluated literally within the @xmath20-@xmath22 weak - coupling bardeen - cooper - schrieffer ( bcs ) framework @xcite , implies a substantial reduction of @xmath0 by nonmagnetic impurities in an apparent contradiction to the anderson s theorem @xcite and to experimental observations on conventional low temperature superconductors @xcite .
it is now understood that the ag theory is in accord with the anderson s theorem if treated within the eliashberg theory taking fully into account the @xmath22 nature of a pairing interaction , whereas the criticism by kim and overhauser should be valid for a non - retarded pairing interaction @xcite .
the problem of retardedness is directly related with a cutoff in the frequency or momentum summation . by working with a more realistic finite bandwidth eliashberg theory , with cutoffs both in the frequency and momentum summations
, we will be able to understand the recent controversy on the ag theory more clearly as will be discussed later . in the ag theory of impure superconductors of infinite bandwidths ,
the thermodynamic properties are independent of the impurity scattering rates because there exists a scaling relation between pure and impure superconductivity @xcite .
this simple relation , however , breaks down if finite bandwidths are taken into consideration as will be discussed below .
therefore , the thermodynamic properties of dirty superconductors are no longer independent of impurity scattering rates when the finite bandwidths are explicitly taken into account .
two of the thermodynamic properties , transition temperature , @xmath0 , and nuclear spin - lattice relaxation rate , @xmath23 , are studied in this paper within the phonon - mediated eliashberg theory .
when an electronic bandwidth is finite , @xmath0 is suppressed as impurity scattering rate , @xmath24 , is increased .
the rate of @xmath0 suppression by impurities is determined by the electronic bandwidth and strength of electron - phonon coupling , and in the limit of @xmath9 , @xmath0 is independent of impurity scattering rate in agreement with the anderson s theorem .
the impurity suppression of @xmath0 is @xmath25 a consequence of a time reversal symmetry breaking , but follows dynamically from the modified eliashberg equation of a finite electronic bandwidth , as is reflected in the fact that nuclear spin - lattice relaxation rate @xmath23 is rather enhanced by the impurity scatterings . as the electronic bandwidth
is reduced , the available electronic states to and from which quasi - particles can be scattered are restricted , and the effective electron - phonon coupling constant , @xmath1 , consequently , is decreased .
the reduced @xmath1 implies decreased @xmath0 and enhanced nmr coherence peak in @xmath26 .
there are several factors that affect the nmr coherence peak of @xmath27 @xcite .
the suppression of nmr coherence peak may be attributed to ( a ) gap anisotropy / non @xmath28-wave pairing of superconducting phase @xcite , ( b ) strong - coupling phonon damping @xcite , ( c ) paramagnetic impurities in samples @xcite , and/or ( d ) strong coulomb interaction @xcite such as paramagnon / antiparamagnon effects @xcite .
the nonmagnetic impurity scatterings , on the other hand , have no influence on @xmath27 for conventional superconductors of infinite bandwidth @xcite , other than the smearing of gap anisotropy .
this is because ( a ) there exists a simple scaling relation between the gap and renormalization functions of pure and impure superconductors of infinite bandwidth , and ( b ) in the expression for @xmath27 , the numerator and denominator are of the same powers in renormalization function , as will be detailed below .
we found that @xmath0 is reduced and the nmr coherence peak is enhanced as @xmath24 is increased for a finite bandwidth , because the effective @xmath1 is reduced .
this interpretation is consistent with the previous study of paramagnetic transition metals by macdonald @xcite . in considering the effects of a finite bandwidth of the @xmath29 electrons on the mass enhancement ,
macdonald found the reduced bandwidths cause a reduction in the mass enhancement due to electron - paramagnon interaction .
the reduced @xmath30 means reduced @xmath1 in agreement with the present work .
this paper is organized as follows : in sec .
ii , we will present the eliashberg equation on imaginary frequencies for impure superconductors with finite bandwidths . within the formalism
, we will also discuss the anderson s theorem including the scaling relation between the pure and impure superconductivity .
we can solve the eliashberg equation either in the imaginary frequency to obtain the gap function , @xmath31 , and renormalization function , @xmath32 , or in the real frequency to obtain @xmath33 and @xmath34 .
it is , however , much easier to solve the eliashberg equation in the imaginary frequency .
we therefore carried out the @xmath0 calculations in the imaginary frequencies .
detailed numerical calculations of as well as qualitative discussions on @xmath0 will be presented in detail in sec .
a brief comment on the theory of @xmath0 of impure superconductivity will be made in view of recent debates on the topic . for calculating @xmath6 , we need @xmath33 and @xmath34 in real frequencies .
it is more efficient to solve the eliashberg equation in the imaginary frequency and perform analytic continuations to real frequency than to solve it in real frequency . using the iterative method for analytic continuations @xcite extended to finite bandwidths
, we calculate the nuclear spin - lattice relaxation rates @xmath6 as the bandwidths , electron - phonon couplings and impurity scattering rates are varied .
the results of these calculations will be presented in sec .
finally , we will summarize our results and give some concluding remarks in sec . v.
the electron - phonon interaction is local in space and retarded in time .
consequently , it s momentum dependence is weak and neglected in the isotropic eliashberg equation , but the frequency dependence is important and fully included .
the isotropic eliashberg equation in the imaginary frequency including the finite electronic bandwidth and impurity scatterings on an equal footing is written as @xcite @xmath35 where @xmath36 here , @xmath37 is the density of states at the fermi level , @xmath38 s the pauli matrices operating in the nambu space , @xmath39 , @xmath40 , where @xmath41 is the matsubara frequency given by @xmath42 , @xmath43 , and @xmath17 is the chemical potential , which should not be confused with the coulomb repulsion .
@xmath44 , @xmath45 and @xmath46 are , respectively , the renormalization function , gap function and energy shift , when analytically continued to real frequency .
a tilde on a variable denotes that it is renormalized by impurity scatterings . from eq .
( [ self ] ) , we obtain the following three coupled equations : @xmath47 + \frac{1}{2\pi\tau } \log \biggl [ \frac { \cos(\widetilde{\theta}_n^1 ) } { \cos(\widetilde{\theta}_n^2 ) } \biggr ] , \label{elia1}\end{aligned}\ ] ] where @xmath48 , ~ \widetilde{\theta}_n^2 = \tan^{-1 } \biggl [ \frac { -w/2 -\mu + \chi_n } { \sqrt { \widetilde{w}_n^2 + \widetilde{\phi}_n^2 } } \biggr ] .
\label{ang}\ ] ] this coupled equation should be solved simultaneously with the following constraint of number conservation which determines @xmath17 : @xmath49 e^{i \delta},\ ] ] where @xmath50 is the band filling factor , @xmath51 , @xmath52 the number of electrons , @xmath53 the number of available states including the spin degeneracy factor , and @xmath54 is a positive infinitesimal .
we took the coulomb pseudopotential @xmath55 in eq .
( [ elia1 ] ) for simplicity . for the half - filled case of @xmath56 , eq .
( [ elia1 ] ) is greatly simplified because @xmath57 and @xmath17 vanish identically .
we have @xmath58 where @xmath59 . \label{ang2}\ ] ] this is the eliashberg equation for the half - filled case including finite bandwidths and impurity scatterings . for discussions on the anderson s theorem , it is convenient to rearrange eq .
( [ elia2 ] ) , moving the last terms of the right hand side to the left , as follows : @xmath60 where @xmath61 . if we multiply both the numerator and denominator of the right hand side of eq .
( [ elia3 ] ) by @xmath62 and identify @xmath63 and @xmath64 , where untilded variables simply imply that they are for pure superconductors , we obtain @xmath65 note that the correct equation that describes the pure superconductors is obtained from eq .
( [ elia2 ] ) by taking the limit @xmath66 , which is just eq .
( [ eliap ] ) with @xmath67 replaced by @xmath68 . when the bandwidth is infinite , we have @xmath67 = @xmath68 = @xmath69 , and eq .
( [ eliap ] ) describes both pure ( @xmath70 and @xmath71 ) and impure ( @xmath72 and @xmath73 ) superconductivity .
they are related by the following scaling relation : @xmath74 this constitutes a proof of the anderson s theorem that the transition temperature and other transport properties remain unchanged under nonmagnetic impurity scatterings for infinite @xmath7 . when the bandwidth is finite , eq . ( [ eliap ] ) is not the correct equation describing pure superconductivity
consequently , the anderson s theorem does not hold in this case , and we expect that the thermodynamic properties , such as transition temperature @xmath0 , will change as the impurity scatterings are introduced
. we will turn to this topic in the following section .
before we solve the eliashberg equation of eq .
( [ elia2 ] ) , let us first try an approximate solution to understand what to expect from detailed numerical calculations . from eq .
( [ ang2 ] ) , @xmath75 .
we plug this into eq .
( [ eliap ] ) , and take @xmath76 so that @xmath77 , to obtain @xmath78 we take @xmath79 for @xmath80 and @xmath81 , and 0 otherwise , which is a standard weak - coupling approximation . then , taking the advantage of the scaling relation of eq .
( [ scale ] ) , we obtain @xmath82 the second equation of eq .
( [ appeqn ] ) , after carrying out the summation over matsubara frequency @xmath83 using @xmath84,\end{aligned}\ ] ] can be written as @xmath85,\ ] ] where @xmath86 is the fermi distribution function . then , @xmath87 , \nonumber \\ & \lambda_{eff } & = \frac{\lambda \left ( 1- 1/\pi \tau \epsilon_f
\right ) } { 1+\lambda \left ( 1- 1/\pi \tau \epsilon_f \right ) }
. \label{leff}\end{aligned}\ ] ] from this , we find @xmath88 where @xmath89 is the transition temperature for pure superconductor of a given bandwidth .
we note that @xmath90 is equal to the transition temperature of infinite bandwidth , @xmath91 , in the approximate treatment that replaces @xmath92 by @xmath93 .
more detailed numerical calculations , however , yield that @xmath94 as shown in fig
. 1 . eq .
( [ apptc ] ) clearly shows that impurities are pair - breaking when the bandwidth is finite in agreement with marsiglio @xcite .
@xmath0 is reduced not because the time reversal symmetry is broken as is the case with magnetic impurities , but because the effective electron - phonon coupling constant is decreased by impurity scatterings for finite bandwidth as can be seen from eq .
( [ leff ] ) . for @xmath9 , eq . ( [ apptc ] ) implies that @xmath0 is not changed as was discussed in the previous section .
it seems appropriate to comment here on the recent debate on the ag theory of impure superconductivity .
this , as already pointed out by radtke @xcite , stems from using non - retarded interaction of weak - coupling scheme @xcite . for this , let us rewrite eq .
( [ eliap ] ) with @xmath95 for infinite bandwidth , before the integral over @xmath96 .
@xmath97 then , we have , for @xmath98 @xmath99 if we proceed the same way leading to eq .
( [ apptc ] ) , we have @xmath100 for @xmath9 .
kim and overhauser @xcite , however , exchanged the range of summation between @xmath96 and @xmath101 , as did ag , and evaluated the resulting equation exactly .
this leads to a contradiction to the anderson s theorem and experimental observations , as was pointed out by them .
the origin of this contradiction is clear .
this procedure amounts to neglecting the retarded nature of electron - phonon interaction , because @xmath102 now becomes frequency independent , which implies an instantaneous interaction in time .
the impurity suppression of @xmath0 we discuss in this paper should be distinguished from the previous studies @xcite .
kim and overhauser found that the nonmagnetic impurities suppress @xmath0 for a non - retarded pairing interaction .
we point out that @xmath0 is suppressed by nonmagnetic impurities for a retarded pairing interaction also if the bandwidth is finite . on the other hand
, marsiglio found that nonmagnetic impurities suppress @xmath0 of finite bandwidth superconductors for non - phonon pairing interactions .
it is now understood that his results are due to both the finiteness of @xmath7 and non - phonon nature of pairing interactoins . for a pairing interaction not of the form of eq .
( [ elia2 ] ) , the proof of the anderson s theorem of sec .
ii does not go through .
therefore , nonmagnetic impurity scatterings can suppress the transition temperature .
the qualitative discussion above is well verified in the detailed numerical calculations . in fig . 1
, we show the transition temperature @xmath0 , calculated from eq .
( [ elia2 ] ) , as a function of impurity scattering rate @xmath24 for several bandwidths .
we took following bickers _ et al . _
@xcite @xmath103 & \text{for}~ |\omega-\omega_0| \leq \gamma_c , \\ 0 & \text{otherwise } , \end{array } \right .
\label{fomega}\end{aligned}\ ] ] with @xmath10 = 0.05 ev , @xmath104 = 0.015 ev .
@xmath105 is a normalization constant to make @xmath106 . using 200 matsubara frequencies , self - consistency
is reached within a few tens of iterations at a given temperature , except for temperatures close to @xmath0 .
the solid , dotted , dot - dashed , short - dashed , and long - dashed curves correspond , respectively , to @xmath7 = 0.1 , 0.5 , 1 , 5 , and 10 ev .
we considered half - filled cases for simplicity , so that the fermi energy @xmath107 .
1(a ) and ( b ) are for @xmath108 = 0.02 and 0.05 ev , respectively .
the long - dashed curves representing @xmath7 = 10 ev , are indistinguishable from infinite bandwidth curves . as we expected from the qualitative analysis above
, the impurity suppression of the transition temperature is more pronounced for narrower bandwidths . as the bandwidth becomes wider , however , the rate of @xmath0 suppression by impurity scatterings is smaller until @xmath109 , beyond which we are almost in the infinite bandwidth limit where @xmath0 is independent of the impurity scattering rate in accordance with the anderson s theorem .
we wish to consider how other transport properties of finite bandwidth superconductors , nmr relaxation rate for example , are altered in the following section .
if the reduction in @xmath0 is due to time reversal symmetry breaking , the nmr coherence peak below @xmath0 will be reduced .
if , on the other hand , it is due to reduction of the effective electron - phonon coupling as is the case we consider here , we expect that the nmr coherence peak will be enhanced .
this is because the strong - coupling effects cause phonon dampings and reduce the coherence peak as was discussed .
to calculate nuclear spin - lattice relaxation rate @xmath23 , we need to perform analytic continuation to obtain the gap and renormalization functions on real frequency , @xmath110 and @xmath111 , from those on imaginary frequency , @xmath112 and @xmath113 .
it was carried out via the iterative method by marsiglio , schossmann , and carbotte ( msc ) @xcite . the msc equation which relates the gap and renormalization functions on imaginary frequency with those on real frequency , extended to half - filled finite bandwidth impure superconductors ,
is given by @xmath114 \nonumber \\ & + & \frac{1}{\beta \omega } \sum_{n\geq 0 } \frac{2 \widetilde{\theta}_n ip_n } { \sqrt{p_n^2 + \widetilde{\delta}_n^2 } } \bigl [ \lambda(\omega -ip_n ) -\lambda(\omega + ip_n ) \bigr ] + \frac{i}{\pi \tau \omega } \frac{\widetilde{\theta}(\omega ) \omega } { \sqrt{\omega^2 -\widetilde{\delta}(\omega)^2 } } , \nonumber \\
\widetilde{\delta}(\omega ) & = & \frac{i}{\widetilde{z}(\omega ) } \int_{-\infty}^{\infty } d\omega ~ \alpha^2 f(\omega ) \frac{2 \widetilde{\theta}(\omega-\omega)~ \widetilde{\delta}(\omega -\omega ) } { \sqrt{(\omega-\omega)^2 -\widetilde{\delta}(\omega-\omega)^2 } } \bigl [ n(\omega ) + f(\omega-\omega ) \bigr ] \nonumber \\ & + & \frac{1}{\beta \widetilde{z}(\omega ) } \sum_{n\geq 0 } \frac{2 \widetilde{\theta}_n \widetilde{\delta}_n } { \sqrt{p_n^2 + \widetilde{\delta}_n^2 } } \bigl [ \lambda(\omega -ip_n ) + \lambda(\omega + ip_n ) \bigr ] + \frac{i}{\pi \tau \widetilde{z}(\omega ) } \frac{\widetilde{\theta}(\omega ) \widetilde{\delta}(\omega ) } { \sqrt{\omega^2 -\widetilde{\delta}(\omega)^2 } } , \label{msc}\end{aligned}\ ] ] where @xmath115 this equation is solved iteratively for @xmath111 and @xmath110 , taking @xmath116 and @xmath117 , solution to the eliashberg equation on matsubara frequencies of eq .
( [ elia2 ] ) , as an input , until self - consistency is reached .
we show in fig .
2 the gap function @xmath118 as a function of @xmath119 at @xmath4 = 0.001 ev .
we took the impurity scattering rate @xmath24 = 0 , and the phonon spectral function @xmath120 as given by eq .
( [ fomega ] ) with @xmath121 ev which corresponds to fig .
2(a ) is for an infinite bandwidth , and ( b ) is for bandwidth @xmath7 = 1 ev . the solid and dashed lines , respectively , stand for real and imaginary parts of the gap function . the obtained results for @xmath122 , where previous studies are available , are in good agreement with the published data @xcite .
the nuclear spin - lattice relaxation rate @xmath123 is given by @xcite @xmath124 , \label{t1}\end{aligned}\ ] ] where @xmath125 is a form factor related with the conduction electron wavefunctions , and @xmath126 is a spin - spin correlation function at frequency @xmath119 and momentum transfer @xmath127 .
the impurity scatterings are included in the self - energy of renormalized green s function .
for a finite bandwidth , it is easy to derive @xmath128 ^ 2 + \biggl [ re \biggl ( \frac{\widetilde{\delta}(\epsilon)~ \widetilde{\theta}(\epsilon ) } { \sqrt{\epsilon^2 -\widetilde{\delta}(\epsilon)^2 } } \biggr ) \biggr]^2 \biggr\}. \label{t1t - s } \\ & \rightarrow & \int_{\delta}^{\infty } d\epsilon \frac{\partial f}{\partial \epsilon}~ \biggl\ { \frac{\epsilon^2 + \delta^2}{\epsilon^2 -\delta^2 } \biggr\ } , \label{t1t - w}\end{aligned}\ ] ] where the @xmath129 and @xmath130 are obtained by solving eqs .
( [ elia2 ] ) and ( [ msc ] ) iteratively .
the standard strong - coupling expression for @xmath23 given , for example , by fibich @xcite can be obtained by putting @xmath131 for infinite bandwidth .
( [ t1t - w ] ) follows for infinite bandwidth weak - coupling limit .
before we present the detailed numerical calculations , let us first analyze the expression for @xmath26 of eq .
( [ t1t - s ] ) qualitatively . for @xmath132 , @xmath133
= 0 in eq .
( [ t1t - s ] ) , and @xmath134 = \left \ { \begin{array}{ll } i \tanh^{-1 } \biggl [ \frac{w}{2\epsilon \widetilde{z}(\epsilon ) } \biggr ] & { \rm for}~ \epsilon \geq \frac{w}{2 \widetilde{z } } , \\
\frac{\pi}{2 } + i \tanh^{-1 } \biggl [ \frac{2\epsilon \widetilde{z}(\epsilon)}{w } \biggr ] & { \rm for}~ \epsilon \leq \frac{w}{2 \widetilde{z}}. \end{array } \right .
\label{ang3}\end{aligned}\ ] ] taking @xmath135 for simplicity , we have @xmath136 = \pi/2 $ ] for @xmath137 , and 0 for @xmath138 . then
, @xmath139 , \nonumber \\ { \rm or,}~~ \frac{(t_1 t)_n}{(t_1 t)_s } & = & \frac { \tanh \bigl [ w/2 t \bigr ] } { \tanh \bigl [ w/2t_c \bigr]}.\end{aligned}\ ] ] @xmath140 , therefore , decreases as the temperature is increased , which is contrasted with the constant value of 1 for the infinite bandwidth case .
the decrease of @xmath140 as the temperature is increased above @xmath0 was observed in the @xmath141v nmr study of v@xmath142si superconductors by kishimoto _
this observation was interpreted in terms of narrow bandwidths in accord with the present work .
the range of integration from 0 to @xmath143 is just what we may expect intuitively .
the @xmath144 region , on the other hand , is difficult to analyze without detailed information on @xmath129 , because the height of nmr coherence peak is mainly determined by the magnitude of imaginary part of @xmath129 in the eliashberg formalism .
we may expect , however , that the coherence peak will be enhanced as @xmath7 is decreased and @xmath24 is increased , because @xmath1 is reduced . in fig .
3 , we show the normalized nmr relaxation rate by the normal state korringa value , @xmath140 , calculated from eqs .
( [ elia2 ] ) , ( [ msc ] ) and ( [ t1t - s ] ) as a function of @xmath145 , where @xmath91 is the critical temperature for infinite bandwidth case .
we took the phonon spectral function @xmath146 as given by eq .
( [ fomega ] ) with @xmath121 ev , which corresponds to fig .
we selected 3 sets of @xmath7 and @xmath24 values , labeled as a , b , and c in fig .
1(b ) , for @xmath23 calculations . @xmath7 and @xmath24 of a , b , and c are , respectively , in unit of ev , 10 and 0 , 0.1 and 0 , and 0.1 and 0.05 , as can be read from fig .
note that as one goes from a to b to c , the normalized nmr relaxation rates have progressively enhanced peaks , and @xmath0 is reduced accordingly .
these results are straightforward to interpret , as already explained before .
the computed values of @xmath1 for a , b , and c are 1.67 , 0.63 , and 0.60 , respectively . as @xmath1 is decreased , @xmath0 should be reduced and nmr coherence peak should be enhanced , because there is no time reversal symmetry breaking in the present problem .
the solid curve of a , having @xmath1 = 1.67 , has substantially reduced coherence peak , in agreement with the previous works @xcite .
note also that the normalized nmr relaxation rates for finite bandwidths decrease as @xmath4 is increased beyond @xmath0 as expected .
in this paper , we have investigated the effects of a finite bandwidth on the thermodynamic properties of impure superconductors within the framework of phonon - mediated eliashberg theory .
we found that the transition temperature and nmr coherence peak are suppressed and enhanced , respectively , by impurity scatterings when the finiteness of bandwidths is explicitly taken into consideration .
these results can be understood in terms of reduced effective electron - phonon coupling @xmath1 .
the motivation for this work was , in part , the observation that the phonon frequency and the fermi energy are comparable and a substantial disorder is present in the fullerene superconductors @xcite .
the nmr coherence peak in @xmath26 was found absent for doped fullerenes @xcite .
we wish to point out that the present theory is @xmath25 concerned with why the nmr coherence peak is absent for a given material .
the present theory shows that if the disorder is increased for a finite bandwidth superconductor , its transition temperature should be reduced and coherence peak should be enhanced , respectively , compared with those of a clean material . in view of our results
, it will be very interesting to systematically investigate how the transition temperature and nmr coherence peak behave as the degree of disorder is varied for doped c@xmath5 . in a@xmath142c@xmath5 ,
almost all other experiments than nmr relaxation rates seem to point to a phonon - mediated @xmath28-wave pairing @xcite . also , due to the orientational disorder @xcite , the fermi surface anisotropy is not strong enough to suppress the coherence peak @xcite .
the present study shows that a quite strong electron - phonon coupling of @xmath147 is needed to suppress the nmr coherence peak in agreement with nakamura @xmath148 @xcite , and allen and rainer @xcite .
the @xmath147 seems too large for doped fullerenes since the far infrared reflectivity measurements of degiorge _
@xcite show that @xmath149 , a classic weak - coupling value . because there are no magnetic impurities in a@xmath142c@xmath5 , the absence of nmr coherence peak is still to be understood .
_ @xcite suggested that the applied magnetic field is responsible for the suppressed nmr coherence peak .
their explanation , however , seems to be more a puzzle than an answer . according to the eliashberg theory
, the energy of applied magnetic field should be at least @xmath150 to suppress the coherence peak .
the magnetic field in their @xmath151c nmr experiment corresponds to @xmath152 .
such a small energy scale in the coherence peak suppression is really a puzzle .
we are currently investigating the strong coulomb interaction effects with a paramagnon approximation @xcite .
we suspect that the strong coulomb interaction may be responsible for suppressed nmr coherence peak in doped fullerenes .
j. r. schrieffer , _ theory of superconductivity _ , addison - wesley ( 1964 ) .
p. b. allen and b. mitrovic , in _ solid state physics _ vol .
37 , edited by f. seitz _ et al .
_ , academic press ( 1982 ) , p. 1 , and references therein .
w. e. pickett , phys .
b * 26 * , 1186 ( 1982 ) .
a. f. hebard , physics today * 45 * no .
11 , 26 ( 1992 ) , and references therein .
j. bardeen , l. n. cooper , and j. r. schrieffer , phys .
rev . * 108 * , 1175 ( 1957 ) .
h. zheng and k. h. bennemann , phys .
b * 46 * , 11993 ( 1992 )
. f. marsiglio , phys . rev .
b * 45 * , 956 ( 1992 ) .
a. a. abrikosov , l. p. gorkov , and i. e. dzyloshinski , _ methods of quantum field theory in statistical physics _ , dover ( 1965 ) , chap .
k. maki , in _ superconductivity _ , edited by r. d. parks , dekker ( 1967 ) , p. 1035 .
y. j. kim and a. w. overhauser , phys .
b * 47 * , 8025 ( 1993 ) ; _ ibid . _
* 49 * , 12339 ( 1994 ) ; a. a. abrikosov and l. p. gorkov , _ ibid .
_ * 49 * , 12337 ( 1994 ) .
p. w. anderson , j. phys . chem .
solid * 11 * , 26 ( 1959 ) .
r. j. radtke , preprint cond - mat/9306037 , and ph.d .
thesis , university of chicago , appendix c , ( 1994 ) .
d. fay and j. appel , phys .
b * 51 * , 15604 ( 1995 ) .
d. e. mclaughlin , in _ solid state physics _
31 , edited by h. ehrenreich _ et al .
_ , academic press ( 1976 ) , p. 1 , and references therein .
y. choi and e. j. mele , phys .
b * 52 * , 7549 ( 1995 )
. y. o. nakamura _
et al . _ , solid state commun . * 86 * , 627 ( 1993 ) . p. b. allen and d. rainer , nature * 349 * , 396 ( 1991 ) .
t. koyama and m. tachiki , phys .
b * 39 * , 2279 ( 1989 ) .
h. hasegawa , j. phys .
. jpn . * 42 * , 779 ( 1977 ) .
a. h. macdonald , phys .
b * 24 * , 1130 ( 1981 ) .
f. marsiglio , m. schossmann , and j. p. carbotte , phys .
b * 37 * , 4965 ( 1988 ) . n. e. bickers _ et al . _ ,
b * 42 * , 67 ( 1990 ) .
t. moriya , j. phys .
. jpn . * 18 * , 516 ( 1963 ) .
m. fibich , phys .
* 14 * , 561 ( 1965 ) ; _ ibid .
_ , 621 ( 1965 ) .
y. kishmoto , t. ohno , and t. tanashiro , j. phys .
* 64 * , 1275 ( 1995 ) .
r. tycko _ et al .
* 68 * , 1912 ( 1992 ) .
s. sasaki , a. matsuda , and c. w. chu , j. phys .
63 * , 1670 ( 1994 ) .
p. w. stephens _
_ , nature * 351 * , 632 ( 1991 ) .
e. j. mele , unpublished .
l. degiorge _
et al . _ ,
nature * 369 * , 541 ( 1994 ) .
v. a. stenger , c. h. pennington , d. r. buffinger , and r. p. ziebarth , phys .
lett * 74 * , 1649 ( 1995 ) . | we study the thermodynamic properties of impure superconductors by explicitly taking into consideration the finiteness of electronic bandwidths within the phonon - mediated eliashberg formalism . for a finite electronic bandwidth ,
the superconducting transition temperature , @xmath0 , is suppressed by nonmagnetic impurity scatterings .
this is a consequence of a reduction in the effective electron - phonon coupling , @xmath1 .
the reduced @xmath1 is reflected in the observation that the coherence peak in @xmath2 , where @xmath3 is the nuclear spin - lattice relaxation time and @xmath4 is the temperature , is enhanced by impurity scatterings for a finite bandwidth .
calculations are presented for @xmath0 and @xmath2 as bandwidths and impurity scattering rates are varied .
implications for doped c@xmath5 superconductors are discussed in connection with @xmath0 and @xmath6 measurements . |
in recent years , data mining algorithms have been used to aid the process of scientific classification .
_ data mining _ is the extraction of potentially useful information from data using machine learning , statistical , and visualization techniques .
pattern recognition algorithms ( or classifiers ) are data mining tools that search for clusters in complex , multi - dimensional spaces of _ attributes _ ( observed and/or measured properties ) .
these algorithms typically operate in one of two modes : supervised ( in which the classifier is trained with known classification instances ) and unsupervised ( in which classification occurs without training examples ) .
algorithms are designed to identify data patterns such as clustering and/or correlations , but their limitations must also be understood : it is up to the scientist to interpret physical mechanisms responsible for producing identified clusters .
clusters found by classifiers can represent source populations ; this happens when the class properties are produced by physical mechanisms pertaining to the sources .
clusters can also result from the way in which source properties are measured ; sampling biases , systemic instrumentation errors , and correlated properties can all force data to cluster and thus give the appearance of class structure when there is none .
data mining algorithms can be applied to gamma - ray burst classification .
two gamma - ray burst classes have been recognized for years
@xcite on the basis of duration and spectral hardness . class i ( long ) bursts are longer , spectrally softer , and have larger fluences than class ii ( short ) bursts .
recent classification schemes have used data collected by batse ( the burst and transient source experiment on nasa s compton gamma - ray observatory ; cgro ) @xcite because this large database ( 2704 bursts observed between 1991 and 2000 ) was collected by a single instrument with known instrumental characteristics .
three attributes of batse gamma - ray burst data have been identified as being significant ( using techniques such as principal component analysis ) in delineating gamma - ray burst classes @xcite : duration t90 ( the time interval during which 90% of a burst s emission is detected ) , fluence s ( time integrated flux in the 50 to 300 kev spectral range ) , and spectral hardness hr321 ( the 50 to 300 kev fluence divided by the 25 to 50 kev fluence ) . logarithmic measures of these values are typically used because classes are more clearly delineated when attributes are defined logarithmically .
historically , bursts with durations @xmath0 seconds have been typically considered to belong to class ii .
data mining techniques allow a third gamma - ray burst class to be identified in batse data .
three classes are preferably recovered instead of two by both statistical clustering techniques @xcite and unsupervised pattern recognition algorithms @xcite .
the third class forms at the boundary between class i and class ii , and primarily contains the softest and smallest fluence bursts from class i. since class ii appears to be relatively unchanged by the detection of the third class , the three classes are called class ii ( short , small fluence , hard bursts ) , class i ( long , large fluence bursts of intermediate hardness ) , and class iii ( intermediate duration , intermediate fluence , soft bursts ; also referred to as intermediate bursts )
. the boundaries between classes are fuzzy , as some bursts are not easily assigned to a specific class .
different data mining algorithms do not necessarily assign individual bursts to the same classes because each classifier operates under different assumptions concerning correlations between data attributes and how these relate to clustering criteria .
some classifiers are designed to work with nominal data while others are not ; some employ bayesian while others employ frequentist statistics ; some assume _ a priori _ distributions of attribute values while others do not .
the results of any classifier can change if the size and makeup of the data set is altered .
data errors can influence the results since few classifiers currently include measurement error information in their analyses .
however , irreproducibility is not necessarily a fault of machine learning methodology .
each classifier provides different insights into the way the data are structured . for
any given data set , there is a good possibility that some critical experiment or observation has not been performed , or that some key measurements have yet to be made , or that the relative importance of some attribute has been underestimated or overestimated .
_ there is no correct way of classifying a dataset because the usefulness of the classification depends on the insights gained from it by the user .
_ in a previous application of supervised classification @xcite to gamma - ray burst data we hypothesized that class iii does not necessarily represent a separate source population . instead ,
instrumental and sampling biases have been proposed as a way in which some class i bursts can take on class iii characteristics . due to a known correlation between hardness and intensity @xcite , small fluence class i bursts
are typically softer than bright class i bursts ; this is supported by principal component analysis @xcite .
since the correlation results from a shift to smaller average peak spectral energy @xmath1 at lower peak flux but not from changes in the average low - energy spectral index @xmath2 or the average high - energy spectral index @xmath3 @xcite , this correlation has been attributed to the softer bursts being generally at larger cosmological redshift .
( this conclusion may not necessarily be correct because a broad range of gamma - ray burst luminosities is suggested from redshifts of gamma - ray burst afterglows @xcite ; however , it should be noted that only a small afterglow sample is available . ) additionally , fluences and durations of some class i bursts can systematically be underestimated @xcite ; we refer to this as the _ fluence duration bias _ @xcite .
simply put , fluences and durations of some class i bursts ( particularly those with the smallest peak fluxes ) can be underestimated due to the unrecognizability of low signal - to - noise emission ; combined with their spectral softness , this gives them characteristics consistent with class iii .
unfortunately , the fluence duration bias has been difficult to quantify .
the amounts by which the fluence and duration of an individual burst are affected depend on the fitted background rates and estimated burst durations at all energies ; to remove the background properly assumes _ a priori _ knowledge of the burst s intrinsic temporal and spectral structure .
such _ a priori _ knowledge can only be acquired in the absence of background , and gamma - ray burst observations are inherently noisy .
very high signal - to - noise estimates of a burst s temporal and spectral structure can only be obtained for a small number of the bursts with the largest fluences .
these well - measured quantities are not entirely intrinsic ; it appears that even the brightest bursts require systematic correction because they are at large redshift ( @xmath4 ) .
our objective is to determine whether or not the fluence duration bias can account for the number of bursts with class iii characteristics . in order to do this ,
we determine the total number of bursts that exhibit class iii characteristics using several different unsupervised classifiers .
then , we statistically model the suspected bias and determine whether it is strong enough to produce the class iii bursts .
a number of pertinent questions will have to be addressed in pursuing this objective : do theoreticians need to develop models for one , two , or three gamma - ray burst classes ? how can data mining techniques be used to aid scientific classification ?
are systematic effects present in data collected by batse or other gamma - ray burst experiments that alter classification structures ? can these effects be understood ?
can information on intrinsic properties of the source population be extracted if these effects are present ? can future instruments be designed to minimize or eliminate these effects ?
we systematically compare the output of various unsupervised algorithms in conjunction with a homogeneous gamma - ray burst data set obtained with one set of instrumental settings .
we use the online gamma - ray burst toolshed @xcite ( shell for expeditions in data mining ) that we are developing to aid our analysis .
this toolshed ( currently ready for pre - beta testers at http://grb.mnsu.edu/grbts/ ) provides a suite of supervised and unsupervised data mining tools and a large database of preprocessed gamma - ray burst attributes .
it allows users to classify data using more than one algorithm in order to identify consistencies in the different classification techniques and thereby gain better insight into the heterogeneous nature of the data . in order to further minimize the effects of instrumental biases ,
we have limited our database to bursts detected with a homogeneous set of batse trigger criteria .
the database consists of bursts from the batse current burst catalog ( http://f64.nsstc.nasa.gov/batse/ grb / catalog / current/ ) .
bursts included are non - overwriting and non - overwritten bursts ( _ e.g. _ those whose batse readout periods did not overlap detectable bursts immediately preceding or following their detection ) triggering at least two batse detectors in the 50 to 300 kev energy range with the trigger threshold set @xmath5 above background on any of the three trigger timescales .
we require all classifiers to use only the three attributes of @xmath6(t90 ) , @xmath6(hr321 ) , and @xmath6(s ) .
we apply four unsupervised toolshed algorithms with different approaches to clustering .
these algorithms are esx , a kohonen neural network , the unsupervised em algorithm , and the unsupervised kmeans algorithm .
esx @xcite is a classifier that forms a three - level tree structure .
the root level of the tree contains summary information for all bursts .
the second ( concept ) level of the tree sub - divides the root level into clusters formed as a result of applying a similarity - based evaluation function .
the third tree level holds the individual bursts .
a kohonen neural network @xcite architecture is represented as a collection of input and output units . during training ,
the input units iteratively feed the burst instances to the output units .
the output units compete for the burst instances .
the output units collecting the most bursts are saved .
the saved units represent the clusters found within the data .
the unsupervised em algorithm @xcite assumes that the attribute space can be subdivided into a predetermined number of normally distributed clusters .
an initial guess is made as to the properties of each random distribution , and this guess is used to calculate probabilities that bursts belong to each cluster .
the cluster characteristics are iteratively adjusted until all clusters are optimally - defined .
the kmeans @xcite algorithm randomly selects @xmath7 data points as initial cluster centers .
each instance is then placed in the cluster to which it is most similar .
once all instances have been placed in their appropriate cluster , the cluster centers are updated by computing the mean of each new cluster .
the process continues until an iteration of the algorithm shows no change in the cluster centers .
predetermined classification significance helps define the number of classes that can be recovered . when allowed to find an optimum number of classes based on a default significance , the aforementioned classifiers typically recover three to four burst classes as opposed to the two traditionally accepted classes .
this indicates that the two traditional classes are not considered to be the optimal solution .
we force all four classifiers to recover two , three , and four classes because we hope that by studying the properties of these force - recovered classes we can determine why the three - class solution is preferred over the two - class solution .
the properties of three force - recovered classes are indicated in table [ tbl-1 ] .
the properties of these classes are similar to those obtained using other clustering techniques @xcite , so we again refer to these as class i ( long ) , class ii ( short ) , and class iii ( intermediate )
. however , these previous results typically place fewer bursts in class iii than class i , whereas three of our four classifiers place the largest number in class iii .
therefore , our analysis finds class iii to be the dominant class . in order to explain why the percentage of class iii members is so large , we examine the placement of class iii bursts when
classifiers are forced to recover only two classes ( short and long ) .
the results are remarkably consistent : all four classifiers fail to clearly delineate the traditionally - accepted short and long classes , and each places a large number of soft class iii bursts in with the hard short class ( see figure 1 ) .
this is surprising , since class iii clustering is not obvious to the naked eye in the hardness _
vs. _ duration parameter space whereas short and long burst clustering is . the hardness _ vs. _ duration boundary is not chosen by the classifiers
because a sharper one exists in the fluence _ vs. _ duration parameter space ( figure 2 ) ; _ the boundary separating short faint bursts from long bright bursts is more significant than that separating short hard bursts from long soft bursts . _
it is surprising that fluence plays such an important role in the classification .
first , fluence is an extrinsic attribute ( since it represents a convolution of a burst s luminosity and distance ) as opposed to hardness and duration , so there is no reason why fluence clustering should relate to any physical differences between burst classes .
second , one would intuitively expect a burst with a longer duration to have a larger fluence , indicating that fluence and duration should be highly - correlated attributes .
thus , clustering in the duration attribute can also cause clustering in the fluence attribute , and the use of fluence as a classification attribute magnifies the clustering importance of duration relative to hardness .
the break between short faint and long bright bursts therefore appears due in part to the use of fluence , an attribute which is of questionable value . to determine if the fluence bias can be removed , we eliminate this attribute and perform the classification using only @xmath6(t90 ) duration and @xmath6(hr321 ) hardness ratio . even without fluence ,
the classifiers again prefer to recover three classes instead of two , and class iii is not diminished in size .
examination of the three class properties indicates that @xmath6(t90 ) has been used almost exclusively to delineate the classes ; hardness is almost ignored by the classifiers .
this is surprising , since the eye tends to delineate two burst classes .
we check this result by supplying _ only the t90 attribute _ to the classifiers . indeed , the classifiers again return three classes rather than two ( class i bursts have t90 @xmath8sec . , class ii bursts have t90 @xmath9 sec . , and class iii bursts have @xmath10 sec . @xmath11 t90 @xmath12 sec . ) .
however , the size of class iii has been diminished in this reclassification and it is no longer the largest class ; this result is consistent with that obtained earlier using only the duration attribute @xcite . we conclude that strong evidence exists for the three - class structure . before accepting the new class as a separate source population , we must try to discount alternative explanations concerning its existence .
it is possible that the classifiers have detected a data cluster resulting from the way that the data have been collected , rather than from a separate and distinct source population .
we consider it unlikely that class iii represents a statistical anomaly since it has been found by four classifiers using different algorithms , and since stringent requirements have been imposed for each classifier to find additional classes .
thus , class iii could result from a systematic effect such as an instrumental or sampling bias .
the suspected bias appears to primarily affect duration and the coupled yet extrinsic attribute of fluence .
this conclusion leads us again to examine the hypothesized fluence duration bias .
this bias could provide a mechanism for underestimating both fluence and duration of some class i bursts ( particularly faint soft ones ) , and could cause these bursts to take on class iii characteristics .
however , with the increased size of class iii , it is reasonable to think that the bias might be strong . in an attempt to quantify the fluence duration bias ,
we have developed a simple model of the bias that can be applied statistically .
the model only influences class i and class iii bursts ( as defined by the em algorithm ) , since the bias has not been hypothesized to alter class ii properties . in a previous work @xcite we estimated the maximum amount
by which the fluences and durations of five bright bursts might need to be corrected if their signal - to - noise ratios were reduced ; our simple model averages these values to obtain _ maximum corrections _ of fluence and duration as functions of @xmath13 ( peak flux measured on the 1024 ms timescale ) .
we do not know how much the fluence of an individual burst might need to be corrected , therefore we assume that the fluence of each burst should be corrected between 0 ergs @xmath14 sec@xmath15 and the maximum fluence correction @xmath16 , and that the duration of each burst should be corrected between 0 seconds and the maximum duration correction t90@xmath17 .
the amount of the maximum correction is dependent upon the signal - to - noise ratio and thus on the peak flux ; the suspected bias is more pronounced for bursts with peak fluxes near the detection threshold .
we naively assume a probability @xmath18 that each burst s measured fluence and duration will be altered with equal probability in the intervals @xmath19 $ ] and @xmath20 $ ] .
thus , the modeled amount by which an individual burst s fluence would be affected by the bias is @xmath21 and the amount by which its duration would be affected is @xmath22 . the problem can be inverted to estimate how much observed burst fluences and durations have been underestimated as a function of @xmath13 . if the fluence duration bias produces class iii properties , then ( 1 ) the faint class i and class iii bursts ( as measured by @xmath13 ) should show evidence of having their fluences ( and durations ) systematically underestimated , and ( 2 ) no evidence of this bias should be present if this combined distribution has been properly corrected for the effect
. we would thus like to compare both the observed distribution and the corrected distribution with the `` true '' distribution .
unfortunately , we do not know the `` true '' distribution .
if we assume that the bias has not affected the fluence and duration distributions of bright bursts ( as measured by @xmath13 ) , then we can compare the corrected and uncorrected distributions of faint bursts to the observed distributions of bright bursts .
the comparison can be made once we identify how the fluence and duration distributions scale with peak flux .
if a given burst s intensity were decreased ( either by decreasing the burst s luminosity or if the burst were observed at a larger distance ) , then its fluence would decrease proportionally to its peak flux .
this generic statement is false only in the presence of sampling and/or instrumental biases .
the effect of time dilation due to cosmological expansion is an example of a sampling bias that can systematically affect fluence count rates differently than peak flux count rates .
since we measure the peak flux and fluence in the same energy channels , the primary source of bias is that the observed peak flux can be as little as @xmath23 of its actual value due to time dilation , whereas the fluence would not be expected to be lessened .
this bias would cause the peak flux of distant bursts to be small relative to the fluence ; note that this bias can not explain class iii characteristics , since class iii bursts have fluences that are small relative to their peak fluxes . in going from bright bursts to faint bursts , a decreasing signal - to - noise ratio can cause fluences of faint bursts to decrease non - proportionally to peak fluxes ; this is an example of a statistical ( rather than systematic ) instrumental bias .
sampling biases can cause the faint burst distribution ( as measured by either peak flux or fluence ) to be different than the bright burst distribution .
trigger biases can cause bursts with certain characteristics to trigger disproportionately relative to other bursts .
however , trigger biases that have been proposed prior to this manuscript do not appear to alter the makeup of the batse dataset by large amounts @xcite . _
we therefore assume in testing the fluence duration bias that it is primarily responsible for causing a burst s fluence to change not in proportion to the change in its peak flux , and that the faint burst distribution of class i + iii bursts would be the same as the bright distribution in the absence of this effect .
_ the distribution of burst fluences at a given peak flux indicates bursts with different time histories ; greater fluence typically belongs to longer bursts with more pulses and smaller fluence typically belongs to shorter bursts with fewer pulses .
if these burst peak fluxes were all decreased by the same amount , then their fluences would decrease proportionally along a line defined by @xmath24 ( where @xmath25 is an arbitrary constant ) .
the difference @xmath26 obtained for each burst indicates the fluence offset of each burst from the line given its peak flux .
the distribution of @xmath27 can be examined for bright bursts ( _ e.g. _ those presumably unaffected by the bias ) and for faint bursts ( _ e.g. _ those affected by the bias ) . in the absence of any biases ,
the faint distribution will be similar to the bright distribution .
if the fluence duration bias is present , then the faint distribution will differ from the bright distribution .
the aforementioned statistical correction should make the corrected " fluence distribution @xmath28 more compatible with the bright distribution than is the uncorrected faint distribution .
figure 3 is a plot of @xmath29 _ vs. _
@xmath30 for the burst sample used in this study .
class i , ii , and iii bursts have been identified using the unsupervised em algorithm
. the proportional decrease of fluence and peak flux is shown for a hypothetical class i burst ( diagonal line ) ; the curving path indicates how the bias might affect the measured fluence of this burst as a function of @xmath13 ( curving line ) in the case where @xmath31 .
the amount by which the fluence would need to be corrected @xmath32 is also shown ( vertical line ) .
we construct eight @xmath27 bins for the set of bright bursts and eight bins for the faint bursts ( the zero point for the @xmath27 scale is arbitrary , so we use @xmath33 ) .
the dividing line between `` bright '' and `` faint '' bursts is set at @xmath34 photon @xmath14 sec@xmath15 since bursts brighter than this value should be essentially unaffected by the proposed fluence duration bias .
the faint uncorrected @xmath27 distribution is moderately different than the bright distribution , with a @xmath35 for 7 degrees of freedom and a corresponding probability of @xmath36 .
the fluence distribution ( as determined from @xmath27 ) has been shifted to lower values consistent with the fluence duration bias . in order to test the correction by the proposed model
, we correct the fluence of each of the @xmath37 bursts by differing amounts @xmath38 .
the @xmath39 of the corrected faint burst distribution is again compared to the `` control '' sample of bright bursts .
since we might have overcorrected some bursts while undercorrecting others , we run the analysis a total of 100 times and average the results .
the corrected @xmath32 distribution is significantly different than the bright distribution ( we obtain @xmath40 for 7 degrees of freedom and a corresponding probability of @xmath41 that the two distributions are identical ) _ indicating that our model has significantly overcorrected for the suspected bias . _
similar results are obtained using the @xmath42 distributions .
since we have apparently overestimated the amplitude of the fluence duration bias @xmath43 for typical bursts , we can decrease our estimate of the bias by introducing a free parameter d in the relationships @xmath44 and @xmath45 , where @xmath46 represents the uncorrected sample while @xmath47 indicates the originally hypothesized bias . in table
[ tbl-2 ] , we demonstrate the effectiveness of the fluence duration bias for different values of @xmath48 chosen arbitrarily . the model fit is only improved when we reduce our estimates of @xmath49 and @xmath50 significantly .
we have shown previously @xcite that the maximum time interval used to calculate burst fluences decreases dramatically near the batse detection threshold ( essentially no bursts in the 3b catalog with @xmath51 photons @xmath14 sec@xmath15 have fluence durations @xmath52 seconds ) .
this indicates that the fluence duration bias causes fluences and durations of very long bursts with small peak fluxes to be underestimated .
our current analysis supports this hypothesis : the @xmath27 and @xmath42 distributions suggest that shorter bursts with small peak fluxes have probably not been affected by the bias , whereas some longer bursts have .
our experience with batse data analysis procedures is also in agreement : fluence duration intervals are rarely chosen to be shorter than many tens of seconds , and the time histories of only a few bursts are particularly susceptible to this bias @xcite .
this should prevent a systematic bias from being introduced for short bursts with small peak flux but not necessarily for long bursts with small peak flux .
these results indicate that the fluence duration bias does not influence faint bursts to the extent hypothesized previously . the shorter class i bursts , which were originally thought most likely to take on class iii characteristics via the bias
, are apparently affected the least .
the fluence duration bias appears to primarily influence the properties of some longer batse bursts .
we conclude that the fluence duration bias is not responsible for the large number of shorter softer bursts comprising class iii .
although the fluence duration bias does not appear to be responsible for the creation of class iii , our analysis of the proportional decrease of fluence and peak flux has unexpectedly provided new insight into measured burst properties .
the faint fluence and duration distributions used in classification are truncated because the samples triggered using short - timescale peak fluxes .
this truncation has the potential of biasing the sample via sample incompleteness . in order to study the potential effects of sample truncation
, we consider the advantages of a _ fluence - limited _ sample relative to a peak flux - limited sample .
an experiment triggering with a short integration window is more likely to detect a short burst than an experiment triggering with a long integration window , because in the latter case the entire burst flux can be recorded in a single temporal bin .
a peak flux - limited sample is thus biased towards shorter bursts relative to longer bursts because it excludes longer bursts having large fluences but with peak fluxes too faint to trigger .
however , a long timescale trigger ( such as one that could trigger on fluence ) would prove to be equally - biased .
a hypothetical experiment triggering on fluence ( _ e.g. _ one with a 10,000 second integration window ) would be more likely to detect a faint long burst ( having little of its fluence in one temporal window ) than would an experiment triggering on peak flux .
a fluence - limited sample would be biased towards longer bursts because it would include faint longer bursts but exclude faint shorter bursts with the same peak flux .
figure 3 demonstrates that an excessive number of short class iii bursts are found near batse s peak flux trigger ; the fluence distribution of these bursts is acutely truncated by the peak flux trigger .
_ thus , class iii occupies a fluence vs. peak flux region where the instrumental ( peak flux ) trigger favors detection of shorter bursts over longer ones .
_ we would like to identify a peak flux measure that does not suffer from truncation of the duration distribution .
the proportional relationship between fluence @xmath53 and peak flux @xmath13 as a burst s luminosity is decreased or as its distance is increased provides a method for identifying such a peak flux measure
. we can re - define fluence to be a peak flux by defining an extremely long temporal window @xmath54 ( @xmath54 is a constant ) that contains the entire flux of the sample s longest burst .
the fluence divided by this temporal window @xmath55 is a peak flux ( having units of photons @xmath14 sec@xmath15 or ergs @xmath14 sec@xmath15 , using an approximate transformation of @xmath56 ergs photon@xmath15 ) @xcite .
the equation governing this proportional decrease in peak flux and fluence is @xmath57 or @xmath58 where @xmath59 has units of flux and is thus a measure of burst intensity .
we define this quantity as the _ dual timescale peak flux _
@xcite since it uses two different timescale measurements .
the minimum value of the dual timescale peak flux @xmath59 can be called the _ dual timescale threshold_. the dual timescale peak flux is merely a multiple of this threshold value , @xmath60 ( or @xmath61 ) , where @xmath7 is a constant .
a dual timescale threshold can be defined as an instrumental setting for a gamma - ray burst experiment ( _ e .
g. _ by requiring @xmath62 to exceed a trigger value ) , as a selection process on previously - detected events in a standard experiment , or with archival data from an experiment triggering independently on one temporal trigger at a time ( such as batse ) .
this latter concept is not new ; several studies have developed their own _ post facto _ batse triggers using archival time series data @xcite .
the dual timescale peak flux treats longer bursts having larger fluences and smaller peak fluxes on an equal basis with shorter bursts having smaller fluences and larger peak fluxes : _ these bursts with different temporal structures have something in common , which is that they have equal probability of detection using the dual timescale peak flux . _ their differences must therefore be defined by a line orthogonal to the dual timescale peak flux , _
e. g. _ one that satisfies the relationship @xmath63 or @xmath64 where @xmath65 has units of time and represents a duration .
we call @xmath65 the _ flux duration _ @xmath65 ; it measures the total time that a burst could emit at its peak flux in order to produce its fluence .
longer bursts typically have large @xmath66 values and shorter bursts should typically have small @xmath66 values .
in fact , the correlation for class i + iii bursts demonstrated in figure 4 has a spearman rank - order correlation significance 0f @xmath67 that @xmath65 and t90 are uncorrelated .
the dual timescale peak flux does not favor bursts of any duration ( longer than the smallest 1024 ms integration window ) , whereas peak flux or fluence do by truncating the distribution and thereby favoring `` faint '' bursts ( _ e.g. _ those near the threshold ) of longer or shorter durations .
since the dual timescale peak flux does not truncate the duration distribution , we can say that the dual timescale peak flux - limited sample retains the duration characteristics of the sample by preserving the duration @xmath66 relative to the intensity @xmath68 .
it was recognized soon after batse s launch @xcite that long - timescale triggers underestimated the intensities of shorter bursts and biased the sample
. however , our analysis demonstrates ( perhaps surprisingly ) that short temporal timescale triggers would also bias the sample against longer bursts .
figure 5 demonstrates how batse s peak flux trigger influences the number of events placed in class iii relative to those placed in class i. shorter bursts ( small @xmath66 ; denoted by region ` c ' ) have been detected while faint longer bursts ( large @xmath66 ; denoted by ` a ' ) have been excluded by the trigger .
a hypothetical fluence trigger allowing the faintest shorter bursts currently detected by batse to trigger would not resolve this problem : shorter bursts ( large @xmath66 ; denoted by ` b ' ) would go undetected by the fluence trigger relative to longer bursts ( small @xmath66 ; denoted by ` d ' ) .
a dual timescale threshold is shown ( diagonal dashed line ) that favors neither longer nor shorter batse bursts .
the threshold excludes most of the bursts previously identified as class iii because these have been favored by the one - second trigger window relative to longer bursts .
it also excludes many class ii bursts which are both typically faint and shorter than the one - second trigger window .
we make a cut on our batse sample that is equally complete for both longer ( large @xmath66 ) and shorter ( small @xmath66 ) bursts and use this as our dual timescale threshold .
this threshold follows the relation @xmath69 , and has been chosen so that even the longest bursts with the largest @xmath66 values are detected by batse s actual peak flux trigger .
this is demonstrated by the diagonal dotted line in figure 4 , and corresponds to a dual timescale peak flux ( via equation 1 ) of @xmath70 photons @xmath14 sec@xmath15 for @xmath71 seconds ( the t90 of the longest burst in the sample ) .
we wish to determine how the sample properties vary with dual timescale peak flux .
we thus divide our sample of class i + iii into four subsamples containing similar numbers of bursts but with different dual timescale peak fluxes : bright bursts ( @xmath72 , or @xmath73 ) , moderately bright bursts ( @xmath74 , or @xmath75 ) , faint bursts ( @xmath76 , or @xmath77 ) , and the faintest bursts ( @xmath78 , or @xmath79 ) .
the first three samples are `` brighter '' than the dual timescale threshold , the faintest sample consists of bursts fainter than the dual timescale threshold and is primarily composed of bursts from class iii .
the bin sizes are chosen so that the three with @xmath80 contain similar numbers of bursts , and so that each bin contains enough bursts to constitute a reasonable statistical sample .
we identify three flux duration intervals from the sample : longer bursts ( @xmath81 ) , middle bursts ( @xmath82 ) , and shorter bursts ( @xmath83 ) .
the bin sizes are again chosen so that each bin contains similar numbers of bursts , and so that each bin contains enough bursts to constitute a reasonable statistical sample .
longer bursts as measured by the flux duration ( @xmath84 seconds ) are also long as measured by t90 ( @xmath85 seconds ) ; the same correlation holds true for middle bursts ( @xmath86 seconds and @xmath87 seconds ) and shorter bursts ( @xmath88 seconds and @xmath89 seconds ) .
the quantity @xmath90 is the burst emission time relative to the flux duration ; this is the amount by which the actual burst emission time is stretched relative to the time interval during which the burst could have emitted at the peak flux rate .
it is interesting to note that @xmath91 for class i + iii bursts .
bursts with @xmath92 and @xmath6(t90 ) values that do not closely follow this relationship have unconvential time histories ( see figure 4 ) .
the attribute @xmath65 is closely related to grb duty cycle @xcite .
the duty cycle @xmath93 measures the persistence of burst emission via the relationship @xmath94 where @xmath95 is the average energy per photon and @xmath96 is the 64 ms peak flux .
a large duty cycle ( @xmath97 ) indicates persistent emission whereas a small duty cycle @xmath98 indicates sporadic emission . using equation ( 2 )
, it can be seen that @xmath99 .
thus , a burst with a large @xmath100 ratio is persistent because it emits at a high rate for a long time relative to its total duration .
we have previously shown that class ii bursts have larger values of @xmath93 and harder spectral indices than class i and class iii bursts@xcite , supporting the hypothesis that these short , hard bursts belong to a different source population . on the other hand ,
class iii bursts are generally softer than class i but have similar @xmath93 values ; the properties of these two classes overlap considerably .
if our hypothesis is correct that class i + iii comprises one population , then we expect that @xmath101 and @xmath65 will deconvolve complex relationships previously measured with the attributes @xmath53 and @xmath13 .
figure 6 demonstrates the relationship between hr321 and @xmath101 for the combined class i + iii burst sample .
a strong correlation exists between hardness and dual timescale peak flux for bursts of all durations ; the hardness ratios are similar for all bursts of the same @xmath101 regardless of whether t90 or @xmath65 is used .
it is also seen that the faintest bursts in the sample ( short bursts fainter than the dual timescale trigger ) appear to extend this relation .
this evidence supports our hypothesis that the bulk of the class iii bursts are short class i bursts that have preferentially been detected by batse s short timescale trigger . if a corresponding sample of longer class i bursts is detected ( by having a lower peak flux trigger threshold and/or by having some bursts trigger on a longer timescale )
, then these bursts most likely would be as soft as the class iii bursts .
we suggest that the fxts ( fast x - ray transients ) found by bepposax @xcite using an x - ray trigger and subsequently identified in batse data @xcite might be long soft bursts that previously escaped detection .
the slope of the @xmath6(hr321 ) _ vs. _ @xmath6(peak flux ) relation is largest when @xmath102 is used as the peak flux measure as opposed to either @xmath29 or @xmath30 ; this is true regardless of whether the sample is peak flux - limited , fluence - limited , or duration - limited .
this result is demonstrated in table [ tbl-3 ] , where hardness _ vs. _ intensity correlations are examined using a spearman rank - order correlation test for the three different intensity measures : the 1024 ms peak flux @xmath13 , the fluence @xmath53 , and the dual timescale peak flux @xmath101 .
small probabilities indicate strong correlations between spectral hardness and the peak flux measure .
spectral hardness ( and presumably redshift ) correlates better with the dual timescale peak flux than with any other peak flux measure , regardless of which measure is used to _
select _ the sample .
furthermore , the @xmath6(hr321 ) _ vs. _ @xmath6(peak flux ) slope is essentially identical for burst samples of different t90 durations when @xmath102 is used ; it does not appear that the same can be said when either @xmath30 or @xmath29 are used as a peak flux measure .
_ thus , @xmath101 appears to more easily deconvolve the attributes of hardness , duration , and peak flux than do either @xmath53 or @xmath13_. we take this to indicate that @xmath101 is a preferred peak flux indicator to @xmath53 and @xmath13 .
there are potentially far - reaching consequences to having @xmath101 as a less - biased temporal flux measure . to date , essentially all statistical studies have used either @xmath30 or @xmath53 as intensity measures ( e.g. @xmath103 _ vs. _ @xmath29 , @xmath104 _ vs. _ @xmath30 , @xmath105 _ vs. _ @xmath30 ) .
these studies are potentially biased because @xmath53 and @xmath13 do not deconvolve the hardness intensity correlation as cleanly as does @xmath101 .
presumably , studies made using @xmath53 and @xmath13 combine longer bursts measured at one value of hr321 with short bursts measured at another value of hr321 .
the use of @xmath101 in future modeling endeavors might improve our understanding of gamma - ray burst properties better than do either fluence or peak flux .
we test our hypothesis that class i bursts and class iii bursts belong to the same population by submitting all bursts in the original sample brighter than @xmath59 to the em algorithm for unsupervised classification , and using the attributes of @xmath53 , t90 , and hr321 .
the classifier preferably recovers six classes as opposed to three ; the original three - class structure is lost as a result of the new trigger . despite this , class ii is still easily identifiable even though it contains only 40 members ( class ii bursts in the batse catalogs thus appear to have been preferentially detected as a result of batse s short timescale trigger ) .
the remaining bursts are placed in five classes with properties not recognizable as belonging to the original class i or class iii
. these classes may provide interesting additional insights into burst properties , but they warrant no further discussion here because they are not identifiable as the original burst classes . thus , strong reasons exist that the class iii cluster arises primarily from the shape of the attribute space defined by batse s peak flux trigger , and not from a separate source population .
our results support the hypothesis that class iii does not represent a separate source population .
we have demonstrated that both fluence and duration are truncated by batse s peak flux trigger .
the truncation effectively oversamples short bursts relative to long bursts . as a result of this truncation ,
the database contains an excess of faint , short ( soft ) bursts .
the use of the dual timescale trigger supports the hypothesis that classes i and iii are really one continuous duration distribution with faint bursts being softer than bright bursts .
the properties of this continuous distribution become somewhat ambiguous at low signal - to - noise , where the fluence duration bias alters burst properties . on the other hand
, class ii appears to represent a separate source population from class i @xcite .
neither sampling biases nor instrumental biases appear to be responsible for creating class ii characteristics from class i bursts .
however , it should be noted that batse s short trigger timescales have aided in the large detection rate of these short events .
we have demonstrated that 1 . gamma - ray burst class iii does not have to represent a separate source population ; it can be produced by the integration time of the instrumental trigger , 2 .
the fluence duration bias by itself , as modeled from a sample of high signal - to - noise bursts , is unlikely to be responsible for the existence of class iii .
class iii is likely produced by an excess of short , low fluence bursts detected by batse s short trigger temporal window .
the excess bursts can be eliminated via a selection process that is dual timescale peak flux - limited , rather than peak flux - limited or fluence - limited .
the dual timescale peak flux measure resulting from this selection process appears to correlate better with hardness ( and therefore with @xmath105 and redshift ) than either peak flux or fluence .
this adds support to the argument that dual timescale peak fluxes correct the temporal limitations introduced by using single timescale peak fluxes .
dual timescale peak fluxes can be established for many combinations of temporal measurements .
the results found here are important to gamma - ray burst astrophysics as well as to the general problem of scientific classification .
data mining tools can help identify complex clusters in multi - dimensional attribute spaces .
the tools are sensitive to clusters and data patterns , as evidenced here because they have allowed us to discover clusters produced artificially as a result of sample incompleteness .
this sensitivity is advantageous , because a better understanding of instrumental response and sampling biases can be used to improve the design of future instruments .
we note that sample incompleteness is generic and applies to the detection of any transient sources identified as the result of a temporal trigger .
examples of transient event statistics that might be biased by a temporal trigger include flare stars , soft gamma repeaters , x - ray bursts , and earthquakes . however , the sensitivity of data mining tools can also cause problems .
data mining is central to the operation of planned virtual observatories , which will electronically combine data collected from a variety of instruments with a range of temporal , spectral , and intensity responses . since sample incompleteness can cause a single instrument with one set of characteristics to find phantom classes , classes identified using multiple instruments should be interpreted cautiously .
the instrumental responses of virtual observatory components will have to be accurately known in order for newly - identified classes to be recognized as separate source populations .
it is important to recognize that data mining techniques have their limitations .
principal component analysis has identified fluence , duration , and hardness as being critical gamma - ray burst classification attributes , while the trigger attribute of peak flux was not chosen .
data mining classifiers failed to recognize that attribute selection had removed the attribute that could have provided the most insight into the gamma - ray burst clustering structure .
we gratefully acknowledge nasa support under grant nra-98-oss-03 ( the applied information systems research program ) and nsf support under grant ast-0098499 ( research in undergraduate institutions ) .
we also thank james neff and robert dukes for valuable discussions .
atteia , j .- l . , barat , c. , boer , m. , dezalay , j .-
, niel , m. , talon , r. , vedrenne , g. , hurley , k. , sommer , m. , sunyaev , r. , kuznetsov , a. , and terekhov , o. 1994 , astron .
, 288 , 213 bagoly , z. , mszros , a. , horvth , i. , balazs , l. g. , and mszros , p. 1998
, , 498 , 342 balastegui , a. , ruiz - lapuente , p. , and canal , r. 2001 , , 328 , 283 bonnell , j. t. , norris , j. p. , nemiroff , r. j. , and scargle , j. d. 1997 , , 490 , 79 dempster , a. p. , laird , n. m. , and rubin , d. b. 1977 , journal of the royal statistical society , series b , 39 , 1 dezalay , j .- p . ,
atteia , j .- l . , barat , c. , boer , m. , darracq , f. , goupil , p. , niel , m. , talon , r. , vedrenne , g. , hurley , k. , terekhov , o. , sunyaev , r. , and kuznetsov , a. 1997 , , 490 , l17 haglin , d. j. , roiger , r. j. , hakkila , j. , pendleton , g. n. and mallozzi , r. 2000 , gamma - ray bursts , r. m. kippen , r. s. mallozzi , and g. j. fishman , new york : aip , 877 hakkila , j. , haglin , d. j. , pendleton , g. n. , mallozzi , r. s. , meegan , c. a. , and roiger , r. j. 2000a , , 538 , 165 hakkila , j. , preece , r. d. , and pendleton , g. n. 2000b , gamma - ray bursts , r. m. kippen , r. s. mallozzi , and g. j. fishman , new york : aip , 83 hakkila , j. giblin , t. w. , roiger , r. j. , haglin , d. j. , and paciesas , w. s. 2002 , proceedings of sci2002 , n. callaos , y. he , and j. a. perez - peraza , orlando : international institute of informatics and systematics press , 479 .
heise , j. , int zand , j. , kippen , r. m. , and woods , p. m. 2001 , gamma - ray bursts in the afterglow era , e. costa , f. frontera , and j. hjorth .
berlin heidelberg : springer , 16 .
horvth , i. , 1998 , , 508 , 757 hurley , k. , 1992 , gamma - ray bursts , w. s. paciesas and g. j. fishman , new york , aip , 3 kippen , r. m. , woods , p. m. , heise , j. , int zand , j. , preece , r. d. , and briggs , m. s. 2001 , gamma - ray bursts in the afterglow era , e. costa , f. frontera , and j. hjorth .
berlin heidelberg : springer , 22 .
klebesadel , r. w. , 1992 , gamma - ray bursts - observations , analyses and theories , c. ho , r. i. epstein , and e. e. fenimore 1992 , cambridge : cambridge univ .
press , 161 kohonen , t. 1982 , proceedings of the sixth international conference on pattern recognition clustering , taxonomy , and topological maps of patterns , m. lang , silver springs , md : ieee computer society press , 114 kommers , j. m. , lewin , w. h. g. , kouveliotou , c. , van paradijs , j. , pendleton , g. n. , meegan , c. a. , and fishman , g. j. 2001 , , 134 , 385 koshut , t. m. , paciesas , w. s. , kouveliotou , c. , van paradijs , j. , pendleton , g. n. , fishman , g. j. , and meegan , c. a. 1996 , , 463 , 570 kouveliotou , c. , meegan , c. a. , fishman , g. j. , bhat , n. p. , briggs , m. s. , koshut , t. m. , paciesas , w. s. , and pendleton , g. n. 1993 , , 413 , l101 lloyd , s. p. 1982
, ieee transactions on information theory , 2 , 129 mallozzi , r. s. , paciesas , w. s. , pendleton , g. n. , briggs , m. s. , preece , r. d. , meegan , c. a. , and fishman , g. j. 1995 , , 454 , 597 mazets , e. p. , and golenetskii , s. v. , 1981 , astrophys .
space sci .
, 75 , 47 meegan , c. a. , fishman , g. j. , wilson , r. b. , horack , j. m. , brock , m. n. , paciesas , w. s. , pendleton , g. n. , and kouveliotou , c. 1992 , nature , 355 , 143 meegan , c. a. , pendleton , g. n. , briggs , m. s. , kouveliotou , c. , koshut , t. m. , lestrade , j. p. , paciesas , w. s. , mccollough , m. l. , brainerd , j. j. , horack , j. m. , hakkila , j. , henze , w. , preece , r. d. , mallozzi , r. s. , and fishman , g. j. 1996 , , 106 , 65 meegan , c. a. , hakkila , j. , johnson , a. , pendleton , g. n. , and mallozzi , r. s. 2000 , gamma - ray bursts , r. m. kippen , r. s. mallozzi , and g. j. fishman , new york : aip , 43 mitrofanov , i. , pozanenko , a. , atteia , j .- l . , barat , c. , chernenko , a. , dolidze , v. , jourdain , e. , kozlenkov , a. , kucherova , r. , and niel , m. 1992 , gamma - ray bursts - observations , analyses and theories , c. ho , r. i. epstein , and e. e. fenimore , cambridge : cambridge univ .
press , 203 mukherjee , s. , feigelson , e. d. , babu , g. j. , murtagh , f. , fraley , c. and raftery , a. 1998 , , 508 , 314 nemiroff , r. j. , norris , j. p. , bonnell , j. t. , wickramasinghe , w. a. d. t. , kouveliotou , c. , paciesas , w. s. , fishman , g. j. , and meegan , c. a. 1994 , , 435 , l133 norris , j. p. , cline , t. l. , desai , u. d. , and teegarden , b. j. , 1984 , nature , 308 , 434 paciesas , w. s. , pendleton , g. n. , kouveliotou , c. , fishman , g. j. , meegan , c. a. , and wilson , r. b. 1992 , gamma - ray bursts , w. s. paciesas and g. j. fishman , new york : aip , 190 paciesas , w. s. , et al .
2002 , hete - ii woods hole conference on gamma - ray bursts , g. r. ricker , new york : aip , in press .
petrosian , v. , lee , t. t. , and azzam , w. j. , 1994 , in gamma - ray bursts , g. j. fishman , j. j. brainerd , and k. j. hurley , new york : american institute of physics , 93 qin , y .- p . , xie , g .- z . ,
liang , e .- w . , and zheng , x .-
2001 , astron . astrophys . , 369 , 537 rajaniemi , h. j. and mhnen , p. 2002
, , 566 , 202 roiger , r. j. , geatz , m. w. , haglin , d. j. , and hakkila , j. 1999 , proceedings of the federal data mining symposium & exposition 99 , w. t. price , fairfax , va : afcea international , 109 roiger , r. j. , hakkila , j. , haglin , d. j. , pendleton , g. n. , and mallozzi , r. s. 2000 , gamma - ray bursts , r. m. kippen , r. s. mallozzi , and g. j. fishman , new york : aip , 38 schmidt , m. , 1999 , , 138 , 409 stern , b. e. , tikhomirova , y. , kompaneets , d. , svensson , r. , and poutanen , j. , 2001 , , 563 , 80 van paradijs , j. , kouveliotou , c. , and wijers , r. a. m. j. , 2000 , ann .
astrophys . , 38 , 379 lrcccc & * property * & * esx * & * kohonen * & * em * & * kmeans * + & * no .
bursts * & 250 & 225 & 422 & 273 + & * @xmath6(fluence ) * & @xmath106 & @xmath107 & @xmath108 & @xmath109 + & * @xmath6(t90 ) * & 1.71 & 1.71 & 1.70 & 1.78 + & * @xmath6(hr321 ) * & 0.31 & 0.31 & 0.21 & 0.26 + & * no .
bursts * & 194 & 239 & 144 & 173 + & * @xmath6(fluence ) * & @xmath110 & @xmath111 & @xmath112 & @xmath113 + & * @xmath6(t90 ) * & 0.02 & 0.13 & @xmath114 & @xmath115 + & * @xmath6(hr321 ) * & 0.44 & 0.39 & 0.58 & 0.53 + & * no . bursts * & 354 & 334 & 232 & 352 + & * @xmath6(fluence ) * & @xmath116 & @xmath117 & @xmath118 & @xmath119 + & * @xmath6(t90 ) * & 1.38 & 1.51 & 1.02 & 1.28 + & * @xmath6(hr321 ) * & 0.06 & 0.07 & 0.04 & 0.06 + lccc @xmath48 & @xmath120 & dof & @xmath121 ( @xmath122 ) + 1 & 34 & 7 & @xmath123 + 0.667 & 23 & 7 & @xmath124 + 0.5 & 19 & 7 & @xmath125 + 0.25 & 13 & 7 & @xmath126 + 0.1 & 12 & 7 & 0.11 + 0 & 14 & 7 & 0.055 + lccc prob . of no correlation between hr321 and : & @xmath13 & @xmath53 & @xmath101 + @xmath13-limited sample & @xmath127 & @xmath128 & @xmath129 + @xmath53-limited sample & @xmath130 & @xmath131 & @xmath132 + @xmath101-limited sample & @xmath133 & @xmath134 & @xmath135 + | unsupervised pattern recognition algorithms support the existence of three gamma - ray burst classes ; class i ( long , large fluence bursts of intermediate spectral hardness ) , class ii ( short , small fluence , hard bursts ) , and class iii ( soft bursts of intermediate durations and fluences ) .
the algorithms surprisingly assign larger membership to class iii than to either of the other two classes .
a known systematic bias has been previously used to explain the existence of class iii in terms of class i ; this bias allows the fluences and durations of some bursts to be underestimated ( hakkila et al . , apj 538 , 165 , 2000 ) .
we show that this bias primarily affects only the longest bursts and can not explain the bulk of the class iii properties .
we resolve the question of class iii existence by demonstrating how samples obtained using standard trigger mechanisms fail to preserve the duration characteristics of small peak flux bursts .
_ sample incompleteness is thus primarily responsible for the existence of class iii . _ in order to avoid this incompleteness , we show how a new dual timescale peak flux can be defined in terms of peak flux and fluence .
the dual timescale peak flux preserves the duration distribution of faint bursts and correlates better with spectral hardness ( and presumably redshift ) than either peak flux or fluence .
the techniques presented here are generic and have applicability to the studies of other transient events .
the results also indicate that pattern recognition algorithms are sensitive to sample completeness ; this can influence the study of large astronomical databases such as those found in a virtual observatory . |
systems of identical particles exhibit unique correlations even without interactions , which originate solely from permutation symmetry.@xcite they manifest themselves in the pair distribution function @xmath1 , i.e. , the relative probability of finding a pair of particles separated by distance @xmath12 .
specifically , normal ideal bosons show enhancement from @xmath13 to @xmath14 , signaling an effective attraction between particles , whereas that of equal - spin fermions is reduced to @xmath15 as predicted by the pauli exclusion principle .
the two distinct correlations are also measurable as bunching and antibunching effects in the experimental detection of particles , as first observed by hanbury brown and twiss with photons back in 1956.@xcite furthermore , @xmath16 below the bose einstein condensation ( bec ) temperature @xmath4 should exhibit a conspicuous reduction@xcite that may be used to monitor the development of coherence in condensates.@xcite more generally , much attention has been focused on the @xmath17-body correlations @xmath18 of quantum gases both experimentally@xcite and theoretically@xcite over many decades . among them , the one - body correlation @xmath19 was used by penrose@xcite and by penrose and onsager@xcite to define bec by @xmath20 , i.e. , the expectation called _ the off - diagonal long - range order _ by yang.@xcite however , there is a fundamental issue on the theoretical side that almost all the calculations have been performed in the grand canonical ensemble,@xcite which is known to suffer from unphysically huge particle - number fluctuations below @xmath4.@xcite ad hoc approximations have been introduced to remove them , such as the introduction of the `` anomalous '' average @xmath9 of the field operator @xmath10 , which is apparently incompatible with the particle - number conservation.@xcite a calculation within the canonical ensemble was performed in this context , but with an approximation by discarding several terms.@xcite hence , it is desirable to calculate one- and two - body correlations completely within the canonical ensemble , observe the approach to the thermodynamic limit , and check the validity of the approximations adopted in the grand canonical calculations .
we carry out such a calculation using a new formula for the reduced density matrices in the canonical ensemble , which is the equivalent of the bloch de dominicis ( or wick s ) theorem in the grand canonical ensemble@xcite for normal - ordered products of operators .
thus , the formula may also be used in perturbative calculations with respect to interactions in the canonical ensemble , such as self - consistent hartree - fock calculations at finite temperatures with a fixed number of particles . in sect .
[ sec : level2 ] we derive an analytic expression for the reduced density matrices of ideal gases in the canonical ensemble
. section [ sec : level3 ] presents numerical results on the one- and two - body correlation functions of homogeneous bose and fermi gases in the canonical ensemble for different particle numbers at several temperatures in comparison with those of the grand canonical ensemble .
concluding remarks are given in sect .
[ sec : level4 ] .
the appendix proves that a superposition of the obtained formula for the reduced density matrices in terms of the particle number reproduces the bloch de dominicis theorem in the grand canonical ensemble for normal - ordered products of operators .
we consider a canonical ensemble of @xmath0 identical particles with spin @xmath21 described by hamiltonian @xmath22 . the reduced @xmath17-particle density matrix ( @xmath23 )
is defined by@xcite @xmath24 here @xmath25 is the partition function with @xmath26 by definition , tr denotes the trace , @xmath27 with @xmath28 and @xmath29 the boltzmann constant and temperature , respectively , and @xmath30 is the field operator of bosons or fermions with @xmath31 denoting space ( @xmath32 ) and spin ( @xmath33 ) coordinates . the purpose of this section is to prove that eq .
( [ rho^(n)-def ] ) for _ non - interacting systems _ is expressible as @xmath34 here @xmath35 and @xmath36 for bosons and fermions , @xmath37 denotes a permutation with @xmath17 elements,@xcite @xmath38 @xmath39 and @xmath40 for even and odd permutations , respectively , and @xmath41 is defined in terms of the one - particle eigenvalues @xmath42 and orthonormal eigenfunctions @xmath43 as @xmath44 where @xmath45 distinguishes one - particle eigenstates .
let us write eq .
( [ rho^(n ) ] ) for @xmath46 explicitly for later purposes : [ rho12 ] @xmath47 @xmath48 .
\label{rho2}\end{aligned}\ ] ] note that @xmath49 is not expressible concisely in terms of @xmath50 , unlike the case of the grand canonical ensemble .
two comments on eq .
( [ rho^(n ) ] ) are in order .
first , eq .
( [ rho1 ] ) enables us to reproduce the following recurrence relation for the partition function:@xcite @xmath51 by setting @xmath52 , subsequently integrating over @xmath53 , and using the sum rule:@xcite @xmath54 where the integration over @xmath53 involves summation over @xmath55 and integration over @xmath56 .
second , eq . ( [ rho^(n ) ] ) corresponds to the bloch de dominicis ( or wick s ) theorem in the grand canonical ensemble for normal - ordered products of operators.@xcite indeed , multiplying eq .
( [ rho^(n ) ] ) by @xmath57 , with @xmath58 the chemical potential , and summing over @xmath0 , we obtain the @xmath17-particle density matrix @xmath59 in the grand canonical ensemble as @xmath60 where @xmath61 with @xmath62 the grand partition function ; see the appendix for details of the derivation . this is exactly the wick decomposition for the average @xmath63 without the expectation @xmath64 , expressed solely in terms of the one - particle density matrix @xmath65 . to prove eq .
( [ rho^(n ) ] ) , we expand the field operators as @xmath66 our hamiltonian is given explicitly in terms of @xmath67 , @xmath68 , and @xmath69 as@xcite @xmath70 and the eigenstates of @xmath71 are expressible as @xmath72 ( @xmath73 ) for bosons ( fermions ) in the canonical ensemble with @xmath74 .
the corresponding eigenvalues are given by @xmath75 the partition function of ideal gases can be written in two different forms as @xmath76 the first sum is over all possible distinct sets of occupation numbers , whereas the second one can be performed independently over each @xmath77 ( @xmath78 ) by using a factor that obeys the following recurrence relation:@xcite @xmath79 with @xmath80 . note that @xmath81 for @xmath35 is the number of ways of arranging the set @xmath82 , which may contain identical states .
on the other hand , @xmath83 for @xmath84 has the effect of removing multiple occupancies from any one - particle state . in both cases
, we can express @xmath83 for every distinct permutation of the state [ w_n-2 ] @xmath85 as @xmath86 with @xmath87 and @xmath88 for fermions .
we prove eq .
( [ rho^(n ) ] ) by induction .
first , the case for @xmath89 is shown to hold by using the orthonormal eigenfunctions@xcite @xmath90 as @xmath91 the last expression is identical to eq .
( [ rho^(n ) ] ) for @xmath89 .
now , suppose that eq .
( [ rho^(n ) ] ) holds true for a certain @xmath17 .
then @xmath92 can be obtained by setting @xmath93 in eq .
( [ rho^(n ) ] ) and integrating over @xmath94 as follows:@xcite [ rho^(n)-proof123 ] @xmath95 let us consider the cases with @xmath96 and those with @xmath97 separately in treating the last sum .
permutations for @xmath96 are expressible as @xmath98 whereas those for @xmath99 can be transformed as@xcite @xmath100 where @xmath101 is some permutation with @xmath102 elements .
this fact enables us to replace the sum over @xmath37 in eq .
( [ rho^(n)-proof1 ] ) by that over @xmath103 with the introduction of factors @xmath104 and @xmath40 for @xmath96 and @xmath97 , respectively .
we thereby obtain @xmath105 the first contribution in the round bracket can be transformed further by using @xmath106 as follows : @xmath107 where we used eq .
( [ z_n - rec ] ) .
for the second contribution in the round brackets of eq .
( [ rho^(n)-proof1 ] ) , the sums over @xmath108 can be transformed with a change of variables @xmath109 as @xmath110 the subsequent summation of @xmath111 over @xmath112 with the constraint @xmath113 yields @xmath114 . substituting the above two results into eq .
( [ rho^(n)-proof2 ] ) , we obtain @xmath115 \notag \\ & \,\times \sum_{\ell_1= 1}^m\cdots\sum_{\ell_{n-1}=1}^m\delta_{\ell_1+\cdots+\ell_{n-1},m } \sum_{\hat p_{n-1}}\sigma^{p_{n-1 } } \notag \\ & \,\times \prod_{j=1}^{n-1 } \phi_{\ell_j}(\xi_j,\xi_{p_j } ' ) , \label{rho^(n)-proof3}\end{aligned}\ ] ] which is exactly eq .
( [ rho^(n ) ] ) with the replacement @xmath116 .
this completes our proof of eq .
( [ rho^(n ) ] ) .
we study one- and two - body correlations of homogeneous ideal gases in the canonical ensemble based on eq .
( [ rho12 ] ) . specifically , we consider identical particles with mass @xmath117 in a box of volume @xmath118 with periodic boundary conditions .
one - particle eigenstates are distinguished in terms of the wave vector @xmath119 by @xmath120 , whose eigenvalues and eigenfunctions are given by @xmath121 equation ( [ phi_n ] ) can be written as @xmath122 note that @xmath123 as can be seen from eq .
( [ z_n - rec ] ) .
thus , the diagonal element @xmath124 does not depend on @xmath125 at all for homogeneous systems .
let us introduce one- and two - body correlation functions of homogeneous systems as [ g^(12)-def ] @xmath126 ^ 2 \rho^{(2)}_{n}(\xi_1,\xi_2;\xi_1,\xi_2 ) .
\label{g^(2)-def}\end{aligned}\ ] ] equation ( [ g^(2)-def ] ) is also called the pair distribution function .
we set @xmath127 in eq .
( [ g^(1)-def ] ) , integrate the resulting expression over @xmath128 , and use eq .
( [ rho1-sum ] ) .
we then find that @xmath129 we also integrate eq .
( [ g^(2)-def ] ) over @xmath130 and use the first relation of eq .
( [ rho^(n)-proof1 ] ) for @xmath46 .
we thereby obtain @xmath131 = 1 .
\notag\end{aligned}\ ] ] noting that ( i ) @xmath132 does not depend on @xmath133 in the absence of magnetic fields and ( ii ) @xmath134 for @xmath135 , we can express the above results in terms of the diagonal element @xmath136 as [ g12-sum ] @xmath137 @xmath138 = 1 .
\label{g^(2)-sum}\end{aligned}\ ] ] we will compare the functions in eq .
( [ g^(12)-def ] ) with the following ones in the grand canonical ensemble:@xcite [ g^(12 g ) ] @xmath139 @xmath140 ^ 2 + 2\frac{n_0}{n}\ell ( k_{\rm q}r)\right\ } , \label{g^(2g)}\ ] ] which are applicable to both bosons ( @xmath35 ) and fermions ( @xmath84 , @xmath141 ) and obtained for bosons below @xmath4 by assuming the anomalous average @xmath142 . here
@xmath143 is the number of condensed particles , which becomes finite for bosons below @xmath4 and is given by @xmath144$],@xcite @xmath145 is defined together with @xmath146 by @xmath147 and @xmath148 denotes @xmath149 with @xmath150 and @xmath151 .
note that @xmath0 in the grand canonical ensemble denotes @xmath152 .
the bec temperature @xmath4 for bosons and the fermi temperature @xmath153 for fermions are given in terms of @xmath146 by@xcite @xmath154 for the sum rules in the grand canonical ensemble , eq .
( [ g^(1 g ) ] ) also satisfies eq .
( [ g^(1)-sum ] ) , whereas eq .
( [ g^(2 g ) ] ) obeys @xmath155=1-\frac{\langle n^2\rangle-\langle n\rangle^2}{\langle n\rangle } , \label{g^(2g)-sum}\end{aligned}\ ] ] where @xmath156 denotes the grand canonical average ; eq .
( [ g^(2g)-sum ] ) can be shown similarly as eq .
( [ g^(2)-sum ] ) .
one of our main interests is how @xmath157 is different from @xmath1 bringing about the additional ( fluctuation ) term in eq .
( [ g^(2g)-sum ] ) . equation ( [ g^(12)-def ] ) can be evaluated numerically from eqs .
( [ rho12 ] ) and ( [ z_n - rec ] ) .
the key quantity here is the function @xmath158 defined in eq .
( [ qq1 ] ) .
it is expressible as [ poisson ] @xmath159 , \label{poisson1}\end{aligned}\ ] ] with @xmath160^{2/3}t / t_{\rm q}$ ] and @xmath161^{1/3}k_{\rm q}{\bm r}$ ] .
thus , @xmath162 is periodic in @xmath163 with period @xmath164 . using the poisson summation formula
, we can also transform eq .
( [ poisson1 ] ) into @xmath165 . \label{poisson2}\end{aligned}\ ] ] since it depends only on @xmath166 , we can obtain @xmath158 by choosing @xmath32 along the @xmath167-axis , for example
. equations ( [ poisson1 ] ) and ( [ poisson2 ] ) were found to give good convergence in the summation over @xmath168 for @xmath169 and @xmath170 , respectively .
round - off errors that may accumulate significantly in evaluating eqs .
( [ rho12 ] ) and ( [ z_n - rec ] ) were removed by high - precision calculations with mathematica .
we confirmed that the sum rule in eq .
( [ g12-sum ] ) is satisfied for each of our numerical calculations for @xmath171 . of fermions with @xmath172 in the canonical ensemble at @xmath173 for @xmath171 in comparison with those in the grand canonical ensemble ( solid lines ) .
[ fr1 ] ] of fermions with @xmath172 in the canonical ensemble at @xmath173 for @xmath171 in comparison with those in the grand canonical ensemble ( solid lines ) .
[ fr2 ] ] we now present numerical results on @xmath136 for @xmath46 .
figures [ fr1 ] and [ fr2 ] show the one- and two - body correlation functions of fermions with @xmath172 , respectively , in the canonical ensemble at @xmath173 for @xmath171 .
they are compared with eq .
( [ g^(12 g ) ] ) for the grand canonical ensemble .
we observe excellent agreement between the two ensembles for fermions .
the reduction of @xmath1 near the origin is due to the pauli exclusion principle and referred to as _ fermi hole _ , whose size is seen to decrease as the temperature is raised owing to the increase in the average kinetic energy per particle .
tiny discrepancies between the two ensembles are seen in @xmath1 for @xmath174 , which reduce in magnitude as @xmath0 increases .
they are responsible for the difference between the sum rules , eqs .
( [ g^(2)-sum ] ) and ( [ g^(2g)-sum ] ) . in other words ,
the fluctuation term on the right - hand side of eq .
( [ g^(2g)-sum ] ) does not affect the short - range behavior of the pair distribution function of fermions . of bosons with @xmath175 in the canonical ensemble at @xmath176 for @xmath171 in comparison with those in the grand canonical ensemble ( solid lines ) .
[ br1 ] ] for bosons with @xmath175 in the canonical ensemble for @xmath171 in comparison with @xmath177 given by eq .
( [ g^(1 g ) ] ) in the grand canonical ensemble ( solid line ) . [ fig4 ] ] of bosons in the canonical ensemble for @xmath171 at ( a ) @xmath178 , ( b ) @xmath179 , ( c ) @xmath180 in comparison with that in the grand canonical ensemble ( solid lines ) .
[ fig5 ] ] for bosons in the canonical ensemble for @xmath171 in comparison with that in the grand canonical ensemble ( solid line ) .
[ fig6 ] ] we turn our attention to bosons with @xmath175 .
figure [ br1 ] shows the radial dependence of the one - body correlation function @xmath19 at @xmath181 in the canonical ensemble for @xmath171 in comparison with eq .
( [ g^(12 g ) ] ) for the grand canonical ensemble .
we also observe little dependence of the correlations on @xmath0 at @xmath182 without condensation , whereas the dependence becomes stronger below @xmath4 as the condensation develops . to see the development of the off - diagonal long - range order monitored by @xmath183,@xcite we also calculated @xmath19 at @xmath184 , i.e. , at the center of the box , which is the furtherest point from the origin , as a function of temperature .
figure [ fig4 ] shows the temperature dependence of the one - body correlation function in the canonical ensemble for @xmath171 in comparison with @xmath185 in the grand canonical ensemble given by eq .
( [ g^(1 g ) ] ) .
we observe that @xmath186 , which monitors the relative number of condensed particles , has a fairly large @xmath0 dependence and slowly approaches @xmath177 in the grand canonical ensemble .
figure [ fig5 ] shows the pair distribution function of bosons with @xmath175 at @xmath181 ( from top to bottom ) in the canonical ensemble with @xmath171 . for comparison
, we also plot eq .
( [ g^(2 g ) ] ) for the grand canonical ensemble .
we observe a substantial @xmath0 dependence in the two - body correlation , especially for @xmath187 . nevertheless , convergence to the grand canonical results as @xmath188 are seen clearly .
the bunching of @xmath14 above @xmath4 compared with @xmath13 is due to the effective attraction between bosons , which is seen to decrease below @xmath4 .
figure [ fig6 ] shows the temperature dependence of @xmath16 in the canonical ensemble with @xmath171 in comparison with @xmath189 given by eq .
( [ g^(2 g ) ] ) for the grand canonical ensemble .
we observe steep decreases of @xmath16 from @xmath3 above @xmath4 towards @xmath190 at @xmath6 as the condensation develops .
moreover , the reduction starts even from above @xmath4 for smaller values of @xmath0 .
we obtained an analytic expression for the reduced density matrices of quantum ideal gases in the canonical ensemble as eq .
( [ rho^(n ) ] ) .
the formula indicates that we can also perform the wick - decomposition - like procedure in the canonical ensemble .
hence , it can be used to study @xmath17-body correlations of ideal gases in the canonical ensemble and also to carry out perturbative calculations in the canonical ensemble .
the formula was subsequently used to clarify one- and two - body correlations of homogeneous bose and fermi gases .
they were also compared with those in the grand canonical ensemble , which were derived for condensed bosons by assuming the anomalous average @xmath11 below the bec temperature @xmath4 .
we confirmed that the correlations in the canonical ensemble approach those in the grand canonical ensemble as the particle number @xmath0 increases towards @xmath191 ; the approach is fast for fermions , whereas it becomes slower for bosons as the temperature is lowered through @xmath4 .
this fact justifies the procedure of introducing the anomalous average @xmath11 below the bec temperature @xmath4 for bosons in the grand canonical ensemble as a mathematical means of removing the unphysical particle - number fluctuations from the grand canonical ensemble . +
* acknowledgments * k. t. is a jsps research fellow , and this work was supported in part by jsps kakenhi grant number 15j01505 .
let us multiply eq .
( [ rho^(n ) ] ) by @xmath192 , perform a summation over @xmath193 , and transform the right - hand side as follows : @xmath194 the first line above is @xmath195 by definition .
we thereby obtain eq .
( [ wick - gc ] ) .
j. viana gomes , a. perrin , m. schellekens , d. boiron , c. i. westbrook , and m. belsley , phys .
a * 74 * , 053607 ( 2006 ) .
j. bosse , k. n. pathak , and g. s. singh , phys .
e * 84 * , 042101 ( 2011 ) .
t. m. wright , a. perrin , a. bray , j. schmiedmayer , and k. v. kheruntsyan , phys .
a * 86 * , 023618 ( 2012 ) .
w. ketterle and h .- j .
miesner , phys .
a * 56 * , 3291 ( 1997 ) .
m. yasuda and f. shimizu , phys .
* 77 * , 3090 ( 1997 ) .
e. a. burt , r. w. ghrist , c. j. myatt , m. j. holland , e. a. cornell , and c. e. wieman , phys .
lett . * 79 * , 337 ( 1997 ) .
i. bloch , t. w. hnsch , and t. esslinger , nature * 403 * , 166 ( 2000 ) .
a. ttl , s. ritter , m. khl , and t. esslinger , phys .
lett . * 95 * , 090404 ( 2005 ) .
m. schellekens , r. hoppeler , a. perrin , j. viana gomes , d. boiron , a. aspect , and c. i. westbrook , science * 310 * , 648 ( 2005 ) .
t. rom , th .
best , d. van oosten , u. schneider , s. flling , b. paredes , and i. bloch , nature * 444 * , 733 ( 2006 ) .
t. jeltes , j. m. mcnamara , w. hogervorst , w. vassen , v. krachmalnicoff , m. schellekens , a. perrin , h. chang , d. boiron , a. aspect , and c. i. westbrook , nature * 445 * , 402 ( 2007 ) .
e. haller , m. rabie , m. j. mark , j. g. danzl , r. hart , k. lauber , g. pupillo , and h .- c .
ngerl , phys .
lett . * 107 * , 230404 ( 2011 ) .
r. g. dall , a. g. manning , s. s. hodgman , w. rugway , k. v. kheruntsyan , and a. g. truscott , nat .
* 9 * , 341 ( 2013 )
. j. schmitt , t. damm , d. dung , f. vewinger , j. klaers , and m. weitz , phys .
* 112 * , 030401 ( 2014 ) .
o. penrose , philos . mag . * 42 * , 1373 ( 1951 ) . | we derive an expression for the reduced density matrices of ideal bose and fermi gases in the canonical ensemble , which corresponds to the bloch de dominicis ( or wick s ) theorem in the grand canonical ensemble for normal - ordered products of operators . using this expression
, we study one- and two - body correlations of homogeneous ideal gases with @xmath0 particles . the pair distribution function @xmath1 of fermions clearly exhibits antibunching with @xmath2 due to the pauli exclusion principle at all temperatures , whereas that of normal bosons shows bunching with @xmath3 , corresponding to the hanbury brown
twiss effect . for bosons below the bose
einstein condensation temperature @xmath4 , an off - diagonal long - range order develops in the one - particle density matrix to reach @xmath5 at @xmath6 , and the pair correlation starts to decrease towards @xmath7 at @xmath6 .
the results for @xmath8 are seen to converge to those of the grand canonical ensemble obtained by assuming the average @xmath9 of the field operator @xmath10 below @xmath4 .
this fact justifies the introduction of the `` anomalous '' average @xmath11 below @xmath4 in the grand canonical ensemble as a mathematical means of removing unphysical particle - number fluctuations to reproduce the canonical results in the thermodynamic limit . |
consider a minimal - delay space - time coded rayleigh quasi - static flat fading mimo channel with full channel state information at the receiver ( csir ) .
the input output relation for such a system is given by @xmath0 where @xmath1 is the channel matrix and @xmath2 is the additive noise .
both @xmath3 and @xmath4 have entries that are i.i.d .
complex - gaussian with zero mean and variance 1 and @xmath5 respectively .
the transmitted codeword is @xmath6 and @xmath7 is the received matrix .
the ml decoding metric to minimize over all possible values of the codeword @xmath8 is @xmath9 [ ld_stbc_def ] a linear stbc @xcite : a linear stbc @xmath10 over a real ( 1-dimensional ) signal set @xmath11 , is a finite set of @xmath12 matrices , where any codeword matrix belonging to the code @xmath10 is obtained from , @xmath13 by letting the real variables @xmath14 take values from a real signal set @xmath15 where @xmath16 are fixed @xmath12 complex matrices defining the code , known as the weight matrices .
the rate of this code is @xmath17 complex symbols per channel use .
we are interested in linear stbcs , since they admit sphere decoding ( sd ) @xcite which is a fast way of decoding for the variables .
a further simplified version of the sd known as the fast sphere decoding ( fsd ) @xcite ( also known as conditional ml decoding ) was studied by biglieri , hong and viterbo .
the quadratic form ( qf ) approach has been used in the context of stbcs in @xcite to determine whether quaternion algebras or biquaternion algebras are division algebras , an aspect dealing with the full diversity of the codes .
this approach has not been fully exploited to study the other characteristics of stbcs . in this paper
, we use this approach to study the fast sphere decoding ( fsd ) complexity of stbcs ( a formal definition of this complexity is given in subsection [ fsdc ] ) . designing stbcs with low decoding complexity has been studied widely in the literature .
orthogonal designs with single symbol decodability were proposed in @xcite , @xcite , @xcite .
for stbcs with more than two transmit antennas , these came at a cost of reduced transmission rates .
to increase the rate at the cost of higher decoding complexity , multi - group decodable stbcs were introduced in @xcite , @xcite , @xcite .
fast decodable codes ( codes that admit fsd ) have reduced sd complexity owing to the fact that a few of the variables can be decoded as single symbols or in groups if we condition them with respect to the other variables .
fast decodable codes for asymmetric systems using division algebras have been recently reported @xcite
. golden code and silver code are also examples of fast decodable codes as shown in @xcite and @xcite .
the properties of fast decodable codes and multi - group decodable codes were combined and a new class of codes called fast group decodable codes were studied in @xcite . in this subsection
we define the hurwitz radon quadratic form ( hrqf ) on any stbc .
we first recall some basics about quadratic forms .
more details can be seen in @xcite .
[ quad_form_def ] let @xmath18 be a field with characteristic not 2 , and @xmath19 be a finite dimensional @xmath18-vector space .
a quadratic form on @xmath19 is defined as a map @xmath20 such that it satisfies the following properties .
* @xmath21 for all @xmath22 and all @xmath23 . * the map @xmath24 $ ] for all @xmath25 is bilinear and symmetric .
if we consider @xmath19 as an @xmath26-dimensional vector space over @xmath18 , then we can also consider the quadratic form as a homogeneous polynomial of degree two , i.e. , for @xmath27 , we have scalars @xmath28 such that @xmath29 for all @xmath30 \in v$ ] .
hence , we can associate a matrix @xmath31 with the quadratic form such that @xmath32 [ hrqf_def ] the hurwitz radon quadratic form is a map from the stbc @xmath33 to the field of real numbers @xmath34 , i.e. , @xmath35 given by @xmath36 where @xmath37 is an element of the stbc and @xmath38 [ hrqf_is_qf_thm ] the map defined by is a quadratic form . the map @xmath39 needs to satisfy the conditions as defined in definition [ quad_form_def ] .
we have @xmath40 and @xmath41\ ] ] should be bilinear and symmetric where @xmath42 and @xmath43 . substituting and simplifying ,
we get @xmath44.\ ] ] it is clearly seen that this map is bilinear and symmetric .
we can associate a matrix with the hrqf .
if we define the matrix @xmath31 where @xmath45 such that @xmath46 , then we can write the hrqf as @xmath47 where @xmath48 $ ] .
notice that @xmath49 is a symmetric matrix and @xmath50 if and only if @xmath51 .
the following example shows that the fsd complexity depends on the ordering of the weight matrices or equivalently the ordering of the variables .
[ silver_code ] let us consider the silver code given by : @xmath52 where @xmath53 and @xmath54 take the alamouti structure , and @xmath55,$ ] @xmath56,$ ] @xmath57 $ ] and @xmath58^{t } = \textbf{u}\left [ s_{3 } , s_{4}\right]^{t},$ ] where @xmath59 is a unitary matrix chosen to maximize the minimum determinant and is given by @xmath60.$ ] let all the variables take values from a signal set of cardinality @xmath61 .
if we order the variables ( and hence the weight matrices ) as @xmath62 $ ] , then the @xmath63 matrix for sd has the following structure @xmath64,\ ] ] where @xmath65 denotes non zero entries .
we can clearly see that the silver code admits fast decoding with this ordering with fsd complexity @xmath66 .
however , if we change the ordering to @xmath67 $ ] , then the @xmath63 matrix for sd has the following structure @xmath68,\ ] ] where @xmath65 denotes non zero entries . with this ordering , the fsd complexity increases to @xmath69 .
the contributions of this paper are as follows : * we give a formal definition of the fsd complexity of a linear stbc ( subsection [ fsdc ] . ) * with the help of hrqf , it is shown that the fsd complexity of the code depends only on the weight matrices of the code with their ordering , and not on the channel realization ( even though the equivalent channel when sd is used depends on the channel realization ) or the number of receive antennas . * a best ordering ( not necessarily unique ) of the weight matrices provides the least fsd complexity for the stbc .
we provide an algorithm to be applied to the hrqf matrix which outputs a best ordering .
the remaining of the paper is organized as follows : in section [ sec2 ] the known classes of low ml decodable codes , the system model and the formal definition of the fsd complexity of a linear stbc are given . in section [ sec3 ] , we show that the fsd complexity depends completely on the hrqf and not on the channel realization or the number of receive antennas . in section [ sec4 ] , we present an algorithm to modify the hrqf matrix in order to obtain a best ordering of the weight matrices to obtain the least fsd complexity . concluding remarks constitute section [ sec5 ] .
_ notations : _ throughout the paper , bold lower - case letters are used to denote vectors and bold upper - case letters to denote matrices .
for a complex variable @xmath70 , @xmath71 and @xmath72 denote the real and imaginary part of @xmath70 , respectively .
the sets of all integers , all real and complex numbers are denoted by @xmath73 and @xmath74 , respectively .
the operation of stacking the columns of @xmath75 one below the other is denoted by @xmath76 .
the kronecker product is denoted by @xmath77 , @xmath78 and @xmath79 denote the @xmath80 identity matrix and the null matrix , respectively . for a complex variable @xmath70 ,
the @xmath81 operator acting on @xmath70 is defined as follows @xmath82.\ ] ] the @xmath81 operator can similarly be applied to any matrix @xmath83 by replacing each entry @xmath84 by @xmath85 , @xmath86 resulting in a matrix denoted by @xmath87 .
given a complex vector @xmath88^{t},$ ] @xmath89 is defined as @xmath90^{t}.$ ]
for any linear stbc with variables @xmath14 given by ( [ ld_stbc ] ) , the generator matrix @xmath91 @xcite is defined by @xmath92 where @xmath93^{t}$ ] . in terms of the weight matrices ,
the generator matrix can be written as @xmath94.\ ] ] hence , for any stbc , can be written as @xmath95 where @xmath96 is given by @xmath97 and @xmath93,$ ] with each @xmath98 drawn from a 1-dimensional ( pam ) constellation . using the above equivalent system model ,
the ml decoding metric can be written as @xmath99 using @xmath100 decomposition of @xmath101 , we get @xmath102 where @xmath103 is an orthonormal matrix and @xmath104 is an upper triangular matrix .
using this , the ml decoding metric now changes to @xmath105 if we have @xmath106 , $ ] where @xmath107 are column vectors , then the @xmath108 and @xmath63 matrices have the following form obtained by the gram - schmidt orthogonalization : @xmath109 , \ ] ] where @xmath110 are column vectors , and @xmath111,\ ] ] where @xmath112 and for @xmath113 @xmath114 in case of a multi - group decodable stbc , the variables can be partitioned into groups such that the ml decoding metric is decoupled into submetrics such that only the members of the same group need to be decoded jointly .
it can be formally defined as @xcite , @xcite , @xcite : [ multi_group_decodability ] an stbc is said to be @xmath115-group decodable if there exists a partition of @xmath116 into @xmath115 non - empty subsets @xmath117 such that the following condition is satisfied : @xmath118 whenever @xmath119 and @xmath120 and @xmath121 .
if we group all the variables of the same group together in , then the @xmath63 matrix for the sd @xcite , @xcite in case of multi - group decodable codes will be of the following form : @xmath122,\ ] ] where @xmath123 is a square upper triangular matrix .
now , consider the standard sd of an stbc .
suppose the @xmath63 matrix as defined in turns out to be such that when we fix values for a set of symbols , the rest of the symbols become group decodable , then the code is said to be fast decodable .
formally , it is defined as follows : [ fast_decodability ] an stbc is said to be fast sd if there exists a partition of @xmath124 where @xmath125 into @xmath115 non - empty subsets @xmath117 such that the following condition is satisfied @xmath126 whenever @xmath127 and @xmath128 and @xmath129 where @xmath130 and @xmath131 are obtained from the @xmath100 decomposition of the equivalent channel matrix @xmath106 = \textbf{q}\textbf{r}$ ] with @xmath107 as column vectors and @xmath132 $ ] with @xmath110 as column vectors as defined in .
hence , by conditioning @xmath133 variables , the code becomes @xmath115-group decodable . as a special case ,
when no conditioning is needed , i.e. , @xmath134 , then the code is @xmath115-group decodable .
the @xmath63 matrix for fast decodable codes will have the following form : @xmath135,\ ] ] where @xmath136 is an @xmath137 block diagonal , upper triangular matrix , @xmath138 is a square upper triangular matrix and @xmath139 is a rectangular matrix .
fast group decodable codes were introduced in @xcite .
these codes combine the properties of multi - group decodable codes and the fast decodable codes .
these codes allow each of the groups in the multi - group decodable codes to be fast decoded .
the @xmath63 matrix for a fast group decodable code will have the following form : @xmath140,\ ] ] where each @xmath141 will have the following form : @xmath142,\ ] ] where @xmath143 is an @xmath144 block diagonal , upper triangular matrix , @xmath145 is a square upper triangular matrix and @xmath146 is a rectangular matrix .
the structure of the @xmath63 matrix for each of the codes defined above depends upon the ordering of the weight matrices .
if we change the ordering of the weight matrices , the @xmath63 matrix may lose its structure and no longer exhibit the desirable decoding properties .
the silver code of example [ silver_code ] illustrates this aspect . in the following subsection
, we give a formal definition of the fsd complexity . in this section
we define the fsd complexity of an stbc .
first we consider a single group decodable case and define the fsd complexity for a particular ordering .
we then extend the definition to a multi - group decodable stbc .
let @xmath10 be an stbc with the weight matrices @xmath147 where all the variables take values from a signal set of cardinality @xmath61 .
let @xmath63 be the matrix obtained by the @xmath148 decomposition used for sd .
let the ordering of the weight matrices used to obtain the @xmath149 matrix be @xmath150 . denote by @xmath151 the number of variables that need to be conditioned , when we use fsd on the @xmath149 matrix . after conditioning @xmath151 variables , let the rest of the variables be @xmath152-group decodable . in case
the code is not fast sphere decodable , then we set @xmath151 to the total number of variables in the matrix .
denote by @xmath153 the submatrix containing the variables of the @xmath154-th group after conditioning @xmath151 variables in the @xmath149 matrix and by @xmath155 the number of variables in the @xmath154-th group , where @xmath156 .
we can use fsd on each of the @xmath153 matrices now .
let us denote by @xmath157 the number of variables that need to be conditioned , when we use fsd on the @xmath153 matrix .
after conditioning @xmath157 variables in @xmath153 , let the rest of the variables be @xmath158-group decodable .
let @xmath159 .
denote by @xmath160 the submatrix containing the variables of the @xmath161-th group after conditioning @xmath157 variables in the @xmath153 matrix and by @xmath162 the number of variables in the @xmath161-th group , where @xmath163 .
it may so happen that @xmath164 for some @xmath154 .
in such cases we set @xmath165 for @xmath166 . we continue this process till we can not apply fsd any more .
this process terminates since there are a finite number of variables and the number of variables are decreasing with each iteration of the fsd .
let the process stop after @xmath167 such iterations . in the @xmath168 -th iteration ,
we have the following : let us denote by @xmath169 the number of variables that need to be conditioned , when we use fsd on the @xmath170 matrix .
after conditioning @xmath169 variables in @xmath170 , let the rest of the variables be @xmath171-group decodable .
let @xmath172 .
denote by @xmath173 the submatrix containing the variables of the @xmath174-th group after conditioning @xmath169 variables in the @xmath170 matrix and by @xmath175 the number of variables in the @xmath174-th group where @xmath176 . on the last iteration , we have @xmath177 = @xmath175 for @xmath176 .
let @xmath178 for @xmath179 , @xmath180 and @xmath181 and @xmath182 .
[ fsd_complexity_def1 ] we define the fsd complexity of a single group decodable stbc for the given ordering to be @xmath183 . in case of a multi - group decodable code with @xmath115 groups ,
the @xmath149 matrix will be a block diagonal matrix .
we then calculate the fsd complexity of each group independently as described above and choose the maximum among them as the fsd complexity of the stbc .
[ fsd_complexity_def2 ] we define the fsd complexity of a multi - group decodable stbc with @xmath115 groups to be @xmath184 where @xmath185 .
we present a few examples to get a better understanding of fsd complexity .
the fsd complexity of this stbc for the given ordering is @xmath214 .
in this section we show that the hrqf matrix is enough to determine the fsd complexity of an stbc and hence the fsd complexity is independent of the channel matrix realization or the number of receive antennas . towards this end , we prove that the zeros in the @xmath149 matrix which determine the fsd complexity are also zeros in the hrqf matrix .
first we define an ordered partition of a set .
[ ordered_partition_def ] we call a partition of @xmath215 into @xmath115 non - empty subsets @xmath117 with cardinalities @xmath216 an ordered partition if @xmath217 , @xmath218 and so on , till @xmath219 now we address the class of multi - group decodable codes .
[ hrqf_multi_group_lemma ] consider an stbc @xmath220 .
let @xmath221 denote the hrqf matrix of this stbc .
if there exists an ordered partition of @xmath116 into @xmath115 non - empty subsets @xmath117 such that @xmath50 whenever @xmath127 and @xmath128 and @xmath129 , then the code is @xmath115-group sphere decodable . in other words , the fsd complexity of the stbc
is determined by the hrqf matrix .
let @xmath149 be the matrix obtained from the qr decomposition of @xmath101 .
for the code to be @xmath115-group sphere decodable , we need to prove that @xmath222 , whenever @xmath127 and @xmath128 and @xmath129 .
we know from @xcite that if @xmath51 is satisfied for some @xmath223 then the corresponding columns in the @xmath224 matrix are orthogonal , i.e. , @xmath225 .
we also know that @xmath50 if and only if @xmath51 . let @xmath226 where @xmath227 and @xmath228 . for any group @xmath229 , we need to prove that @xmath222 for @xmath230 and @xmath231 .
consider the first group @xmath232 .
we have @xmath50 for @xmath233 and @xmath234 .
we need to prove that the @xmath63 matrix has zero entries at the same locations .
the proof for this is by induction . + for @xmath235 and for any @xmath236 , @xmath237 since @xmath238 .
now , let @xmath239 for all @xmath240 for any @xmath241 such that @xmath233 .
we have , @xmath242\\ % \end{equation * } % \begin{equation * } & = \frac{1}{\parallel \textbf{r}_{i}\parallel}\left [ \langle \textbf{h}_{i } , \textbf{h}_{j } \rangle - \sum_{l=1}^{i-1 } \langle \textbf{q}_{l } , \textbf{h}_{i}\rangle \langle \textbf{q}_{l } , \textbf{h}_{j } \rangle \right ] = 0,\end{aligned}\ ] ] since @xmath225 as @xmath243 and @xmath244 for @xmath240 , by induction hypothesis . now consider the @xmath245-th group @xmath246 .
let the induction hypothesis be true for all groups @xmath247 .
consider @xmath248 where @xmath230 and @xmath231 .
we have , @xmath249 \\ % \end{equation * } % \begin{equation * } & = \frac{1}{\parallel \textbf{r}_{i}\parallel}\left [ \langle \textbf{h}_{i } , \textbf{h}_{j } \rangle - \sum_{l=1}^{i-1 } \langle \textbf{q}_{l } , \textbf{h}_{i}\rangle \langle \textbf{q}_{l } , \textbf{h}_{j } \rangle \right ] = 0 , \end{aligned}\ ] ] since @xmath225 as @xmath243 and @xmath244 for @xmath240 by the induction hypothesis .
we now consider an example to illustrate the above lemma .
now we move on the class of fast decodable codes .
[ hrqf_fast_decode_lemma ] consider an stbc @xmath220 .
let @xmath221 denote the hrqf matrix of this stbc .
if there exists an ordered partition of @xmath124 where @xmath125 into @xmath115 non - empty subsets @xmath117 such that @xmath50 whenever @xmath127 and @xmath128 and @xmath129 , then the code is fast decodable or conditionally @xmath115-group decodable .
the proof follows from the proof of lemma [ hrqf_multi_group_lemma ] by replacing @xmath258 with @xmath259 .
we now consider an example to illustrate the above lemma .
[ hrqf_fast_decode_ex ] consider the silver code as mentioned in example [ silver_code ] .
if we order the variables ( and hence the weight matrices ) in the following fashion @xmath260 $ ] , then the hrqf matrix @xmath49 and the @xmath63 matrix will have the following structure : @xmath261,\ ] ] @xmath64,\ ] ] where @xmath65 denotes the non - zero entries .
as it can be seen , the upper triangular portion of the matrix @xmath49 , has a structure that admits fast decodability which is conditionally @xmath202-group decodable if considered as the @xmath63 matrix .
we now turn to the class of fast group decodable codes .
[ hrqf_fast_group_decode_lemma ] consider an stbc @xmath220 .
let @xmath221 denote the hrqf matrix of this stbc .
if there exists an ordered partition of @xmath116 into @xmath115 non - empty subsets @xmath117 with cardinalities @xmath216 such that @xmath50 whenever @xmath127 and @xmath128 and @xmath129 , and if any group @xmath262 admits fast decodability , i.e. , there exists an ordered partition of @xmath263 where @xmath264 , into @xmath265 non - empty subsets @xmath266 such that @xmath267 whenever @xmath268 and @xmath269 and @xmath129 , @xmath270 , then the code is fast group decodable .
the proof follows from the proofs of lemmas [ hrqf_multi_group_lemma ] and [ hrqf_fast_decode_lemma ] .
we now consider an example to illustrate the above lemma .
[ hrqf_fast_group_decode_ex ] consider the fast group decodable stbc @xcite given in .
@xmath271\ ] ] ' '' '' let the ordering of the variables ( and hence the weight matrices ) be @xmath272 $ ] .
this stbc is two group decodable with @xmath273 in one group and @xmath274 in the other .
the second group is conditionally five group decodable .
the hrqf matrix @xmath49 and the @xmath63 matrix are given in and respectively , where @xmath65 denotes the non zero entries .
@xmath275\ ] ] @xmath276\ ] ] as we have seen from lemmas [ hrqf_multi_group_lemma ] , [ hrqf_fast_decode_lemma ] and [ hrqf_fast_group_decode_lemma ] , the fsd complexity of the stbc depends only upon the hrqf matrix @xmath49 and not on the @xmath224 matrix , i.e. , the fsd complexity is independent of the channel matrix and the number of receive antennas .
it can be completely captured into a single matrix obtained from the set of weight matrices and their ordering .
as seen in example [ silver_code ] , the ordering of weight matrices determines the fsd complexity of an stbc .
we have also seen that the hrqf matrix completely determines the fsd complexity of an stbc . in this section
we present an algorithm that uses the hrqf matrix as an input and manipulates it in order to obtain a best possible ordering of weight matrices .
we do so by using row and column permutations of the hrqf matrix .
the rows and columns of the hrqf matrix are in one to one correspondence with the ordering of the weight matrices .
hence , if we change the ordering of the weight matrices , the hrqf matrix changes accordingly and vice verse .
for example , any transposition in the ordering of the weight matrices will result in swapping the corresponding rows and columns ( since hrqf matrix is symmetric ) of the hrqf matrix .
[ hrqf_remark ] note that we can not perform such a manipulation on the @xmath149 matrix since it depends not only on the order of weight matrices but on the channel matrix as well .
also , all the entries of the @xmath149 matrix do not depict the hr orthogonality of the weight matrices , i.e. , the @xmath277-th entry of the matrix may not be zero even if the @xmath241-th and @xmath278-th weight matrices are hr orthogonal .
hence , the @xmath149 matrix needs to be calculated each time the ordering of the weight matrices is changed which is not so in the case of the hrqf matrix .
the algorithm to get a best possible ordering is given in algorithm [ alg1:alg1 ] in the next page .
[ alg1:alg1 ] - @xmath279 - @xmath280 ; @xmath281 before we delve into the proof of correctness of the algorithm , let us make a few observations regarding the structure of the hrqf matrix for various scenarios of decoding .
let @xmath10 be an stbc with weight matrices @xmath150 and corresponding variables @xmath282 .
the code @xmath10 can be multi - group decodable , fast - decodable , fast group decodable or none of these .
the structure of the upper triangular portion of the hrqf matrix for the first three of these are shown in the equations , and respectively . in case
the code does not allow any of these forms of decoding , the ordering of the variables is immaterial .
let @xmath136 be a best ordering of the variables .
without loss of generality , let the first variable in this ordering be labelled @xmath283 .
irrespective of the type of decoding provided by the code , we notice that all the variables that need to be jointly decoded with @xmath283 are adjacent to it in the hrqf matrix .
these are followed by the set of zeros which indicate all the variables which are hurwitz - radon orthogonal with the previous set .
these are further followed by the variables that are to be conditioned in order to obtain this group - decodable structure . in case of multi - group decodable codes , this is a null set . _
proof of correctness : _ we need to show that the algorithm produces a best possible ordering for fsd complexity .
the algorithm uses the structural property of the hrqf matrix for obtaining a best ordering .
given any ordering @xmath284 , with the first variable as @xmath283 , the algorithm partitions the variables into three sets - @xmath285 and @xmath286 . in case of multi - group decoding and fast group decoding
, @xmath287 contains all the variables that belong to the same group as @xmath283 . in case of
fast decoding , it represents the set of variables that need to be jointly decoded with @xmath283 after conditioning .
@xmath288 contains all the variables hr orthogonal with @xmath287 .
@xmath286 contains the variables that need to be conditioned .
this will be an empty set in case of multi - group decoding and hence multi - group decoding can be considered as fast decoding with no conditioned variables .
the algorithm is recursively run on the sets @xmath287 and @xmath288 to further order the variables ( as in case of fast group decoding ) .
first , we fix the first variable in the given ordering .
let it be @xmath98 for some @xmath289 .
the algorithm starts with only @xmath98 in @xmath287 , all variables which are hr orthogonal with @xmath98 in a temporary set @xmath290 and the rest of the variables in @xmath286.the variables hr orthogonal with @xmath98 can be easily identified as they correspond to the zero entries in the first row . for ease of manipulation , we move all the variables in the set @xmath290 adjacent @xmath98 in the ordering .
this is equivalent to grouping all the zeros in the first row and placing them adjacent to the ( 1,1 ) entry , as these denote all the variables hr orthogonal with @xmath98 .
the current ordering of variables is - @xmath291 . for the working of the algorithm ,
the following set operations are synonymous with the following matrix operations on the hrqf matrix : * moving a variable @xmath292 into @xmath287 - suppose the variable @xmath292 is the @xmath245-th element in the current ordering , we move the @xmath245-th column to the @xmath293-th column shifting the rest of the columns to the right and then we move the @xmath245-th row to the @xmath293-th row shifting the rest of the rows downwards .
update the ordering accordingly . * moving a variable @xmath292 into @xmath286 - suppose the variable @xmath292 is the @xmath245-th element in the current ordering , we move the @xmath245-th column to the last column shifting the rest of the columns to the left and then we move the @xmath245-th row to the last row shifting the rest of the rows upwards .
update the ordering accordingly .
* moving a variable @xmath292 into @xmath288 - no change .
only the cardinalities of @xmath288 and @xmath290 will change accordingly .
the algorithm works in two stages . *
first , we find the largest @xmath125 such that @xmath294 * in the second stage , since the variables in @xmath287 and @xmath288 will be decoded separately , we consider the submatrices representing them as hrqf matrices of some stbc and run the first step of the algorithm on them recursively .
let @xmath295 be the first variable in @xmath290 .
it is currently the second variable in the overall ordering .
we now proceed to find all the variables that are hr orthogonal with @xmath295 and not hr orthogonal with @xmath287 , since these will need to be jointly decoded with the variables in @xmath287 .
this can be found as follows .
if any variable is hr orthogonal with @xmath295 , it will have a zero entry in the column corresponding to @xmath295 .
hence we traverse down the column represented by @xmath295 to find the next zero entry .
let it be found in the @xmath245-th row corresponding to the variable @xmath296 .
we also need to ensure that this variable is not hr orthogonal with @xmath287 .
so , we traverse the @xmath245-th row from column 1 to column @xmath297 and check for any non - zero entries . in case any of them are found , it means that this variable needs to be jointly decoded with @xmath287 .
we add @xmath296 to @xmath287 .
we now repeat the procedure on the variable @xmath295 again until all the variables have been exhausted . since there are a finite number of variables , this process terminates .
note that it is possible that a few members from @xmath290 may move into @xmath287 during this process .
these variables will be accounted for later .
we now add @xmath295 to @xmath288 .
we have now managed to create two sets which are hr orthogonal and the current ordering of the variables is @xmath298 .
we now proceed to the next variable in @xmath290 .
let this be @xmath292 .
we need to ensure that @xmath292 is hr orthogonal with @xmath287 .
this can be easily checked by counting the consecutive number of zeros in its column from the top of the column .
if it is equal to @xmath297 , then @xmath292 is hr orthogonal with @xmath287 .
we add @xmath292 to @xmath288 in this case . otherwise , we move @xmath292 to @xmath286
. it may so happen that @xmath292 may be hr orthogonal with @xmath287 and the already existing members of @xmath288 .
this can be checked by counting the consecutive numbers of zeros in its column from the top of the column and matching cardinalities .
in such a case , we repeat the procedure of finding any variable that needs to be jointly decoded with @xmath287 on this variable as well .
if any such variable exists , we move that variable and the pre - existing variables of @xmath288 into @xmath287 .
this is needed for fast group decoding scenarios .
we continue this procedure till we have exhausted all the variables of @xmath290 .
we have now produced three sets @xmath287 , @xmath288 and @xmath286 .
@xmath287 and @xmath288 are hr orthogonal w.r.t .
one another if we condition them on the variables of @xmath286 .
we still need to process @xmath287 and @xmath288 further .
we repeat the same procedure on these variables considering them as independent hrqf matrices .
if the cardinality of @xmath287 is the same as that of the hrqf matrix row / column size , then we stop this procedure as it can not be ordered any further and set the fsd complexity to @xmath297 .
this process terminates as the sizes of @xmath287 and @xmath288 keep reducing with each iteration and there are a finite number of variables .
we calculate the fsd complexity of this ordering as - @xmath299 where @xmath300 and @xmath301 are the fsd complexities of @xmath287 and @xmath288 respectively .
since we are dealing with fast sphere decoding and grouping the variables as sets @xmath285 and @xmath286 are the only possible ways to obtain a structure that will allow fs decoding , these are all the orderings we need to consider if we take @xmath98 as the first variable .
hence , we have exhausted all the possible orderings which can provide fs decoding with @xmath98 as the first variable , and a best possible fsd complexity for this case has been found by the above steps .
now , if we run through all the variables of the stbc making each one of them the first variable in turn and repeating the above procedure , we can find out a best possible fsd complexity for all possible orderings .
since all the variables have been given a chance to be the first variable , we have exhausted all possible orderings that offer fs decoding .
and hence , the ordering provided by the algorithm is a best possible ordering for fsd complexity .
@xmath302 [ hrqf_alg_remark ]
since the algorithm recursively orders each set , it is capable of even ordering variables in scenarios where any group obtained from conditioning of variables admits conditional decoding .
we now illustrate the working of the algorithm with an example .
[ algo_fast_decode ] consider the silver code presented in example [ silver_code ] .
if we order the variables as @xmath303 $ ] , we get the following hrqf matrix and the @xmath149 matrix for this ordering : @xmath304;\ ] ] @xmath305.\ ] ] the fsd complexity for this ordering is @xmath306 . and
this ordering does not admit fast decoding as well .
when we run the algorithm on the given hrqf matrix , the two sets @xmath287 and @xmath288 are formed , which are hr orthogonal with each other . in this case ,
@xmath307 and @xmath308 .
the conditioned variables will be present in the set @xmath309 .
the hrqf matrix at this stage is as given by .
the ordering of the variables at the end of this stage is @xmath310 $ ] .
now , the variables from both sets @xmath287 and @xmath288 are run through the algorithm again .
so , the top left @xmath311 matrix and the next block diagonal @xmath312 matrix are both fed to the algorithm . since this is already the best possible ordering of these sets ,
the matrix @xmath49 remains the same after this stage . and
the final ordering obtained is @xmath310 $ ] .
the @xmath63 matrix for this ordering is given by @xmath313\ ] ] @xmath314\ ] ] the fsd complexity for this @xmath63 matrix is @xmath66 which is the best possible complexity for the silver code .
[ algo_multi_group ] consider the fast group decodable code presented in example [ hrqf_fast_group_decode_ex ] .
if we order the variables as @xmath315 $ ] , we get the following hrqf matrix and the @xmath149 matrix : @xmath316,\ ] ] @xmath317.\ ] ] the fsd complexity for this @xmath63 matrix is @xmath318 but the best possible fsd complexity is @xmath319 .
when we run the algorithm on the given hrqf matrix , the two sets @xmath287 and @xmath288 are formed , which are hr orthogonal with each other . in this case , @xmath320 and @xmath321 . the set @xmath286 is empty as this provides a group decoding scenario .
now , the variables from both sets are run through the algorithm again .
so , the top left @xmath322 matrix and the bottom right @xmath311 matrix are both fed to the algorithm .
the top left matrix is ordered according to the fast decoding algorithm as presented in the previous example .
since the bottom right matrix is a @xmath311 matrix , it is returned without change .
the final ordering of variables obtained is @xmath323 $ ] .
the @xmath221 and the @xmath149 matrix for this ordering is as shown below .
@xmath324,\ ] ] @xmath325.\ ] ] this ordering gives us the fsd complexity of @xmath319 .
in this paper we have analysed the fsd complexity of an stbc using quadratic forms .
we have shown that the hrqf completely categorizes the fsd complexity of an stbc and hence it is independent of the channel and the number of receive antennas .
we have provided an algorithm to obtain a best ordering of weight matrices to get the best decoding performance from the code .
this work was supported partly by the drdo - iisc program on advanced research in mathematical engineering through a research grant , and partly by the inae chair professorship grant to b. s. rajan .
k. p. srinath and b. s. rajan , low ml - decoding complexity , large coding gain , full - rate , full - diversity stbcs for 2x2 and 4x2 mimo systems , ieee journal of selected topics in signal processing : special issue on managing complexity in multiuser mimo systems , vol .
3 , no . 6 , pp .
916 - 927 , dec . 2009 .
t. p. ren , y. l. guan , c. yuen and r. j. shen , fast - group - decodable space - time block code ,
proceedings ieee information theory workshop , ( itw 2010 ) , cairo , egypt , jan .
6 - 8 , 2010 , available online at http://www1.i2r.a-star.edu.sg/cyuen/publications.html .
o. tirkkonen , a. boariu and a. hottinen , minimal non - orthogonality rate 1 space - time block code for 3 + tx antennas , proceedings of ieee international symposium on spread - spectrum techniques and applications , new jersey , pp .
429 - 432 , sep .
6 - 8 , 2000 . | decoding of linear space - time block codes ( stbcs ) with sphere - decoding ( sd ) is well known . a fast - version of the sd known as fast sphere decoding ( fsd ) has been recently studied by biglieri , hong and viterbo . viewing a linear stbc as a vector space spanned by its defining weight matrices over the real number field
, we define a quadratic form ( qf ) , called the hurwitz - radon qf ( hrqf ) , on this vector space and give a qf interpretation of the fsd complexity of a linear stbc .
it is shown that the fsd complexity is only a function of the weight matrices defining the code and their ordering , and not of the channel realization ( even though the equivalent channel when sd is used depends on the channel realization ) or the number of receive antennas .
it is also shown that the fsd complexity is completely captured into a single matrix obtained from the hrqf . moreover , for a given set of weight matrices , an algorithm to obtain a best ordering of them leading to the least fsd complexity is presented . the well known classes of low fsd complexity codes ( multi - group decodable codes , fast decodable codes and fast group decodable codes ) are presented in the framework of hrqf . |
given integer @xmath0 and a polynomial @xmath1 $ ] and some @xmath2^{d}$ ] , we let @xmath3 denote the number of solutions @xmath4 to the congruence @xmath5 questions concerning the distribution of solutions to polynomial congruences have been considered in a number of works ( for example @xcite ) . in @xcite fouvry gives an asymptotic formula for the number of solutions to systems of polynomial congruences in small cubic boxes for a wide class of systems ( see also @xcite ) .
shparlinski @xcite uses the results of @xcite and ideas of @xcite to obtain an asymptotic formula for the number of solutions to the same systems when the solutions lie in a very general class of sets . for the case of a single polynomial , @xmath6 in @xmath7 variables , shparlinski
@xcite shows that for suitable @xmath8 , when the modulus @xmath9 is prime , @xmath10 provided @xmath6 is irreducible over @xmath11 and @xmath12 denotes the lebesgue measure on @xmath13^d$ ] .
this gives an asymptotic formula for @xmath3 provided @xmath14 and a nontrivial upper bound for @xmath3 when @xmath15 .
we follow the method of @xcite to give an upper bound for @xmath3 without any restrictions on our polynomial @xmath6 when the modulus @xmath0 is composite .
we first establish an upper bound for @xmath3 when @xmath8 is a cube .
this gives a generalization of theorem @xmath16 of @xcite .
although we follow the same argument , the difference is our use of a multidimensional version of vinogradov s mean value theorem ( theorem 1.1 of @xcite ) . to extend the bound from cubes to more general sets @xmath8
, we approximate @xmath8 by cubes using ideas based on theorem @xmath17 of @xcite .
we let @xmath12 denote the lebesgue measure on @xmath13^d$ ] , @xmath18 the euclidian norm and define the distance between @xmath19^d$ ] and @xmath20^d$ ] to be @xmath21 as in @xcite , we say that @xmath2^{d}$ ] is _ well - shaped _ if there exists @xmath22 such that for every @xmath23 the measures of the sets @xmath24^d \backslash \omega \ : \
\dist(\vec{u},\omega ) < \varepsilon \right\},\ ] ] @xmath25^d \backslash \omega ) < \varepsilon \right\}\ ] ] exist and satisfy @xmath26 from lemma 1 of @xcite all convex subsets of @xmath13^d$ ] are well - shaped and from equation @xmath27 of @xcite , if the boundary of @xmath8 is a manifold of dimension @xmath28 with bounded surface area then @xmath8 is well - shaped , for suitably chosen @xmath29 + for @xmath30 we write @xmath31 if @xmath32 . given a @xmath7-tuple of non - negative integers @xmath33 , we set @xmath34 and @xmath35 .
we let @xmath36 denote the number of distinct @xmath7-tuples , @xmath37 with @xmath38 so that @xmath39 we will always suppose @xmath0 is an integer greater than @xmath17 .
given @xmath40 $ ] , we let @xmath41 denote the degree of @xmath6 and @xmath7 the number of variables . writing @xmath42
we define @xmath43 we use @xmath44 and @xmath45 to mean that there exists some absolute constant @xmath46 such that @xmath47 for all values of @xmath48 within some specified range . whenever we use @xmath49 and @xmath50 , unless stated otherwise the implied constant will depend only on @xmath7 , @xmath41 and the particular @xmath51 in @xmath52 similarily @xmath53 denotes a term which is sufficiently small when our parameter is large enough in terms of @xmath7 , @xmath41 and @xmath51 .
we can now present our main results : [ multi ] for positive @xmath54 , integer @xmath0 and @xmath55\ ] ] of degree @xmath56 with @xmath57 , let @xmath58 denote the number of solutions to the congruence @xmath59 with @xmath60\times \dots \times [ k_d+1,k_d+h]\times [ l+1,l+r].\ ] ] then uniformly over all @xmath61 @xmath62 as @xmath63 arguing from heuristics , we expect the bound for @xmath58 to be around @xmath64 which can be directly compared with theorem 1 .
similarly , by considering the first term in theorem 1 we immediatley see when this bound for @xmath58 is worse than the trivial bound @xmath65 .
also , if @xmath9 is prime and @xmath66 $ ] is not multilinear , i.e @xmath6 is not linear in each of its variables , then theorem [ multi ] is trivial .
this may be seen by the following argument .
first we may show by slightly adjusting the proof of theorem 1 of @xcite that for @xmath67 $ ] of degree @xmath68 @xmath69 supposing @xmath70 $ ] is not multilinear , then after re - ordering the variables we may suppose for some @xmath71 that @xmath72=\sum_{i=0}^{k_0}x_{d}^if_{i}[x_1,\dots , x_{d-1}]\ ] ] with @xmath73 and consider separately the values of @xmath74 such that @xmath75\equiv0 \pmod p\ ] ] and @xmath76 \not \equiv 0 \pmod p.\ ] ] for the first case we use the assumption that @xmath77 is prime and induction on @xmath7 to bound the number of values @xmath78 such that @xmath79\equiv0 \pmod p$ ] by @xmath80 and bound the number of solutions to @xmath81\equiv y \pmod p\ ] ] in remaining variables @xmath82 trivially by @xmath83 . for the second case , we bound the number of @xmath74 such that @xmath79 \not \equiv 0 \pmod p$ ] trivially by @xmath84 and bound the number of solutions in the remaining variables @xmath82 by applying to the equation . combining the above two cases
gives @xmath85 which can be compared with theorem 1 .
taking @xmath86 in theorem [ multi ] we get , [ cubes ] for any cube @xmath87^{d}$ ] of side length @xmath88 , @xmath89 $ ] of degree @xmath68 with @xmath57 we have @xmath90 as @xmath91 taking @xmath92 in theorem [ multi ] we get , [ cubes 1 ] suppose @xmath89 $ ] of degree @xmath68 with @xmath57 is of the form , @xmath93 for some @xmath94 $ ] , then for any cube @xmath87^{d}$ ] of side length @xmath88 , we have @xmath95 as @xmath96 , where @xmath36 corresponds to @xmath97 in the definition .
we use the above corollaries to estimate @xmath3 for well - shaped @xmath8 .
[ multi vws ] suppose @xmath89 $ ] satisfies the conditions of corollary [ cubes ] and @xmath98^{d}$ ] is well - shaped with @xmath99 .
then we have @xmath100 as @xmath101 suppose @xmath102 $ ] satisfies the conditions of corollary [ cubes 1 ] and @xmath98^{d}$ ] is well - shaped .
then we have @xmath103 as @xmath104
making a change of variables we may assume @xmath105 .
suppose for integer @xmath106 we have @xmath107 satisfying ( [ polynomial equation 1 ] ) with + @xmath108 then @xmath109 for some @xmath110 hence there exists @xmath111 such that @xmath112 with @xmath113 equal to the number of solutions to the congruence @xmath114 with each co - ordinate of @xmath115 between @xmath16 and @xmath116 + + since @xmath117 we may write ( [ polynomial equation 2 ] ) in the form @xmath118 with @xmath119 since @xmath57 , we choose @xmath120 with @xmath121 and @xmath122 . considering ( [ linear ] ) as a linear equation in @xmath123 , if we let @xmath123 , @xmath124 take arbitrary values then @xmath125 is determined uniquely @xmath126 .
since we have @xmath127 there are at most @xmath128 solutions to ( [ linear ] ) in integer variables @xmath129 with @xmath130 for @xmath131 with each @xmath132 let @xmath133 denote the number of solutions in integers , @xmath123 , to @xmath134 with each @xmath135 having components between @xmath16 and @xmath136 and we write @xmath137 when @xmath138 let @xmath139 be the collection of sets @xmath131 such that @xmath140 and @xmath141 so that the cardinality of @xmath139 is bounded by .
we see that @xmath142 since if @xmath143 is a solution to , then the integers @xmath144 , defined by @xmath145 are a solution to and the @xmath143 are a solution to @xmath146 so if we let @xmath147 , then we see that the solution to , @xmath148 is counted by the term @xmath149 in . by and
, we have @xmath150 for some @xmath151 .
although for any @xmath152 we have the inequality @xmath153 since if we let @xmath154 and @xmath155 then for @xmath123 defined as in we have @xmath156^{r}}\exp\left(2\pi i \displaystyle\sum_{1\leq |\mathbf{i}|\leq k}\alpha_{\mathbf{i } } ( \lambda_{\mathbf{i } } -u_{\mathbf{i}})\right)d\boldsymbol{\alpha } \\ & = \displaystyle\int_{[0,1]^{r}}|s(\boldsymbol{\alpha})|^{2s}\exp\left(-2\pi i\displaystyle\sum_{1\leq |\mathbf{i}| \leq k}\alpha_{\mathbf{i}}u_{\mathbf{i}}\right)d\boldsymbol{\alpha } \\ & \leq \displaystyle\int_{[0,1]^{r}}|s(\boldsymbol{\alpha})|^{2s}d\boldsymbol{\alpha}=j_{s , k , d}(h)\end{aligned}\ ] ] where the integral is over the variables @xmath157 , @xmath158 .
hence by and we have @xmath159 by theorem 1.1 of @xcite we have for @xmath160 @xmath161 for any @xmath162 provided @xmath136 is sufficiently large in terms of @xmath163 and @xmath106 . inserting this bound into gives @xmath164 and the result follows taking @xmath165 .
as in @xcite we begin with choosing @xmath166 with each co - ordinate irrational .
for integer @xmath167 let @xmath168 be the set of cubes of the form @xmath169\times \dots \times \left[a_{d}+\frac{u_{d}}{j},a_{d}+\frac{u_{d}+1}{j}\right ] , \ \ \
u_i \in \mathbb{z}.\ ] ] since each @xmath170 is irrational , no point ( [ solutions ] ) lies in two distinct cubes ( [ cuubes ] ) .
given integer @xmath171 , let @xmath172 and consider the set @xmath173 since @xmath8 is well - shaped , we have @xmath174 let @xmath175 be the cubes of @xmath168 lying inside @xmath176 and we suppose @xmath177 then by ( [ omega big ] ) we obtain , @xmath178 ( -0.80,0.8 ) ellipse ( 6.77 and 8.8 ) ; ( -0.80,0.8 ) ellipse ( 5.35 and 7.38 ) ; ( -4,-4 ) rectangle ( 4,4 ) ; ( -4,4 ) rectangle ( 0,8 ) ; ( -4,-4 ) rectangle ( -2,-6 ) ; ( 2,-4 ) rectangle ( 4,-6 ) ; ( 0,4 ) rectangle ( 2,6 ) ; ( 2,4 ) rectangle ( 4,6 ) ; ( 0,6 ) rectangle ( 2,8 ) ; ( 4,-2 ) rectangle ( 6,0 ) ; ( 0,-4 ) rectangle ( -2,-6 ) ; ( 2,-4 ) rectangle ( 0,-6 ) ; ( -4,-2 ) rectangle ( -6,0 ) ; ( -4,0 ) rectangle ( -6,2 ) ; ( -4,2 ) rectangle ( -6,4 ) ; ( -4,4 ) rectangle ( -6,6 ) ; ( -1,-6 ) rectangle ( 0,-7 ) ; ( 0,-6 ) rectangle ( 1,-7 ) ; ( -4,6 ) rectangle ( -5,7 ) ; ( 4,2 ) rectangle ( 5,3 ) ; ( 4,1 ) rectangle ( 5,2 ) ; ( 4,0 ) rectangle ( 5,1 ) ; ( 4,-3 ) rectangle ( 5,-2 ) ; ( 4,-4 ) rectangle ( 5,-3 ) ; ( -5,-4 ) rectangle ( -4,-3 ) ; ( -5,-3 ) rectangle ( -4,-2 ) ; ( -6,-3 ) rectangle ( -5,-2 ) ; ( -7,0 ) rectangle ( -6,1 ) ; ( -7,1 ) rectangle ( -6,2 ) ; ( -7,2 ) rectangle ( -6,3 ) ; ( -7,3 ) rectangle ( -6,4 ) ; ( 1,-6 ) rectangle ( 2,-7 ) ; ( -6,4 ) rectangle ( -7,5 ) ; ( -6,-1 ) rectangle ( -7,0 ) ; ( -2,-6 ) rectangle ( -1,-7 ) ; ( 2,6 ) rectangle ( 3,7 ) ; ( 4,3 ) rectangle ( 5,4 ) ; ( 5,0 ) rectangle ( 6,1 ) ; ( -4,8 ) rectangle ( -3,9 ) ; ( -5,-5 ) rectangle ( -4,-4 ) ; ( -6,6 ) rectangle ( -5,7 ) ; ( -5,7 ) rectangle ( -4,8 ) ; ( -7,5 ) rectangle ( -6,6 ) ; ( -3,8 ) rectangle ( -2,9 ) ; ( -2,8 ) rectangle ( -1,9 ) ; ( -1,8 ) rectangle ( 0,9 ) ; ( -7,-1 ) rectangle ( -6,-2 ) ; ( 2,-6 ) rectangle ( 3,-7 ) ; ( 4,-5 ) rectangle ( 5,-4 ) ; ( -4,-4 ) rectangle ( 4,4 ) ; ( -4,4 ) rectangle ( 0,8 ) ; ( -4,-4 ) rectangle ( -2,-6 ) ; ( 2,-4 ) rectangle ( 4,-6 ) ; ( 0,4 ) rectangle ( 2,6 ) ; ( 2,4 ) rectangle ( 4,6 ) ; ( 0,6 ) rectangle ( 2,8 ) ; ( 4,-2 ) rectangle ( 6,0 ) ; ( 0,-4 ) rectangle ( -2,-6 ) ; ( 2,-4 ) rectangle ( 0,-6 ) ; ( -4,-2 ) rectangle ( -6,0 ) ; ( -4,0 ) rectangle ( -6,2 ) ; ( -4,2 ) rectangle ( -6,4 ) ; ( -4,4 ) rectangle ( -6,6 ) ; ( -1,-6 ) rectangle ( 0,-7 ) ; ( 0,-6 ) rectangle ( 1,-7 ) ; ( -4,6 ) rectangle ( -5,7 ) ; ( 4,2 ) rectangle ( 5,3 ) ; ( 4,1 ) rectangle ( 5,2 ) ; ( 4,0 ) rectangle ( 5,1 ) ; ( 4,-3 ) rectangle ( 5,-2 ) ; ( 4,-4 ) rectangle ( 5,-3 ) ; ( -5,-4 ) rectangle ( -4,-3 ) ; ( -5,-3 ) rectangle ( -4,-2 ) ; ( -6,-3 ) rectangle ( -5,-2 ) ; ( -7,0 ) rectangle ( -6,1 ) ; ( -7,1 ) rectangle ( -6,2 ) ; ( -7,2 ) rectangle ( -6,3 ) ; ( -7,3 ) rectangle ( -6,4 ) ; ( 1,-6 ) rectangle ( 2,-7 ) ; ( -6,4 ) rectangle ( -7,5 ) ; ( -6,-1 ) rectangle ( -7,0 ) ; ( -2,-6 ) rectangle ( -1,-7 ) ; ( 2,6 ) rectangle ( 3,7 ) ; ( 4,3 ) rectangle ( 5,4 ) ; ( 5,0 ) rectangle ( 6,1 ) ; ( -4,8 ) rectangle ( -3,9 ) ; ( -5,-5 ) rectangle ( -4,-4 ) ;
( -6,6 ) rectangle ( -5,7 ) ; ( -5,7 ) rectangle ( -4,8 ) ; ( -7,5 ) rectangle ( -6,6 ) ; ( -3,8 ) rectangle ( -2,9 ) ; ( -2,8 ) rectangle ( -1,9 ) ; ( -1,8 ) rectangle ( 0,9 ) ; ( -7,-1 ) rectangle ( -6,-2 ) ; ( 2,-6 ) rectangle ( 3,-7 ) ; ( 4,-5 ) rectangle ( 5,-4 ) ; also , since a cube of side length @xmath179 has diameter @xmath180 , we see that the cubes of @xmath175 cover @xmath181 and hence @xmath182 but for @xmath183 , we have @xmath184 and since @xmath8 is well - shaped @xmath185 so we get @xmath186 combining this with ( [ c upper bound ] ) gives @xmath187 let @xmath188 and for @xmath189 we let @xmath190 be the set of cubes from @xmath191 that are not contained in any cubes from @xmath192 then we have @xmath193 and for @xmath194 the cubes from both @xmath190 and @xmath195 are contained in @xmath176 .
this gives @xmath196 and hence by ( [ c bound ] ) @xmath197 we have @xmath198 since if @xmath199 then @xmath200^d \setminus \omega_\varepsilon)\geq \varepsilon.\ ] ] although @xmath201 for some @xmath202 and since @xmath203 has diameter @xmath204 we have @xmath205 since the union of the cubes from @xmath195 is contained in the union from @xmath206 we get ( [ omega in ] ) . hence @xmath207 and using corollary [ cubes ] , as @xmath208 @xmath209 we use ( [ omega big ] ) to bound @xmath210 and from ( [ b bound ] ) , for @xmath211 @xmath212 hence we get @xmath213 recalling that @xmath214 to balance the two terms involving @xmath215 , we choose @xmath216 substituting this choice into ( [ optimize ] ) gives , @xmath217 the same choice for @xmath218 is essentially optimal , @xmath219 this gives @xmath100 where we have replaced @xmath220 with @xmath221 since @xmath222 theorem 3.4 follows since for the choice of @xmath218 in ( [ m choice ] ) , for @xmath99 @xmath223 which tends to infinity as @xmath104 +
using the same constructions from theorem 3.5 , we have @xmath224 hence by corollary [ cubes 1 ] @xmath225 for the first sum by , @xmath226 for the second sum , @xmath227 substituting the above bounds into gives @xmath228 for @xmath229 we choose @xmath215 to balance the first and last terms then choose @xmath218 to balance the remaining terms , so that @xmath230 @xmath231 which gives @xmath232 and @xmath233 if @xmath234 then we choose @xmath215 to balance the last two terms and take @xmath218 as small as possible subject to the condition @xmath232 .
this gives @xmath235 @xmath236 and @xmath237 combining the above two bounds completes the proof .
using the methods of theorem 3.4 and theorem 3.5 , we have not been able to to give bounds for @xmath3 which are nontrivial when @xmath238 .
this seems to be caused by two factors , the bound from corollary [ cubes ] and the bounds for @xmath239 , which affect the estimates ( [ omega big ] ) and ( [ b bound ] ) . for certain cases with prime modulus we may be able to do better than theorem 3.5 . for example , the same method may be combined with other bounds replacing corollary [ cubes 1 ] for more specific families of polynomials .
this has the potential to obtain sharper estimates for such polynomials and also to increase the range of values of @xmath240 for which an analogue of theorem 3.5 would apply .
for example , bourgain , garaev , konyagin and shparlinski @xcite consider the number @xmath241 of solutions to the congruence @xmath242 they show that if @xmath243 then we have the bound @xmath244 for some constant @xmath245 depending only on @xmath246 ( lemma 2.33 of @xcite ) .
+ in @xcite , the same authors consider the number @xmath247 of solutions to the congruence @xmath248 @xmath249 and show that @xmath250 for some constants @xmath251 and @xmath245 depending only on @xmath246 ( theorem 17 of @xcite ) .
+ another possible way to improve on our results for certain classes of well - shaped sets is to use weyl s formula for tubes ( equation ( 2 ) of @xcite ) and steiner s formula for convex bodies ( equation ( 4.2.27 ) of @xcite ) to give an explicit constant in ( [ well - shaped ] ) for certain subsets of @xmath13^d$ ] for which these formula are valid .
this would have the effect of improving on the bounds ( [ omega big ] ) and ( [ b bound ] ) and hence the bound for @xmath3 and possibly the range of values of @xmath240 for which this bound would be valid .
the author would like to thank igor shparlinski for suggesting this problem and for his guidance while working on it and writing the current paper .
chang , j. cilleruelo , m. z. garaev , j. hernndez , i. e. shparlinski and a. zumalacrregui , ` concentration of points and isomorphism classes of hyperelliptic curves over a finite field in some thin families ' , ( available from http://arxiv.org/abs/1111.1543 ) .
j. cilleruelo , m. garaev , a. ostafe and i. e. shparlinski , ` on the concentration of points of polynomial maps and applications ' mathematische zeitschrift , ( to appear ) .
s. parsell , s. prendiville and t. wooley , ` near - optimal mean value estimates for multidimensional weyl sums ' , ( available from http://arxiv.org/abs/1205.6331 ) .
w. schmidt , ` irregularities of distribution .
ix ' , _ acta arith .
_ , * 27 * ( 1975 ) , 385396 . r. schneider , ` convex bodies : the brunn - minkowski theory ' , encyclopedia of math . and its applications * 44 * , cambridge u. press , 1993 . i. e. shparlinski , ` on the distribution of points on multidimensional modular hyperbolas ' , _ proc .
japan acad .
, ser.a _ , * 83 * ( 2007 ) , 59 . | we use a generalization of vinogradov s mean value theorem of s. parsell , s. prendiville and t. wooley and ideas of w. schmidt to give nontrivial bounds for the number of solutions to polynomial congruences , when the solutions lie in a very general class of sets , including all convex sets . |
according to recent _ kepler _ statistics , most of the population of known transiting planets ( including planetary candidates ) around gfkm stars is dominated by planets of size between one and four earth radii ( @xcite ; @xcite ; @xcite ) . from small to large , those planets can be basically categorized into earth - size planets , super - earths and mini - neptunes ( @xcite ) .
many of them are expected to harbor atmospheres , several of which have been characterized , including gj1214b ( @xcite ) , gj436b ( @xcite ) , hd 97658b ( @xcite ) , gj 3470b ( @xcite ) , and hat - p-11b ( @xcite ) . when the planet is smaller than jupiter and especially neptune or uranus , the atmosphere might no longer be solely dominated by hydrogen ( e.g. , @xcite ) .
compositional diversity of the atmospheres emerges , as seen on the terrestrial planets in the solar system .
however , identifying the compositions of those atmospheres is difficult because the transit spectroscopic measurements are influenced by hazes or clouds .
in fact , four out of the above five planets show featureless transmission spectra ( @xcite ; @xcite ; @xcite ; @xcite ) , except hat - p-11b that exhibits water absorption in a hydrogen environment ( @xcite ) .
theoretically , the observable outer envelopes of the atmospheres of the sub - jupiter - size planets could exhibit a variety of compositions ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ) .
the primordial constituents of the atmosphere are mostly determined by the metallicity of the protoplanetary disk ( e.g. , @xcite ; @xcite ) , the details of gas accretion during planetary formation , location of the planetary formation , and subsequent migration tracks ( e.g. , @xcite ) .
the atmospheric bulk compositions will also be greatly influenced by many processes during evolution after planetary formation .
those processes include planetary and asteroid impacts , volcanism and outgassing ( @xcite ) , plate tectonics ( @xcite ) , atmosphere - surface interaction , atmosphere - ocean exchange , atmospheric escape ( @xcite ; @xcite ; @xcite ) , and possible biospheric processes ( @xcite ; @xcite ) .
atmospheric dynamics and chemistry will also play significant roles in determining the observable atmospheric composition ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) . as a result
, we expect to observe a large compositional diversity among sub - jupiter - size exoplanets . there could be hydrogen - dominated atmospheres as on jupiter and hot jupiters ; more metal - rich ( but still hydrogen - dominated ) atmospheres like uranus and neptune ; evaporated atmospheres dominated by helium ( proposed for gj436b , @xcite ) ; or carbon dioxide atmosphere like those of venus and mars ; water worlds ( proposed for gj1214b , @xcite ; @xcite ; @xcite ) ; nitrogen and oxygen atmosphere like earth ; nitrogen and methane atmosphere like saturn s moon titan ; atmosphere of sulfur compounds like jupiter s moon io , carbon monoxide or hydrocarbon atmospheres ( @xcite ) , or some metallic - based or silicon - based atmospheres if the planet is very close to the host star ( @xcite ; @xcite ; @xcite ) .
more quantitatively , given the metallicity and effective temperature , thermochemical and photochemical calculations have predicted a variety of possible bulk constituents for exo - planetary atmospheres ( e.g. , @xcite ; @xcite ) .
most known exoplanets are close to their host stars and expected to be tidally locked with a synchronized rotation around the central star .
weather on those planets is intriguing .
however , to date , there has not been a systematic study on bulk composition effects on the atmospheric dynamics and general circulation of tidally locked planets .
previous studies mostly focused on the hydrogen atmospheres of hot jupiters ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) , and earth - like environments for earth - size planets in the habitable zone ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ) .
@xcite has considered the effect of different atmospheric opacity sources on the general circulation for gj436b , but the bulk composition of the atmosphere is still hydrogen .
recently , the dynamical simulations by @xcite for super - earth gj1214b illustrate the effects of metallicity and possible compositions on the atmospheric dynamics .
it is found that atmospheres with a lower mean - molecular weight have smaller day - night and equator - to - polar temperature contrast than higher mean - molecular weight atmospheres .
they also show that a water atmosphere exhibits an equatorial prograde jet , but a carbon - dioxide atmosphere is dominated two high - latitude jets , suggesting significant bulk composition effects on the atmospheric dynamics .
there are three important properties of an atmospheric species that could influence the atmospheric dynamics : opacity , molecular weight , and heat capacity .
opacity fundamentally controls the heating and cooling rates in the atmosphere , which drives the dynamics .
molecular weight greatly affects the scale height of an atmosphere , the buoyancy frequency , atmospheric wave properties such as gravity wave speed , and the deformation radius , which is a typical dynamical length scale in the atmosphere .
heat capacity also affects the radiative heating and cooling rates , buoyancy , wave speed and deformation radius , but in a different way from that of the molecular weight .
given the complicated roles the bulk composition could play , it is difficult to disentangle the underlying mechanisms if we consider all those effects together . instead
, here we use an idealized general circulation model ( gcm ) to study individual effects .
specifically , we investigate the effects of the molecular weight and heat capacity while holding fixed the day - night heating scheme .
note that gas opacity is commonly dominated by trace species that do not significantly affect the molecular mass , e.g. , water and carbon dioxide in the case of earth ; water , carbon monoxide , and methane in the case of hot jupiters .
jupiter itself has significant opacity contribution from the background hydrogen , but also from trace hydrocarbons including methane , ethane , and acetylene ( @xcite ) .
therefore , we leave a thorough investigation of the opacity effect for the future .
ccccccc @xmath3 & 2.02 & 28.85 & 29.43 & 30.21 & 34.28 & + @xmath4 & 4.00 & 20.79 & 20.79 & 20.79 & 20.79 & + @xmath5 & 16.04 & 35.71 & 51.17 & 71.80 & 94.40 & + @xmath6 & 18.02 & 33.60 & 36.88 & 41.27 & 51.18 & + @xmath7 & 28.01 & 29.14 & 30.87 & 33.18 & 36.25 & + @xmath8 & 28.01 & 29.13 & 30.66 & 32.70 & 35.97 & + @xmath9 & 32.00 & 29.39 & 31.74 & 34.87 & 37.74 & + @xmath10 & 44.01 & 37.22 & 44.54 & 54.31 & 60.35 this study has important implications for observations .
we aim to build quantitative theories for the bulk composition effects using a series of analytical expressions and predict some observable signatures , including the spatial maps of wind and temperature , day - night temperature differences and thermal phase curves .
it is expected that the atmospheric dynamics will have significant effects on these observables
. for tidally locked planets , the day - night temperature difference and the shape of the thermal phase curve can be greatly influenced by the day - night heat redistribution efficiency and the presence of an equatorial prograde jet ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) .
the atmospheric wind speed might be directly measured through ultra - high resolution doppler mapping techniques that have been applied to hot jupiter hd209458b ( @xcite ) and hd189733b ( @xcite ; @xcite ) . on the other hand ,
measuring phase curves and winds might provide some hints on the properties of the bulk composition in the atmosphere .
this may serve as an alternative way to overcome the current difficulties in determining the atmospheric composition of transiting super - earths / mini - neptunes that exhibit featureless spectra , such as gj1214b .
this paper is organized as follows . in section 2
, we will first provide some theoretical background on the effects of compositions on atmospheric dynamics .
we will introduce the atmospheric circulation model and qualitatively summarize the main results from our simulations in section 3 .
then we will dig into the underlying physical mechanisms by which the temperature and wind patterns are affected by molecular weight ( section 4 ) and molar heat capacity ( section 5 ) .
we will also derive a series of analytical expressions to interpret the simulation results .
this paper ends with a conclusion and implications for the observations in section 6 .
the analytical treatment of temperature , wind and thermal phase shift are detailed in the appendices .
thermochemical and photochemical models ( e.g. , @xcite ; @xcite ) predict a variety of possible bulk constituents , depending on the metallically of the atmosphere , carbon to oxygen ratio , stellar flux , and atmospheric evolution .
plausible constituents exhibit a wide range of mean - molecular weight ( @xmath11 ) and molar heat capacity at constant pressure ( @xmath12 ) , summarized in table 1 and fig .
1 . molecular weight can vary by a factor of @xmath1320 , from @xmath14 to @xmath15.the heat capacity , however , can only vary by a factor of @xmath134 , from @xmath16 for helium to about @xmath17 for high - temperature methane , where @xmath18 is the molar gas constant .
the molecular weight and molar heat capacity are independent properties of a gas molecule .
the latter is solely controlled by the degrees of freedom of the molecule . the monatomic gas , helium , which could conceivably comprise the background atmospheric gas as a result of atmospheric escape processes ( @xcite ) , has the lowest @xmath12 ( @xmath16 ) .
diatomic gases , such as hydrogen , carbon monoxide , nitrogen and oxygen , have molar heat capacities ranging from @xmath19 at low temperature up to @xmath20 at high temperature for the vibrational modes to be fully excited .
multi - atomic molecules , such as a triatomic linear molecule carbon dioxide and more complex - structured molecules like water and methane , usually contain a larger number of degrees of freedom and therefore span a large range of @xmath12 .
for example , the molar heat capacity of methane can reach more than @xmath17 at 2000 k , although thermochemical equilibrium prefers forming carbon monoxide instead of methane at high temperature ( e.g. , @xcite ) . from thermochemical calculations , co and @xmath21 will never dominate the bulk atmosphere , but their contribution to the mean molecular weight and molar heat capacity could be significant . through what physical mechanisms
could the molecular weight and molar heat capacity influence the atmospheric dynamics on a tidally locked planet , including observational signatures such as temperature and wind speed ?
for a tidally locked planet in the moderate temperature range , if the magnetohydrodynamics ( mhd ) effect can be ignored , the photospheric dynamics is approximately governed by the hydrostatic primitive equations . in pressure ( @xmath22 )
coordinate , the equations are : @xmath23 where @xmath24 is density ; @xmath11 is molecular weight ; @xmath25 is temperature ; @xmath12 is molar heat capacity in units of @xmath26 ; @xmath27 is vertical velocity ; @xmath28 is geopotential where @xmath29 is gravitational acceleration and @xmath30 is altitude ; @xmath31 is coriolis parameter where @xmath32 is the planetary rotation rate and @xmath33 is latitude ; @xmath34 is the horizontal gradient at constant pressure ; @xmath35 is the horizontal velocity at constant pressure where @xmath36 is the zonal ( east - west ) velocity and @xmath37 is the meridional ( north - south ) velocity ; @xmath38 is the radiative heating / cooling rate in units of @xmath39 .
@xmath40 is the material derivative ; @xmath41 is a drag term representing missing physics such as sub - grid turbulent mixing ( e.g. , @xcite ; @xcite ) .
both the molecular weight and molar heat capacity appear in the thermodynamic equation ( 1d ) and the molecular weight is also present the equation of state ( 1e ) .
a change of the thermodynamic properties in the atmosphere ( e.g. , temperature ) will directly feed back to the geopotential field and lead to a change of atmospheric motion .
the major effects on the atmospheric dynamics can be categorized into radiative effects and dynamical effects .
the radiative heating and cooling is the primary driving force of the atmospheric dynamics on tidally locked planets . in principle
, the three - dimensional profiles of heating and cooling ( e.g. , the strength and pressure - dependence of the dayside heating and nightside cooling ) are affected by opacities and therefore composition . as noted before
, we do not investigate the opacity effect in this study . in order to isolate the effect of composition on the opacity from the other bulk composition effects , the simulations in this study share the same reference properties such as
atmospheric opacity , planetary effective temperature , planetary mass , planetary radius , the same distance from the same host star and same cooling mechanism , but only differ in bulk compositions .
the temperature change due to radiative heating and cooling also depends on the molecular weight and molar heat capacity ( eq . 1d ) .
a higher molecular weight implies less molecules at a given pressure , resulting a greater temperature change rate , or equivalently a smaller radiative timescale .
a higher molar heat capacity suggests that the atmosphere is more difficult to be heated and cooled , leading to a larger radiative timescale .
for example , the radiative timescale in a hydrogen atmosphere is more than 15 times longer than the @xmath1 atmosphere solely due to the effects of @xmath11 and @xmath12 .
the dynamical effect is also significant .
a typical dynamical timescale in the atmosphere is the wave propagation timescale @xmath42 ( @xcite ) , where @xmath43 is a typical horizontal length scale ( e.g. , radius of the planet ) , @xmath44 is the buoyancy frequency ( brunt - v@xmath45is@xmath45l@xmath45 frequency ) of the atmosphere , and @xmath46 is the pressure scale height .
@xmath47 approximates the horizontal phase speed of long - vertical - wavelength gravity or kelvin waves , which is greatly influenced by the molecular weight and molar heat capacity : @xmath48^{1/2}.\ ] ] the gravity wave speed @xmath47 inversely scales as @xmath49 , implying that higher molecular - mass atmospheres have slower gravity / kelvin wave speeds than lower - molecular - mass atmospheres . for an isothermal atmosphere ( @xmath50 ) ,
@xmath47 inversely scales as @xmath51 , implying that atmospheres whose molecules have more degrees of freedom likewise have slower gravity wave speeds . nevertheless , because @xmath12 varies over a smaller range than @xmath11 for plausible compositions ( fig .
1 ) , in practice , the effect of @xmath12 is generally secondary to the effect of @xmath11 on the wave speeds .
an exception occurs if the temperature profile is close to an adiabat , because in that situation , @xmath52 is close to @xmath53 in eq .
small variations in @xmath12 can cause large fractional variations in the factor in square brackets , and therefore large fractional changes to the wave speeds and possibly the entire dynamic regime .
we will see this effect in section 5 . in an isothermal atmosphere ,
the wave propagation timescale @xmath54 can be estimated : @xmath55 showing that the typical dynamical timescale of the atmosphere increases with molecular weight and molar heat capacity .
a useful atmospheric dynamic length scale is the rossby deformation radius , within which the atmospheric flow is more significantly influenced by gravity and buoyancy effects instead of planetary rotation . in the equatorial region where the coriolis parameter is small , the rossby deformation radius ( @xmath56 ) can be approximated as : @xmath57^{1/4}.\ ] ] the meridional gradient of the coriolis parameter @xmath58 for planetary radius @xmath59 and latitude @xmath33 . for a given temperature - pressure profile
, the deformation radius always decreases as @xmath11 and @xmath12 increase .
a smaller rossby deformation radius implies that the typical dynamical features in the system , such as the stable vortex size , are smaller .
ccccccc i & 5 & standard cases : @xmath11 and @xmath12 from five typical constituents : @xmath0 , @xmath60 , @xmath61 , @xmath2 and @xmath1 .
& + ii & 7 & keep @xmath12 as @xmath19 but vary @xmath11 of 2 , 4 , 8 , 16 , 24 , 32 and 44 @xmath62 , hereafter mass-2 " , etc .
& + iii & 7 & keep @xmath63 same as the mass-2 " case but vary @xmath11 of 2 , 4 , 8 , 16 , 24 , 32 , and 44 @xmath62 . & + iv & 4 & keep @xmath11 as 16 @xmath62 but vary @xmath12 of @xmath19 , @xmath64 , @xmath65 and @xmath66 . the interaction between the radiative effect and dynamical effect
is complicated and often leads to noticeable effects on atmospheric dynamics .
@xcite showed that hot jupiters should experience two distinct circulation regimes depending on the relative importance of wave propagation and radiative relaxation .
if the radiative timescale is sufficiently short compared with wave propagation timescale , the waves will be damped and jet formation will be inhibited ( @xcite ; @xcite ; @xcite ; @xcite ) . therefore the general circulation pattern in the atmosphere is dominated by a substellar - to - antistellar ( day - to - night ) divergent flow pattern instead of the equatorial zonal jet . in this situation , there is a large day - night temperature difference and small atmospheric hot spot shift in the thermal phase curve . on the other hand ,
if the radiative timescale is sufficiently long compared with wave propagation timescale , the atmospheric flow will self - organize to form a zonal jet pattern .
if we assume the radiative timescale increases with pressure , as generally it does ( e.g. , @xcite ; @xcite ) , the general circulation pattern in the atmosphere is expected to exhibit a regime shift " from a divergent flow in the upper atmosphere to a jet - dominated regime in the deeper atmosphere .
just such a transition in fact occurs in gcm simulations of hot jupiters ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) .
an intuitive measure of this transition is the ratio of radiative timescale versus the wave propagation timescale : @xmath67 , and both the molecular weight and molar heat capacity play important roles in this ratio .
to quantitatively understand the above effects on a tidally locked planet , we simulate the atmospheric circulation using a three - dimensional ( 3d ) general circulation model ( mitgcm ) that has been extensively used in studying atmospheric dynamics on planets in and out of the solar system ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ) . for simplicity ,
we consider a planet with zero obliquity and in a circular orbit around its host star .
we assume the radius , mass and rotation period of the idealized planet are 2.7 earth radii , 6.5 earth mass , and 1.365@xmath68 s , respectively .
those properties are similar to gj1214b , whose bulk atmospheric composition is still under debate ( @xcite ; @xcite ; @xcite ) .
we adopt newtonian cooling approximation as a simple radiative heating and cooling scheme . the heating term in eq .
( 1d ) is formulated as : @xmath69 if gj1214b has a hydrogen - dominated atmosphere , we estimated the radiative timescale @xmath70 on the order of @xmath71 s based on a simple gray cooling scheme : @xmath72 where @xmath73 is the pressure level at which the opacity of the atmosphere reaches unity
. motivated by earlier studies using newtonian cooling ( @xcite , @xcite ) , we assume that the vertical profile of radiative timescale increases as a power law of pressure , ranging from @xmath74 s at @xmath75 pa to @xmath76 s at @xmath77 pa ( fig .
2 ) : @xmath78 as noted before , we assume the same opacity for all compositions , implying that @xmath73 is constant . based on eq .
( 6 ) , the radiative timescale @xmath63 for other compositions can be scaled based on molecular weight and molar heat capacity : @xmath79 the spatial distribution of the radiative equilibrium temperature @xmath80 is : @xmath81 where @xmath82 is longitude and we assume a homogeneous equilibrium temperature in the nightside @xmath83 .
the equilibrium temperature difference @xmath84 is 600 k at the top of the atmosphere and decreases towards zero at the bottom in a linear fashion with @xmath85 .
the mean temperature @xmath86 is calculated using a simplified radiative scheme from @xcite ( fig .
the vertical temperature profile is designed to be stably stratified , meaning that its vertical gradient , @xmath87 , does not exceed @xmath52 for all compositions . and
@xmath88).,scaledwidth=45.0% ] we assume a linear drag scheme in the momentum equation ( eq .
1a ) , @xmath89 . the drag coefficient ( @xmath90 ) is assumed as @xmath91 at the bottom of the domain ( @xmath77 pa ) and decreases linearly with decreasing pressure to zero at @xmath71 pa , above which the atmosphere is essentially drag - free . for simplicity , we assume in each simulation that the value of @xmath12 is a constant ( using the value at 600 k , table 1 ) .
thus , we are neglecting the temperature dependence of the heat capacity . in most cases
we expect this temperature dependence to be a secondary effect , so this should be a reasonable assumption for the present study .
our tests show that this assumption is actually sufficient to diagnose the role of @xmath12 on the atmospheric dynamics .
we specifically designed four sets of experiments to investigate the effects and mechanisms .
table 2 summarizes our simulation experiments .
the experiment i includes the standard cases with five typical constituents : @xmath0 , @xmath60 , @xmath61 , @xmath2 and @xmath1 . in experiment ii , we keep @xmath12 fixed and vary @xmath11 to study the molecular weight effect at constant molar heat capacity .
the experiment iii is very similar to ii but we fix the radiative timescale @xmath63 to separate the radiative effect from dynamical effect of @xmath11 .
the experiment iv , in which we keep @xmath11 constant and vary @xmath12 in the simulations , is designed to investigate the molar heat capacity effect at constant molecular weight .
the radiative timescale @xmath63 in i , ii and iv is allowed to vary between simulations but it is fixed in iii . the standard simulations ( experiment i ) exhibit important features and trends in the longitude - latitude temperature and wind maps and the zonal - mean zonal wind ( @xmath92 ) distributions ( fig .
3 ) can be decomposed as @xmath93 , where @xmath94 is the zonal - mean value averaged over a latitude circle and @xmath95 is the local departure from the zonal average or so - called eddy " value . ] .
the temperature and wind maps from experiment ii ( fig .
4 ) look very similar to that from experiment i , implying that the molecular weight effect dominates over the molar heat capacity effect . in order to tie the simulation results to the observations and better understand the theoretical trends , we define several quantities : the day - night temperature contrast
, phase offsets of flux peak before the time of secondary eclipse in the thermal phase curve , root - mean - square ( rms ) of the wind speed @xmath96 , averaged zonal - mean zonal wind speed in the equatorial region @xmath97 , and the pressure where the zonal - mean zonal jet speed reaches a maximum ( hereafter the jet core " region ) .
the above quantities exhibit the same trends in experiments i and ii as a function molecular weight from @xmath0 to @xmath1 ( fig .
5 ) . in the following ,
we summarize the major trends and qualitatively describe the possible mechanisms .
\(1 ) the temperature contrast at the same pressure level between the dayside and nightside increases with molecular weight ( fig .
we adopted the normalized day - night temperature difference ( @xmath98 ) on a global scale , similar to that in @xcite and @xcite : @xmath99 ^ 2}}{\overline{[t_{eq}(\lambda , \phi , p)-\overline{t}(\phi , p)]^2}}\bigg)^{1/2}d\phi.\ ] ] @xmath100 represents homogenous longitudinal temperature distribution and @xmath101 represents a temperature distribution close to the radiative equilibrium temperature pattern .
the normalized temperature difference varies from 0.2 in the @xmath0 case to about 1 in the @xmath1 case ( fig .
5 ) . as the molecular weight increases , the radiative timescale decreases and the dynamical timescale ( wave propagation timescale ) increases ,
leading to a more efficient relaxation of the thermal field towards the radiative equilibrium state and increase of the day - night temperature contrast . a smaller wind speed in higher - molecular - weight atmosphere ( fig .
5 ) also acts less efficiently to redistribute heat and smooth the thermal field by advection .
\(2 ) the phase offset before the secondary eclipse , which corresponds to a hot spot in the atmosphere shifted east of the sub - stellar point ( fig .
4 , left column ) , decreases with increasing molecular weight ( fig . 5 ) . as discussed in trend ( 1 ) , the decrease of the radiative timescale and the zonal wind speed with molecular weight helps maintain the radiative equilibrium thermal field , weakening the ability of advection to distort the temperature field .
\(3 ) the rms wind speed @xmath96 decreases with molecular weight ( fig . 5 ) .
here we calculated @xmath96 based on horizontal wind speeds @xmath36 and @xmath37 weighted by area over a 120 degree latitude band at each pressure level : @xmath102 however , this decreasing trend is less related to the changes of radiative and dynamical timescales . a smaller @xmath96 in the @xmath1 atmosphere than the @xmath0 atmosphere
is more associated with that the atmospheric scale height is smaller in the @xmath1 atmosphere than that in the @xmath0 atmosphere .
an atmosphere composed of heavier molecules will have a smaller day - night pressure - gradient force , leading to a smaller wind field in general .
a more rigorous discussion on the wind speed will be presented in section 4.3 .
\(4 ) the latitudinal width of the equatorial jet decreases with increasing molecular weight ( fig .
3 ) . fig .
5 presents the e - folding width of the equatorial zonal jet as function of molecular weight . in the @xmath0 case ,
the zonal - mean zonal wind is dominated by a broad eastward equatorial jet , which , over at least certain ranges of pressure , extends nearly to the poles .
the jet width shrinks with increasing molecular weight .
3 shows that , for the case of @xmath61 , the equatorial jet is sufficiently narrow to allow room for broad regions of westward equatorial flow at mid - to - high latitudes . for @xmath2 and @xmath1 ,
the equatorial jet is even narrower , and and two new high - latitude eastward wind regions appear .
the same trends emerge in fig .
4 when molecular mass is varied while holding molar heat capacity constant .
this trend is related to the decrease of dynamical length scale with increasing molecular weight .
\(5 ) at the same pressure , the general circulation pattern gradually shifts from the zonal - jet - dominated regime to the day - to - night circulation regime as the molecular weight increases ( fig .
4 , left column ) .
correspondingly , the jet core pressure occurs deeper in the atmosphere ( right column of fig .
4 , fig . 5 ) .
this is closely related to the interplay between the radiative timescale and wave propagation timescale .
@xcite showed that when the radiative timescale is significantly less than the wave - propagation timescale , the atmosphere tends to be dominated by a day - night flow .
whereas strong zonal jets emerge when the radiative timescale is comparable to or longer than the wave propagation timescale . because the ratio of the radiative timescale to the wave propagation timescale @xmath67 decreases with increasing molecular weight , the transition between the day - to - night circulation regime and the jet - dominated regime moves deeper into the atmosphere .
\(6 ) we defined the root - mean - square of the equatorial zonal - mean zonal wind speed @xmath97 and the eddy component @xmath103 over a 60 degree latitude band : @xmath104 if the zonal velocity @xmath36 is much larger than the meridional velocity @xmath37 , @xmath96 in the equatorial domain ( eq .
11 , but averaged in the equatorial region ) can be related to @xmath97 and @xmath103 via @xmath105 .
the maximum zonal - mean zonal wind @xmath97 in the atmosphere ( jet core speed " ) decreases with molecular weight .
this might be associated with the trends ( 3 ) and ( 5 ) .
the molar heat capacity has a secondary effect , which can be seen from the small deviation between the typical composition simulations and the cases with molar heat capacity fixed ( fig .
5 ) , for instance , the small departure between the he case and the case with the same molecular weight ( 4 @xmath62 ) but @xmath12 fixed ( hereafter mass-4 " ) , and also the difference between the @xmath1 case and the mass-44 case .
more discussion will be presented in section 5 .
next we will present analytical theories and quantitatively discuss each of those trends . in order to elaborate the dependence of these trends on the molecular weight more clearly
, our discussion will mainly focus on the experiment ii .
the dayside - nightside temperature difference is a fundamental aspect of the overall circulation and moreover directly controls the amplitude of the phase variation in infrared phase curves a major observable .
our simulations show that the day - night temperature difference increases with molecular weight ( fig .
this results from the effect of the molecular weight on both the dynamics and on the radiative timescale via eq .
( 5 ) . to separate these effects ,
experiment iii shows how the day - night temperature difference varies with molecular mass when the radiative timescale held fixed as the hydrogen case ( eq .
the longitude - latitude maps of temperature are shown in fig .
the integrated dayside - nightside temperature difference @xmath98 , shown in fig .
7 , indicate that even when radiative time constant is held fixed , the day - night temperature difference increases with molecular mass though more weakly than in the experiments i and ii . here , we discuss the physical reasons for all of these trends and compare them to theory . in particular , @xcite and @xcite presented analytical theories for the day - night temperature difference on hot jupiters ; @xcite constructed their theory for a 1.5-layer shallow - water model , and @xcite extended it to the three - dimensional primitive equations . here , we show that the @xcite theory explains our simulated trends well .
we also present a modified form of the @xcite theory that , while incorporating the same assumptions , is more analytically compact .
@xcite and @xcite pointed out that the day - night temperature difference is controlled by several intrinsic timescales in the atmosphere , including the rotational timescale , radiative timescale , wave propagation timescale , horizontal and vertical advection timescale , and the frictional drag timescale ( if drag is present ) . by combining the horizontal momentum equation and the thermodynamic energy equation under the `` weak temperature gradient '' ( wtg ) approximation ( @xcite ) ,
we can obtain analytical expressions for a proxy of day - night temperature difference ( @xmath106 ) and the rms wind velocity @xmath96 , on synchronously rotating planets . in the momentum equation ,
the day - night pressure gradient force , which results from the day - night temperature difference and is responsible for driving the circulation , is primarily balanced against the largest of the coriolis , horizontal - advection , vertical - advection , and frictional - drag forces per mass , respectively .
the four possible balances lead to four distinct expressions for @xmath98 , along with conditions that determine which of these four regimes applies in any given portion of the parameter space ( @xcite ) .
this separation into regimes is helpful for clarifying our understanding of the behavior in each regime .
nevertheless , it would be useful to obtain a single analytic expression for @xmath98 that interpolates smoothly between the regimes .
here we united the above four regimes and achieved a single and close - form expression of @xmath98 in appendix a : @xmath107 where the non - dimensional parameters @xmath108 and @xmath109 are defined as : @xmath110 here , @xmath111 is the frictional drag timescale .
@xmath112 is the timescale for horizontal wave propagation .
the quantity @xmath113 is a reference advective timescale due to the `` equilibrium cyclostrophic wind '' @xmath114 , which is a hypothetical , reference wind speed that would result from acceleration of the wind from day to night due to the day - night pressure gradient if the day - night temperature difference were in radiative equilibrium and if the rossby number exceeds unity ( compare to @xcite , eq . 48 ) . in the above
, @xmath115 is the difference in log pressure between some deep pressure where the day - night temperature difference is small ( 10 bars in the theory and simulations from @xcite ) and some smaller pressure of interest in the observable atmosphere .
see @xcite for more details about the motivations for and assumptions underlying the theory , and comparisons with a grid of gcm simulations for hot jupiters covering a wide range of @xmath63 and @xmath111 .
this theory implies that the day - night temperature contrast is smaller if the radiative timescale is larger or the dynamical timescale is smaller . to view the dependence on molecular weight
, we evaluate each term in @xmath108 and @xmath109 . assuming the dynamical length scale @xmath43 is constant ( representing a planetary radius associated with the day - night temperature contrast , for example ) , then @xmath116 in an isothermal atmosphere ( eq .
3 ) and the advective timescale @xmath117 is scaled as @xmath118 . if @xmath63 is considered constant , then the second term in the expression for @xmath108 ( eq .
15 ) is proportional to @xmath11 , and @xmath119 ( eq .
whereas if @xmath63 scales inversely with @xmath11 as expected from eq .
( 8) , then the second term in the expression for @xmath108 is proportional to @xmath120 and @xmath121 .
therefore , eq . ( 14 ) predicts that the day - night temperature contrast should increase with the molecular weight , and this dependence should be weaker in the case where @xmath63 is constant than in the case where @xmath63 decreases with increasing molecular weight .
both predictions are qualitatively consistent with the trends in our simulations ( fig . 7 ) .
this theory implies that in the limit of very high molecular mass ( @xmath122 ) , the atmospheric temperature structure tends toward radiative equilibrium .
this can be seen from the fact that , in this limit , the denominator in the second term on the righthand side of eq .
( 14 ) becomes infinite , which therefore implies that @xmath123 .
the convergence toward radiative equilibrium as @xmath124 is fastest if we allow @xmath63 to scale inversely with @xmath11 ( eq .
8) but occurs even if @xmath63 were taken to be independent of molecular weight .
what is the physical reason for the convergence toward radiative equilibrium at high molecular mass ?
@xcite and @xcite showed that the day - night temperature difference is regulated by a wave - adjustment mechanism : planetary - scale waves that propagate from the dayside to the nightside act to adjust isentropes up or down in an attempt to flatten them , which , acting in isolation , attempts to produce a state with minimal day - night temperature differences .
this process operates efficiently when the waves can propagate from the dayside to the nightside faster than radiation or friction can act to damp them . on the other hand ,
if friction or radiation damps the waves before they propagate from the dayside to the nightside , then the wave - adjustment mechanism is suppressed , and the day - night temperature difference becomes large . in the limit of high molecular mass
, the wave speeds @xmath47 become slow , and therefore the day - night wave propagation timescale @xmath54 becomes long ( eq .
3)which gives the atmosphere ample time to respond to radiation for any plausible , finite value of @xmath63 .
in other words , the limit @xmath124 corresponds to the limit where the wave - adjustment mechanism acts less and less efficiently , allowing it to be easily suppressed .
( 14 ) also quantitatively predicts the day - night temperature difference from first principles . in our study , drag is not important above the bottom frictional layer ( inviscid limit ) therefore @xmath111 can be neglected .
we use the mean temperature profile @xmath86 ( fig .
2 ) to estimate @xmath54 . at 700 pa , taking the horizontal length scale as @xmath125 m , @xmath126 k , @xmath127 500 k , and radiative timescale based on eq .
( 8) , we obtain the theoretical curve of @xmath98 as function of molecular weight ( fig .
the predictions agree fairly well with our simulation results .
the theory also predicts that @xmath98 should decrease with increasing pressure . among the parameters in @xmath108 and @xmath109
, @xmath63 increases by several orders of magnitude from the top to the bottom of the atmosphere , while the other parameters such as @xmath54 , @xmath117 and @xmath115 have much weaker pressure dependence . from eq .
( 14 ) , @xmath98 should monotonically decrease with pressure . since @xmath84 also decreases with increasing pressure
, the day - night temperature contrast should be smaller in the deeper atmosphere , consistent with our simulation results .
we also investigated the significance of the radiation effect versus the dynamical effect . in experiment iii
, we fixed the radiative timescale for all simulations regardless of molecular weight .
the resulting day - night temperature difference is much lower than the simulations where @xmath63 changes with @xmath11 ( fig .
7 ) , suggesting an important role of the radiative relaxation . for the analytical prediction
, we use a constant radiative timescale @xmath128 s but keep the same previous parameters , the theoretical prediction agrees very well with the experiment iii simulations ( fig . 7 ) .
due to the strong equatorial prograde jet , the temperature maximum is advected eastward with respect with the sub - stellar point , causing the peak of thermal flux to become shifted before the secondary eclipse .
this phase offset was first predicted by @xcite and later confirmed by thermal phase curve observations ( e.g. , @xcite ) .
@xcite systematically investigated the relationship between the horizontal advection and the thermal phase curve in a semi - analytic model . based on our 3d simulations , for each pressure level
, we first integrated the emergent thermal flux over the face - on hemisphere as a function of phase ( @xmath129 ) as the planet orbits / rotates : @xmath130 here , we define @xmath129 such that @xmath131 corresponds to the time of primary transit ( when the planet s nightside faces earth ) and @xmath132 corresponds to the time of secondary eclipse ( when the planet s dayside faces earth ) , and @xmath133 and @xmath134 correspond , respectively , to the times when the planet s western and eastern terminators lie at the sub - earth longitude . note that longitude is defined such that @xmath135 is the substellar point and @xmath136 is the anti - stellar point .
then we removed the mean of the integrated flux and obtain a normalized thermal phase curve @xmath137 as a function of orbital phase ( in degrees ) after normalization : @xmath138 where @xmath139 denotes the averaged flux over the full phase .
the normalized thermal phase curves at 3000 pa for the standard cases are illustrated in fig .
8 . for clarity , the primary and secondary
eclipses are not shown here .
there is a clear anti - correlation between the phase offset before the secondary eclipse ( @xmath132 ) and the amplitude of the phase curve . from hydrogen to @xmath1 ,
the phase shift is less as the amplitude increases .
the thermal phase curve of the @xmath1 case is almost symmetric about the secondary eclipse , and the amplitude of the maximum flux is actually larger than twice of the mean emergent flux .
although the phase curve is an integrated effect of the temperature distribution , we found that the degree of thermal phase shift ( @xmath140 ) with respect to the secondary eclipses can be approximated by longitudinal offset of temperature maximum ( @xmath141 ) with respect to the substellar point in our simulations .
( black ) , @xmath60 ( blue ) , @xmath61 ( green ) , @xmath2 ( orange ) and @xmath1 ( red ) .
the primary transit occurs at phase @xmath142 and the secondary eclipse occurs at phase @xmath143.,scaledwidth=45.0% ] our previous theory for day - night temperature difference ( eq .
14 ) only predicts a global - average quantity @xmath98 but does not provide the longitudinal variation of the temperature , especially the offset of the temperature field . in order to quantitatively
investigate the thermal phase shift , we construct in appendix b a simple kinematic model to study how the longitudinal temperature distribution is influenced by a constant zonal wind .
our procedure is similar to that in @xcite but different in the radiative relaxation schemes . to ensure consistency with our 3d simulations , we adopt the newtonian cooling scheme where the relaxation rate is proportional to temperature , while in @xcite the relaxation rate is proportional to @xmath144 . in appendix
b we show that the temperature distribution and therefore the thermal phase curve can be characterized by three parameters in the atmosphere : @xmath145 and @xmath146 , where @xmath147 is the ratio between the radiative timescale @xmath63 and the advection timescale @xmath148 .
the advection timescale here should be defined as @xmath149 for a specified zonal - mean zonal wind @xmath92 .
therefore , this is not a first - principle theory because we do not have a theory to predict the zonal - mean zonal wind . with the @xmath92 from numerical simulations , the analytical temperature distributions and thermal phase curves
agree fairly well with the 3d gcm results ( appendix b ) .
our analytical theory provides a simple but physically based tool for future analysis of transit thermal phase curves . in appendix b
we also show that the longitudinal hot spot offset @xmath141 satisfies : @xmath150 where @xmath151 and @xmath152 @xmath141 can not be expressed explicitly but can be solved easily . when @xmath141 is small , @xmath153 .
given our expectation that the longitudinal offset of the hot spot ( @xmath141 , as measured in degrees ) is similar to the phase shift of the flux peak in the lightcurve ( @xmath140 , again expressed in degrees ) , eq .
( 19 ) can thus be used to estimate the phase shift in the lightcurve for a particular model .
note that @xmath141 only depends on @xmath146 .
it implies the thermal phase shift @xmath140 might be mainly dominated by @xmath146 as well ( see appendix b for more discussion ) . if we take @xmath154 where @xmath92 is calculated from the simulations ( eq .
12 , but averaged from latitude @xmath155 to @xmath156 ) , the predicted phase shifts from eq .
( 19 ) agree fairly well with the thermal phase shift in all simulation cases for several pressure levels ( fig .
it suggests that our kinematic thermal model has captured the essential physics to explain the zonal distribution of thermodynamic quantities in 3d simulations .
our simple theory reveals an essential relationship between the phase shift and two fundamental timescales in the system .
if the radiative timescale is very short relative to the advective timescale , i.e. , @xmath157 , then @xmath158 , meaning that there is no thermal phase shift .
the atmosphere thermal structure is entirely controlled by the radiation and retains the equilibrium temperature distribution , @xmath159 . on the other hand ,
if the radiative timescale is very long compared with the advective timescale , then the phase offset shifts toward 90 degrees , and the predicted amplitude of the phase curve variation goes to zero .
the trend of the horizontal temperature distribution is mainly due to the molecular weight effect , as evidenced by the similarities between experiment i ( fig .
3 ) and experiment ii ( fig .
the radiative timescale is inversely proportional to the molecular weight ( eq .
the advective timescale @xmath148 due to the zonal - mean zonal wind increases with molecular weight ( fig . 5 ) .
therefore , the ratio @xmath160 must decrease with increasing molecular weight , and correspondingly , the thermal phase shift decreases , as shown in experiments i and ii ( fig . 10 ) . since @xmath63 increases dramatically ( fig .
2 ) with pressure and @xmath148 has a weaker dependence on pressure ( suggested by @xmath92 in fig .
4 ) , the theory implies that the phase shift is larger at greater depth in the atmosphere .
our simulations confirm that the phase shifts at 700 pa are smaller than that at 3000 pa ( fig .
however , this prediction can not be applied to the jet core region where the atmosphere is overwhelmingly dominated by jets and waves and the thermal phase curve does not look like a sinusoidal - type curve but a more irregular pattern . despite the success of eq . ( 19 ) in explaining the simulation results , we should emphasize that it is not a fully predictive theory , since it can only be evaluated with knowledge of zonal - mean zonal wind speeds . a fully predictive , first principles theory of the hot spot offset would require a quantitatively accurate theory for the zonal - mean zonal wind speed , which is currently lacking .
the same theory for the day - night temperature difference in appendix a also predicts the rms wind speed @xmath96 ( hereafter @xmath161 ) : @xmath162 where @xmath108 and @xmath109 are from eq .
( 15 ) and ( 16 ) , respectively .
again , this is essentially identical to the prediction from @xcite except that it combines their distinct expressions from the multiple regimes ( for possible balances in the momentum equation ) into a single expression that smoothly varies between the regimes and is valid across the entire parameter space .
recall that the the hypothetical equilibrium cyclostrophic wind " @xmath114 is inversely propositional to the square root of the molecular weight ( section 4.1 ) .
the ratio of the wind speed to @xmath163 solely depends on a single parameter @xmath164 . at 700 pa ,
the simulations show that the rms wind speed decreases with increasing molecular weight ( fig .
11 ) . with the same parameters in section 4.1 ,
the predicted wind speeds are in a good agreement with the high molecular weight simulations but underestimate the low - molecular - weight simulation results ( e.g. , @xmath165 ) by a factor of 2 to 3 ( fig .
the analytical expression shows a non - monotonic behavior in the low molecular weight regime , which is not consistent with the numerical simulations , suggesting some caveats in our analytical theory .
this same discrepancy between model and theory have been seen in hydrogen atmosphere simulations ( hot jupiters ) in @xcite with large drag timescale and radiation timescale , and fig .
6 of @xcite .
the reason is attributed to the lack of treatment of the equator - to - pole temperature gradient , which could allow stronger zonal winds in the simulations than are predicted in the theory ( @xcite ) . when the molecular weight is small ( such as hydrogen ) , the day - night temperature is small ( fig .
7 ) and the equator - to - pole temperature difference has a significant contribution to the horizontal winds . neglecting that effect in the current theory results in a factor of 2 - 3 less in the analytically predicted wind speeds compared to the simulation results .
nevertheless , in the high - molecular - weight regime , the day - night temperature contrast is more important and our theory behaves very well in reproducing the simulation results ( fig .
11 ) . as a function of molecular weight at 700 pa from experiments (
blue filled circles ) , ii ( black filled circles ) and iii ( open circles ) .
blue and yellow dashed lines are the predictions from eq .
( 22 ) and ( 24 ) , respectively.,scaledwidth=45.0% ] it is worth considering more theoretical details in the limit of high molecular weight .
we showed in section 4.1 that , at high molecular weight limit ( @xmath124 ) , the temperature converges toward the radiative - equilibrium temperature structure ( @xmath166 in fig .
this results from the fact that the wave speeds in the high molecular weight limit become slow , which means that the wave - adjustment mechanism that regulates the day - night temperature differences ( @xcite ; @xcite ) becomes easily suppressed .
thus , we can simply replace @xmath167 with @xmath84 in the momentum equation ( a.1a in appendix a ) . for a drag - free atmosphere ,
the momentum equation then becomes : @xmath168 this directly yields wind velocity : @xmath169 where @xmath170 is the rossby number associated with the equilibrium cyclostrophic wind " @xmath163 .
the wind speed predicted by eq .
( 24 ) agrees quantitatively well with the simulations for the high mass cases ( fig .
11 ) . strictly speaking , eq .
( 23 ) and ( 24 ) are only valid when @xmath171 , which occurs in the limit where the molecular weight is sufficiently large , i.e. , @xmath124 , so that @xmath172 and @xmath173 , leading to @xmath174 in eq . ( 24 ) . in other words , in this limit , since the pressure - gradient force ( right hand side of eq .
23 ) is sufficiently weak , we expect the winds to be weak the term quadratic in speed , @xmath175 , will become smaller than the linear term @xmath176 .
therefore the inverse dependence of pressure - gradient force on @xmath11 translates directly into an inverse dependence of @xmath161 on @xmath11 . however , in reality , the horizontal temperature difference @xmath167 in the cases with @xmath177 in experiments i and ii is already fairly close to the radiative - equilibrium temperature contrast @xmath84 ( fig .
. therefore , eq . ( 23 ) and ( 24 ) are also a good approximation for atmospheres with moderate molecular weight . based on eq .
( 24 ) , if the rossby number @xmath178 , @xmath179 ; if @xmath180 , then @xmath174 . as the molecular weight increases , the pressure - gradient force deceases , leading to a smaller @xmath163 and smaller rossby number @xmath181 . as a result ,
the dependence of rms wind speed on molecular weight will gradually shift from the @xmath182 regime to the @xmath183 regime . in our simulations ,
the wind speed @xmath161 decreases with @xmath11 in a trend between @xmath182 and @xmath183 ( fig .
decrease with pressure from experiment i.,scaledwidth=45.0% ] the physical reason that the wind speed decreases with increasing molecular mass in the high mass limit is straightforward . in a hydrostatically balanced atmosphere ,
the pressure - gradient force is essentially the result of a vertical integral over the day - night temperature difference . at high molecular mass ,
the scale height is small , and thus there is not much altitude difference between the photosphere pressure and the pressure of a given deep isobar where the day - night temperature difference is small .
the larger the molecular mass , the smaller is the altitude range over which the day - night temperature difference acts .
this leads to a smaller pressure - gradient force , which manifests as the inverse @xmath11 dependence in the right hand side of eq .
( 23 ) . as a result , the rms wind speed , which is the solution of the eq .
( 23 ) and directly driven by the pressure - gradient force , should decrease with increasing molecular weight accordingly .
vertically , because the pressure - gradient force decreases towards deeper into the atmosphere , @xmath96 decreases towards higher pressure for all cases ( fig .
( 24 ) also shows that , in a drag - free atmosphere , the theoretical wind speeds in the high molecular weight regime are solely controlled by three parameters : the planetary rotation rate , equilibrium day - night temperature difference and molecular weight of the atmosphere .
importantly , this implies that the wind speeds become independent of molar heat capacity and the mean temperature of the atmosphere . the high - molecular - mass limit which corresponds to the limit where the day - night temperature difference converges toward radiative equilibrium only applies when the wave timescale is sufficiently long compared to the radiative timescale .
nevertheless , this condition readily becomes satisfied when the molecular mass is large , @xmath177 in our simulations , for example .
once that limit is reached , the wave and radiative timescales no longer directly influence the wind speeds . our simulations seem to support this argument .
the wind speeds from our experiments i , ii and iii exhibit the same trend ( fig . 11 ) .
the agreement of experiments i and ii suggests that the molar heat capacity effect is not important .
the agreement of experiments ii and iii in which the radiative timescale is fixed for all cases evidently shows that the radiation timescale does not control the wind speed .
we also did an extra experiment ( not shown here ) based on the standard cases with mean temperature enhanced by a factor of 3 but keeping @xmath84 .
the resulting @xmath96 is roughly the same as that in experiment i , implying that the mean thermal state has negligible effect on the wind speed . in the zonal - mean zonal flow pattern ( fig . 3 and 4 )
, the number and width of the zonal jets depends on the molecular mass .
hydrogen atmospheres exhibit only one broad jet an eastward jet centered at the equator
while higher molecular - mass atmospheres have a narrower equatorial jet , flanked by regions of westward flow at higher latitudes . at the highest values of molecular weight ( e.g. , @xmath2 and @xmath1 ) , eastward zonal - mean zonal wind emerges close to the poles
the latitudinally alternating structure in the zonal - mean zonal wind map of the @xmath1 case is consistent with the simulations by @xcite with a fully realistic radiative transfer for gj1214b , although their polar jets have higher speeds .
the number of alternating banded structures could be roughly estimated from eq .
( 39 ) in @xcite based on the rhines scale argument : @xmath184 , implying an inverse correlation between the wind speed and number of bands in the atmosphere .
it predicts that a @xmath1 atmosphere that has a smaller wind speed than the @xmath0 case would exhibit more banded structure ( fig .
3 ) . the zonal - mean zonal flow patterns in our simulations
as well as that in @xcite are consistent with this argument .
the half - width of the equatorial jet can be estimated by the equatorial deformation radius @xmath185 that measures a decaying length scale of long equatorial rossby waves ( @xcite ) .
we assume an isothermal temperature profile and use eq .
( 4 ) to estimate the jet width : @xmath186 the predicted wind speeds generally follow the trend of the simulations but are slightly higher than the e - folding width of the simulated jets ( fig .
13 ) . a strong radiative relaxation acts to suppress the jets formation and damp the jets ( @xcite ) .
these authors showed that , if the radiative time constant is short , rossby and kelvin waves , which cause the phase tilts and drive the equatorial jet formation in a synchronously rotating atmosphere , will be suppressed as a result , the equatorial superrotating jet is damped .
@xcite argued that the suppression of jet formation should occur when the radiative time constant is shorter than the timescale for the kelvin wave to propagate over a hemisphere , i.e. , @xmath187 . in this case , the atmosphere exhibits a day - to - night flow pattern .
if the radiative timescale is longer than the wave propagation timescale ( @xmath188 ) , the atmosphere is dominated by an equatorial zonal jet . in our case ,
the radiative timescale is small in the upper atmosphere and increases with pressure ( e.g. , fig .
2 ) , and therefore it is possible that the general circulation pattern shifts from a day - to - night divergent flow in the upper atmosphere to a jet - dominated flow in the deeper atmosphere ( @xcite ) . this is consistent with our simulations where the equatorial zonal jet only occurs below some pressure level ( fig .
3 ) . , with a unity - slope dashed line .
experiments i to iv in table 2 corresponds to symbols of blue , black filled , black open , orange filled , respectively .
@xmath54 is zonal averaged for all experiments.,scaledwidth=45.0% ] this transition from a zonal flow at depth to a divergent day - to - night flow aloft suggests that the zonal - mean zonal wind maximizes at some specific pressure near this transition . in the regime of zonal jets
deep in the atmosphere , the zonal wind has similar sign at many longitudes along a given latitude circle , and thus the zonal - mean zonal wind is strong .
our models suggest that at and below the photosphere level , the zonal - mean temperature is generally warmer at the equator than at high latitudes , and this implies through a generalized thermal - wind balance that , starting from some deep level of assumed weak winds , the zonal - mean zonal winds will increase with height in the jet - dominated regime .
however , in the regime of day - to - night flow , there is a significant cancellation of the eastward- and westward - flowing branches along any given latitude circle , which causes the zonal - mean zonal wind to become weak , even if the local values of the day - night eddy winds themselves are large .
thus , in the divergent flow regime , one might expect the zonal - mean zonal winds to decrease with increasing altitude .
the implication is a maximum of zonal - mean zonal wind with pressure near the transition between these regimes .
interestingly , at altitudes above this maximum , the decrease of zonal - mean zonal winds with altitude would necessarily imply a reversal of the zonal - mean meridional temperature gradient ( with the poles being warmer than the equator in the zonal mean ) .
understanding the details of this phenomenon are beyond the scope of the present paper , although we note that the phenomenon is commonly associated with wave - driven circulations that are common in planetary stratospheres throughout the solar system .
still , our simulations are qualitatively consistent with both a regime transition from jet to divergent flow with increasing altitude and likewise generally exhibit a maximum zonal - mean zonal wind which occurs at approximately similar pressure to the transition ( see figs .
3 - 4 ) , suggesting validity to our overall argument .
we expect the jet speed maximum occurs at the pressure level where the radiative timescale is roughly equal to the wave propagation timescale .
empirically , the pressure levels at which the equatorial zonal - mean jet speed maximizes ( the jet core pressure " ) in our simulations correlate well with which the radiative timescale is equal to a quarter of the gravity wave propagation timescale ( fig .
14 ) , i.e. , @xmath189 . based on the approximations of @xmath54 ( eq . 3 ) and @xmath63 ( eq
. 7 and 8) in section 2 , using @xmath190 , we obtain the analytical expression of jet core pressure @xmath191 : @xmath192 note that this expression is related to the prescribed formulation of radiative timescale in this study @xmath193 ( eq . 8) , where @xmath194 is a constant .. ] however , it is straightforward to generalize the above analytical expression to a more realistic case , given the pressure dependence of radiative timescale in the real atmosphere .
( 26 ) predicts that the atmospheric zonal jet occurs deeper in the atmosphere as molecular weight increases , consistent with figs . 3 and 4 .
quantitatively , if the molar heat capacity is fixed ( experiment ii ) , the jet core pressure @xmath191 scales as @xmath195 from eq .
this analytical expression matches our simulations quite well ( fig .
the radiation effect has a significant impact on the jet core pressure level . the simulations with the radiative timescale fixed ( black open circles in fig .
15 ) exhibit a jet - core peak at much lower pressures than their control cases ( black dots ) .
if radiative timescale is taken to be independent of molar heat capacity and molecular mass , the jet core pressure @xmath191 is : @xmath196 where @xmath197 and @xmath198 are the molar heat capacity and molecular weight for molecular hydrogen , respectively .
this analytical expression agrees well with our simulations ( fig .
15 ) . in this case
, the jet core pressure @xmath191 scales as @xmath199 from eq .
( 27 ) .
the generation of equatorial prograde jets has been studied for decades but this problem has not been completely solved yet . the equatorial jet formation on tidally locked exoplanets might be closely related to the standing rossby wave pattern ( @xcite ) and the resonance of the rossby waves ( @xcite ) .
normally the zonal - mean jet speed is difficult to estimate without knowing the details of the wave - mean flow interactions ( @xcite ) .
no theory currently exists that can quantitatively predict the zonal - mean zonal wind speed on synchronously rotating planets .
nevertheless , the @xcite theory , and our alternate form of that theory presented in section 4.3 and appendix a , does provide a theoretical prediction for the characteristic horizontal wind speed @xmath96 . in the maximum zonal jet speed region ,
the contribution of zonal velocity @xmath36 is much larger than the meridional velocity @xmath37 at the equator , and @xmath96 ( eq .
11 ) in the equatorial region can be approximated by : @xmath201 where @xmath191 denotes to the jet core pressure ( eq .
as noted in section 3 , @xmath202 is related to the equatorial zonal - mean zonal wind ( @xmath97 , eq .
12 ) and the eddy component ( @xmath103 , eq . 13 ) via @xmath203 .
furthermore , in the jet core region , we found that the @xmath200 is significantly larger than @xmath204 .
in fact , the zonal - mean kinetic energy occupies about 95% of the total kinetic energy in the jet core region , i.e. , @xmath205 .
therefore , we can approximate the maximum zonal - mean jet speed ( the jet core speed " ) @xmath200 using @xmath202 .
note that this approximation only works in the jet core region where the kinetic energy is primarily dominated by the zonal - mean component .
@xmath202 should decrease with molecular weight because ( 1 ) at the same pressure level , @xmath96 decreases with @xmath11 ( fig .
11 ) ; and ( 2 ) the pressure of the equatorial jet core @xmath191 increases with @xmath11 ( fig .
15 ) but @xmath96 decreases with increasing pressure ( fig .
12 ) . using eq .
( 22 ) and ( 24 ) , we can analytically predict @xmath202 and thus the maximum zonal - mean zonal wind @xmath200 . for hydrogen atmosphere , @xmath200 is on the order of 1000 @xmath206 and that for @xmath1 atmosphere is on the order of 100 @xmath206 .
the predicted equatorial jet core speeds as a function of molecular weight are approximately consistent with simulations within a factor of 2 ( fig .
the effect of heat capacity can be inferred from the difference between the standard cases ( experiment i ) and the cases with heat capacity fixed ( experiment ii ) . because the variation of the molar heat capacity around 1000 k is generally within a factor of three ( from @xmath207 for helium to @xmath208 for carbon dioxide , see fig .
1 ) , the effect is generally limited .
furthermore , the two important parameters in our analytical theory in appendix a , @xmath108 and @xmath109 , are a function of @xmath209 . in an isothermal atmosphere , @xmath210 ( eq . 8) and @xmath211 ( eq .
3 ) , implying that @xmath209 does not depend on @xmath12 . as a result ,
both day - night temperature contrast and global wind speed are not strongly influenced by the molar heat capacity , as seen in fig .
7 and 10 , respectively . however , the molar heat capacity seems to have effects on the zonal - mean zonal wind pattern in our simulations , as seen in experiment iv based on the mass-16 " case with different heat capacity ( fig .
it shows that the zonal - mean equatorial jet speed decreases with increasing heat capacity , but the variation is within a factor of 2 .
the most significant impact is the jet core pressure .
cases in experiment iv exhibit a dramatic change of the jet core pressure while the molar heat capacity only varies by a factor of 2 ( from @xmath19 to @xmath66 ) . the circulation pattern at 700 pa ( fig .
17 ) transits from the equatorial - jet - dominated regime in the @xmath19 case to a day - to - night flow regime in the @xmath66 case .
that the jet core pressure moves deeper into the atmosphere in the simulations contradicts our theory of jet core pressure ( eq .
26 ) in section 4.5 , which predicts that the jet core pressure @xmath191 should scale with @xmath212 and thus be smaller for a higher @xmath12 atmosphere . as @xmath12 increases , the radiative timescale @xmath213 , eq .
8) increases more than the wave timescale ( @xmath211 , eq .
3 ) in an isothermal atmosphere , implying a less radiative - control atmosphere and the zonal jet should form at a _ lower _ pressure level . furthermore
, @xmath214 is a very weak dependence given the variation of @xmath12 within a factor of 2 . neither the sign nor the magnitude of the @xmath12 dependence in fig .
17 are consistent with eq .
( 26 ) , albeit which has successfully explained the principal trend of the regime shift and its dependence on the molecular weight ( fig .
15 ) . what leads to the discrepancy between experiment iv and our theory in section 4.2 ?
we hypothesize that it is the assumption of an isothermal atmosphere used for wave propagation timescale ( eq .
a more realistic estimate of @xmath54 should be based on the @xmath47 given in eq .
if the atmospheric temperature is close to adiabatic , a small change of the molar heat capacity might greatly reduce the gravity wave speed @xmath47 in the atmosphere and inhibit equatorial jet formation .
in this situation , the dependence of jet core pressure on @xmath12 could be significant , as in experiment iv .
we separately analyzed the wave propagation timescales on the dayside and the nightside , and this analysis suggests that the planetary - scale kelvin / rossby waves are more easily suppressed on the nightside than the dayside . using the local temperature and its gradient with pressure , we locally calculated @xmath47 at each ( longitude , latitude ) point over the globe . for each pressure level , the _
mean _ wave propagation timescales for _ each hemisphere _ is approximated by @xmath215 where @xmath216 is obtained by spatially averaging the @xmath47 values for each hemisphere .
these averages include only points within @xmath217 latitude of the equator .
( 18 ) shows that @xmath54 decreases on both the dayside and the nightside as the molar heat capacity increases , but the change of @xmath54 is much more dramatic on the nightside ( more than an order of magnitude ) than on the dayside ( less than 50% ) .. this is based on the assumption that that @xmath47 is constant with height . if @xmath47 varies with height ( as in fig .
18 , nightside ) , then the appropriate wave speed would involve a an appropriate vertical integral of @xmath44 with height over the range of heights associated with that wave mode
. therefore the extremely short wave speeds ( and corresponding long wave timescales ) at @xmath218 pa in fig .
18 ( orange dashed curve ) might not be realistic ( although the wave speed could still be somewhat short ) , because the wave speed might be significantly influenced by adjacent layers that are more stratified .
] this is because the nightside hemisphere is less vertically stratified than the dayside
the nightside atmosphere has a larger vertical temperature gradient ( @xmath53 ) . ]
the inequity originates from that the day - night temperature contrast ( @xmath219 ) decreases with increasing pressure ( section 4.1 ) .
] based on eq .
( 2 ) , the wave propagation speed is smaller in the nightside .
the fractional change of @xmath47 due to @xmath12 increase could also be larger on the nightside .
the quantity @xmath47 is more sensitive on the nightside than on the dayside to variations in @xmath12 , allowing @xmath12 to exert a greater influence on wave dynamics on the nightside than on the dayside .
this sensitivity results simply from the fact that the thermal profile @xmath220 is closer to an adiabat on the nightside than on the dayside , especially in the @xmath221 case ( fig .
18 ) . invoking our previous theory that the jet core pressure correlates with the pressure where the radiative timescale is roughly equal to a quarter of the wave propagation timescale in section 4.2
, we expect the jet core pressure to be given by the pressure where the vertical profiles of @xmath222 and @xmath54 intersect .
the radiative timescales are plotted on top of @xmath54 as a function of pressure in fig .
interestingly , the trend of the crossing points as function of the molar heat capacity shows an opposite behavior between the dayside and nightside . on the dayside , the crossing point moves towards the upper atmosphere as the molar heat capacity increases . on the other hand
, on the nightside , the crossing point moves deeper into the atmosphere , consistent with the jet core pressure trend in experiment iv , implying that nightside wave dynamics might be more important in controlling the transition from the day - to - night flow to the jet - dominated flow when the atmospheric mean temperature is close to adiabatic .
sub - jupiter size planets dominate the population of currently known planets . due to the limitation of the observational techniques , we expect most discovered planets will be close to their host stars .
this emerging population shall exhibit a larger - than - ever variety of chemical compositions in their atmospheres .
here we took a preliminary step to study the influence of their bulk composition on the planetary - scale atmospheric circulation and the temperature distributions .
future transit measurements , possible direct imaging observations and doppler imaging techniques will provide a vast amount of information on their atmospheres and may test the systematic behaviors predicted in this study .
we have mainly investigated the influences on atmospheric dynamics from two independent properties of the bulk composition : molecular weight and molar heat capacity .
as the molecular weight increases , the atmosphere tends to have a larger day - night temperature contrast , a smaller rms wind speed , a smaller eastward phase shift in the thermal phase curve , and a narrower equatorial super - rotating jet that moves to deeper into the atmosphere .
the zonal - mean zonal wind is smaller and more likely to exhibit a latitudinally alternating pattern in a higher - molecular - weight atmosphere .
we also found if the vertical temperature profile is close to adiabatic , the molar heat capacity will play a significant role in controlling the transition from a divergent flow in the upper atmosphere to a jet - dominated flow in the lower atmosphere . ,
solid ) , dayside gravity wave propagation ( dotted ) and nightside gravity wave propagation ( dashed ) , as a function of pressure .
the colors represent four molar heat capacity cases : @xmath19 ( black ) , @xmath64 ( blue ) , @xmath65 ( green ) , and @xmath66 ( orange ) . from low to high heat capacity cases
, the crossing points between the radiative timescale and wave propagation timescale moves towards upper atmosphere in the dayside but towards deeper atmosphere in the nightside.,scaledwidth=45.0% ] the underlying physical mechanisms have been discussed both qualitatively and quantitatively .
the influences of molecular weight and molar heat capacity mainly originate from the interplay between the radiation and wave dynamics .
the molecular weight also has a direct impact on the wind speed .
we summarize our diagnosis as below : \(1 ) most of the trends in our simulations of possible bulk compositions are due to the change of molecular weight , mainly because , for plausible compositions , molecular weight varies over a much wider range than the molar heat capacity .
\(2 ) the temperature contrast between the dayside and nightside increases with the molecular weight . as the molecular weight increases
, the atmosphere exhibits a temperature structure that becomes closer to radiative equilibrium . @xcite and @xcite showed that the day - night temperature difference is controlled by a competition between the tendency of zonally propagating planetary - scale waves to mute the day - night temperature difference , and the tendency of the day - night radiative - heating contrast to increase it .
the wave speeds decrease with increasing molecular weight , which lessens the ability of the waves to adjust the thermal structure .
therefore , this wave - adjustment mechanism becomes more easily suppressed leading to larger day - night temperature differences at high molecular weight .
\(3 ) the wind speeds decrease with increasing molecular weight .
this results from the fact that large molecular weight implies small scale height , which implies a smaller day - night difference in the geopotential ( on isobars ) than would occur in a low - molecular - mass atmosphere .
this causes a weaker pressure - gradient force in the horizontal momentum equation , and therefore leads to weaker winds .
\(4 ) the peak of thermal phase curve is shifted prior to the secondary eclipse due to a strong superrotating jet at equator .
the phase shift is controlled by the ratio of atmospheric radiative timescale versus advection timescale of the zonal - mean zonal wind .
the phase shift is smaller if the molecular weight is larger because the radiative timescale decreases and advection timescale increases with the molecular weight .
\(5 ) the pressure level at which the zonal - mean jet speeds reach a maximum roughly coincides with the pressure level where the radiative timescale is equal to a quarter of the wave propagation timescale in our simulations , implying that the jet core pressure is strongly controlled by both radiative effects and dynamical effects . in the jet core region
, the maximum zonal - mean jet speed @xmath200 can be approximately by @xmath202 , which decreases with molecular weight .
\(6 ) as molecular weight increases , the equatorial jet width becomes narrower .
more alternating wind patterns emerge in the higher latitude in the zonal - mean zonal wind distribution .
the @xmath1 atmosphere exhibits five alternating wind patterns in the zonal - mean zonal wind map , consistent with the rhines scale argument in @xcite .
@xcite predicted that the half - width of the equatorial - jet scales with the equatorial deformation radius .
we find that this theory explains the jet widths in our simulations ; when the molecular mass is greater , the equatorial deformation radius is smaller , and the equatorial jet is narrower .
\(7 ) consistent with previous work ( @xcite ) , the longitude - latitude wind pattern shifts from a day - to - night divergent flow pattern in the upper atmosphere to the jet - dominated pattern in the lower atmosphere . as molecular weight increases
, the transition occurs deeper in the atmosphere .
\(8 ) the effect of molar heat capacity is generally small except when the atmospheric temperature profile is close to adiabatic , which occurs most easily on the nightside .
the effect becomes important primarily when the molar heat capacity is large .
in this situation , a slight increase of the molar heat capacity might greatly reduce wave speed in the atmosphere , particularly in the nightside atmosphere , and significantly influence the jet formation in the upper observable atmosphere . as a result ,
the pressure of the transition from a day - night divergent flow to a jet - dominated flow moves deeper into the atmosphere .
several analytical theories have been presented in this study .
we presented a modified form of the theory from @xcite that yields predictions of the day - night temperature difference and root - mean - square wind speeds @xmath96 as a function of radiative time constant , frictional drag time constant , rotation rate , and other parameters .
this theory explains well the dependence of day - night temperature difference and wind speed with molecular weight .
we presented an analytic estimate of the eastward offset of the dayside hot spot if the equatorial jet speed is known .
we showed that the equatorial jet width is consistent with the equatorial deformation radius .
we also estimated the pressure ( @xmath191 ) at which the zonal - mean equatorial jet reaches a maximum .
we furthermore showed that the analytic estimate of @xmath96 provides a reasonable estimate to the dependence of the peak zonal-_mean _ zonal wind @xmath200 with molecular weight .
this should provide constraints for future theories of what controls the equilibrated speed of the zonal - mean zonal wind , which have yet to be constructed .
all of the above estimates are closely related to possible observables on tidally locked exoplanets . in terms of observational implications , aside from the trends that were summarized above and might be tested in statistics with a large sample of data , we also emphasize two points here , related to the hazes and clouds which are recently identified on several super - earths .
first , we found a correlation between the atmospheric molecular weight and the jet speed and the width of alternating zonal - mean wind pattern in the atmosphere .
that might shed light on a new way to infer the atmospheric composition from the wind pattern observations which might be achieved from a precise doppler imaging measurement ( e.g. , @xcite ) .
this method might be typically useful for the planetary atmospheres covered by the hazes and clouds , which flatten the spectrum and make it difficult to infer broad spectral features diagnostic of the composition , as in the case of gj1214b ( @xcite ) .
the general circulation simulations with more realistic radiative transfer scheme from ( @xcite ) also show that the jet pattern is significantly distinct between the water atmosphere and the carbon dioxide atmosphere .
second , if the atmospheric radiative equilibrium situation is close to adiabatic , as seen in some high molar heat capacity simulations , the nightside atmosphere has a large tendency to trigger convection above the zonal jet region once the temperature gradient drops so that the richardson number @xmath223 is under 0.25 ( @xcite ) . in general , haze and cloud particles might be preferentially formed in the nightside of a colder environment .
if the convection occurs in the nightside atmosphere , it will enhance the vertical mixing of those particles in the atmosphere ( @xcite ; @xcite ; @xcite ) .
it might help to loft the haze / cloud particles to the upper atmosphere and stabilize the low - pressure haze layers observed on several super earths / mini - neptunes ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ) .
finally , it should be noted that this work is highly idealized with a simple newtonian cooling radiative scheme in order to highlight the underlying physical mechanisms .
the opacity issue , although crucially important , is very complicated and beyond the scope of this paper . to the first order , different opacity corresponds to different pressure level where @xmath224 in eq .
( 6 ) and implies different pressure dependence of the radiative timescale .
the quantitative prediction related to the radiative timescale would thereby change accordingly .
but the underlying mechanisms discussed in this study remain the same .
we also did not explore the other parameters such as the situation of non - synchonized rotating planets ( @xcite ) , eccentric orbits ( @xcite ; @xcite ; @xcite ) , or planets with non - zero obliquities . in those situations ,
a whole range of atmospheric behaviors will emerge and the dynamics in a larger parameter space is far more rich than what we have discussed here .
as shown in @xcite in shallow water equations and @xcite in the three - dimensional primitive equations , we can obtain analytical expressions for a proxy of day - night temperature difference ( @xmath225 ) and horizontal wind speed @xmath161 , which should be considered as the rms wind velocity @xmath96 , on tidally locked planets . the horizontal pressure - gradient force associated with
the horizontal day - night temperature difference drives the flow .
this force will be balanced by the largest of the remaining forces , namely , frictional drag , coriolis force , horizontal momentum advection , and vertical momentum advection ( @xcite ) .
this leads to distinct expressions for the day - night temperature difference and wind speed in each regime . here
, we construct a modified version of this theory in which the four regimes are combined into a single expression that naturally transitions from one regime to another as appropriate .
we adopted the same assumptions as in @xcite to simplify the primitive equations .
we assume that the day - night temperature difference extends vertically over a log pressure range of @xmath226 where @xmath227 is the pressure level deep in the atmosphere at which the temperature is horizontally homogeneous . the magnitude of the horizontal acceleration term in the momentum equation ( eq .
1a ) can be estimated as @xmath228 ( @xcite ) .
for the thermodynamical equation ( eq .
1d ) , following @xcite , we adopted the weak - temperature gradient " ( wtg ) approximation ( @xcite ) , in which the horizontal advection of the temperature is neglected .
assume @xmath229 and the scaling equation set is : @xmath230 see @xcite for the derivation of eq .
( a. 1a ) and eq .
( a. 1b ) .
here we retained a factor of 2 in both equations in order to obtain a more quantitative expression for the theory - simulation comparison .
we should emphasize that here we do not intend to solve the primitive equations ( eq .
1 ) . the formulation of eq . ( a.1 ) is only for a purpose of combining the possible dominant terms : advection , rotation and drag , to achieve a uniform expression for the same theory that has been derived in @xcite .
hopefully the uniform expression is more general and might apply to the transitions between those four limiting regimes .
combine the above three equations yields the solutions for @xmath167 and @xmath161 : @xmath231 where @xmath232 can be considered as the cyclostrophic wind speed induced by an equilibrium day - night temperature difference .
the non - dimensional parameters @xmath108 and @xmath109 are defined as : @xmath110 where @xmath42 is the timescale for wave propagation .
@xmath233 is the advective timescale due to the equilibrium cyclostrophic wind " @xmath163 .
note that eq .
( a.3 ) can also be written as : @xmath162 therefore , the ratio of the wind speed to the equilibrium cyclostrophic wind " @xmath163 solely depends on a single parameter @xmath164 .
+ the parameter @xmath108 is related to the coriolis term which is characterized by the timescale of @xmath234 , and the drag term which is characterized by the timescale of @xmath111 .
@xmath109 is related to the advection term characterized by the timescale of @xmath235 . by comparing the magnitudes of the three timescales ,
we can obtain the limits of eq .
( a.4 ) and ( a.5 ) .
\(1 ) when drag dominates , @xmath111 is much smaller than @xmath234 and @xmath235 , and the parameter @xmath108 is reduced to @xmath236 : @xmath237 \(2 ) in the coriolis - dominated regime , rotation dominates , and therefore @xmath234 is much smaller than @xmath111 and @xmath235 .
the parameter @xmath108 is reduced to @xmath238 in the inviscid limit : @xmath239 \(3 ) in the advection - dominated regime , @xmath235 is much smaller than @xmath234 and @xmath111 , which is equivalent to @xmath240 . in the above three regimes ,
the day - night temperature difference is : @xmath241 similarly , for the wind speed @xmath161 , we have corresponding limits : @xmath242 one can show that the above expressions for day - night temperature difference are consistent with the ones derived in @xcite .
we emphasize that the characteristic wind speed @xmath161 we derived here should not be confused with the zonal - mean zonal wind @xmath92 , and should instead be considered as the rms wind velocity @xmath96 , although in the jet core region , the two values might be close ( section 4.6 ) .
we construct a simple kinematic model for the thermal phase shift in the presence of a specified zonal jet . for simplicity ,
we first consider an 1d solution . in the zonal direction ,
the thermodynamic equation can be reduced to a 1d equation with respect to longitude @xmath82 .
we adopt the newtonian cooling approximation for the radiative scheme by assuming a constant radiative timescale @xmath63 and a constant advective " timescale @xmath243 where @xmath92 is the zonal - mean zonal wind .
the kinematic equation is : @xmath244 where @xmath80 is the equilibrium temperature , the distribution of which could be given by eq .
( 9 ) for a tidally locked planet . since we have decoupled the latitude and altitude with longitude in this 1d problem , we assume : @xmath245 where we assume the equilibrium temperature in the nightside is flat : @xmath246 .
@xmath86 is the global mean temperature and @xmath247 is the temperature difference between the sub - stellar and the anti - stellar points . here we define @xmath248 as the longitude of the substellar point and solve eq .
( b.1 ) in steady state in the domain @xmath249 $ ] : @xmath250 where @xmath147 is the ratio between the radiative timescale and the advective timescale .
the general solution is : @xmath251 where @xmath252 .
@xmath253 , @xmath254 and @xmath255 are determined by the continuity of solutions at terminators @xmath256 and periodic boundary condition @xmath257 .
the final solution of eq .
( b.1 ) can be expressed as : @xmath258 where @xmath152 , @xmath259 and different ratios between the radiative timescale and the advective timescale @xmath147 .
the red , orange , green , blue and black colors corresponds to @xmath260 and 10 , respectively .
the vertical dashed lines indicate the longitudes of the temperature maximum , i.e. , thehot spot".,scaledwidth=45.0% ] the analytical temperature distribution is not a purely sinusoidal function ( fig .
the temperature peak ( hot spot " ) on the dayside shifts relative to the equilibrium temperature pattern . when the zonal wind is eastward , @xmath148 is positive , leading to an eastward shifted hot spot in the downwind direction .
the shape of temperature curve is significantly controlled by @xmath147 ( fig .
the longitudinal shift of the hot spot becomes larger as @xmath146 increases .
the day - night temperature difference is roughly reduced by a factor of @xmath261 or @xmath262 compared with the radiative equilibrium situation . as @xmath146 increases , the dayside temperature decreases and nightside temperature increases , resulting a smaller day - night temperature difference . although both this theory and the theory in appendix a can be used to explain the temperature difference between the dayside and nightside , they are very different . the theory in appendix a is a fully predictive theory from the first principle .
but the theory we introduced here requires information of zonal - mean zonal wind @xmath92 , the theory of which has not been developed yet .
we can see two extreme cases where @xmath146 approaches either of the two limits : in the strong radiation regime where @xmath263 or in the weak radiative regime where @xmath264 .
if the radiative timescale is very short relative to the advective timescale , i.e. , @xmath265 , the temperature should be strongly controlled by the radiative relaxation .
one can show that @xmath266 and @xmath267 and the solution retains the equilibrium temperature distribution @xmath268 in the dayside and @xmath269 in the nightside . on the other hand ,
if the radiative timescale is very long compared with the advective timescale , i.e. , @xmath264 , the advection will efficiently distribute the heat across longitudes .
one can show that @xmath270 and @xmath271 . in this limit ,
the temperature becomes constant with longitude and @xmath272 .
it is straightforward to generalize the solution for a 3d atmosphere with the 3d equilibrium temperature distribution in eq .
the solution is : @xmath273 where @xmath274 is given by eq .
( b.6 ) . to compare with our 3d numerical results
, we adopted @xmath275 and @xmath63 the same as that in the experiment i cases ( section 3 ) .
however , the advection timescale is assumed as @xmath276 where @xmath277 is the planetary radius and @xmath92 is the zonal - mean zonal wind .
since a first - principle analytical theory of @xmath92 has not yet been developed , we used the equatorial - mean @xmath92 from the numerical results from experiment i. with those parameters ( table 3 ) , the analytical temperature distribution agrees well with the gcm results ( fig . 20 ) .
ccccccc @xmath3 & 294.7 & 471.8 & 5.1 & 43756.6 & 1999.1 & 8553.9 + @xmath10 & 294.7 & 471.8 & 0.03 & 3043.4 & 176.6 & 96845.8 and bottom for @xmath1 .
the advection timescale @xmath148 used in eq .
( b.7 ) is derived from the simulations .
the parameters used in the analytical calculation are in table 3.,scaledwidth=80.0% ] ( black ) and @xmath1 ( red ) for experiment i at 700 pa based on the temperature distributions in fig .
the solid curves are derived from the 3d simulation and dashed curves are calculated based on the analytical solution from eq .
the parameters used in the analytical calculation are in table 3 .
the primary transit occurs at phase @xmath142 and the secondary eclipse occurs at phase @xmath143.,scaledwidth=45.0% ] the analytical thermal phase curves can be derived based on analytical thermal phase curve based on the 2d temperature solution ( eq .
b.7 ) and the procedure described in section 4.2 ( eqs . 17 and 18 ) .
the analytical phase curves agree well with that derived from 3d simulations ( fig .
the phase curve amplitudes from the analytical solution are slightly larger than the numerical results but the overall curve shapes , especially the phase shifts , match pretty well .
our theory explains the numerical results better when @xmath146 is smaller , as in the @xmath1 case .
this physically based , analytical 2d temperature solution captures key features for calculating the thermal phase curves , which can be characterized by three parameters in the atmosphere : @xmath145 and the ratio @xmath146 between @xmath63 and @xmath148 .
( b.7 ) might be useful in atmospheric phase curve analysis for synchronously rotating exoplanets . )
is computed from eq .
( b.8 ) and the red is from @xmath278.,scaledwidth=45.0% ] by setting the derivative of temperature @xmath279 in eq .
( b.5 ) , one can derive the longitude of the local temperature maximum @xmath141 on the dayside .
this leads to the condition : @xmath280 the analytical solution of @xmath141 can not be shown in an explicit form , but it depends solely on @xmath146 . as @xmath146 increases , the hot spot shifts more to the east ( fig .
19 ) . in an eastward flow ,
the hot spot location is eastward of the substellar point , but by a smaller amount than the value @xmath252 . in other words ,
the hot spot location is westward of @xmath281 .
when the hot spot shift is small ( @xmath141 is small ) , @xmath282 is close to unity and @xmath283 is generally a good approximation of the hot spot location .
this situation occurs when the radiative timescale is smaller than the advective timescale ( @xmath146 small ) . when the advective timescale is smaller than the radiative timescale ( @xmath146 large ) , the hot spot location predicted from eq .
( b.7 ) could be 10 to 20 degrees west of the longitude @xmath281 .
the analytical thermal phase curves are derived based on analytical thermal phase curve based on the 2d temperature solution ( eq .
b.7 ) and the procedure described in section 4.2 ( eqs . 17 and 18 ) .
we found a relationship between the longitudinal hot spot offset @xmath141 and the thermal phase shift ( @xmath140 ) before the secondary eclipse . when expressed in degrees , both @xmath141 and @xmath140 increase with @xmath146 ( fig .
22 ) in a similar trend .
when the phase shift is less than about 55 degrees , @xmath141 from eq .
( b. 8) agrees perfectly with the phase shift and it might under - predict @xmath140 to up to 10 degrees when the phase shift is large . on the other hand ,
the simpler predictor @xmath252 generally over - predicts @xmath140 when the phase shift is larger than 30 degrees .
although there are three parameters that shape the thermal lightcurve , we found that the thermal phase shift is almost totally controlled by @xmath146 , whereas @xmath284 and @xmath84 have negligible influence in the parameter regime we explored in this study .
this research was partially supported by bisgrove scholar program to x.z . and nsf grant ast1313444 to a.p.s .
we thank t. kataria , t. komacek and x. tan for helpful discussions .
some of the simulations were performed on the stampede supercomputer at tacc through an allocation by xsede . | super earths and mini neptunes likely have a wide range of atmospheric compositions , ranging from low - molecular mass atmospheres of @xmath0 to higher molecular atmospheres of water , @xmath1 , @xmath2 , or other species
. the effect of atmospheric composition and molecular mass on the transit spectra have been extensively investigated , but almost no attention has been paid to the question of how molecular mass and composition affect the underlying atmospheric dynamics . here , we systematically investigate the effects of atmospheric bulk compositions on temperature and wind distributions for tidally locked sub - jupiter - sized planets , using an idealized three - dimensional general circulation model ( gcm ) .
the bulk composition effects are characterized in the framework of two independent variables : molecular weight and molar heat capacity .
the effect of molecular weight dominates .
as the molecular weight increases , the atmosphere tends to have a larger day - night temperature contrast , a smaller eastward phase shift in the thermal phase curve and a smaller zonal wind speed .
the width of the equatorial super - rotating jet also becomes narrower and the jet core " region , where the zonal - mean jet speed maximizes , moves to a greater pressure level .
the zonal - mean zonal wind is more prone to exhibit a latitudinally alternating pattern in a higher - molecular - weight atmosphere .
the effects of the molar heat capacity are generally small .
but if the vertical temperature profile is close to adiabatic , molar heat capacity will play a significant role in controlling the transition from a divergent flow in the upper atmosphere to a jet - dominated flow in the lower atmosphere .
six possible observables exhibit trends with molecular weight in our simulations :
( i ) day - night temperature difference ; ( ii ) thermal phase shift ; ( iii ) root - mean - square of the wind speed ; ( iv ) latitudinal widths of the equatorial jets ; and ( v ) pressure of the jet core region ; and ( vi ) maximum zonal - mean zonal jet speed
. we also present analytical theories that quantitatively explain the above trends .
those observables could be used to indirectly determine the atmospheric compositions on tidally locked sub - jupiter - sized planets . |
let @xmath0 be infinite dimensional hilbert spaces and @xmath1 a bounded linear operator .
if @xmath2 , the range of @xmath3 , is not closed it is well known that the linear operator equation @xmath4 is ill - posed , in the sense that @xmath5 , the moore - penrose generalized inverse of @xmath3 , is not bounded @xcite .
the moore - penrose generalized inverse is strongly related to the least - squares ( ls ) solutions of ( [ eq:0 ] ) .
in fact equation ( [ eq:0 ] ) has a ls solution if and only if @xmath6 belongs to @xmath7 , the domain of @xmath5 , which is defined as @xmath8 . in that case , @xmath9 is the best approximate solution ( i.e. the ls solution of minimum norm ) and the set of all ls solutions of ( [ eq:0 ] ) is given by @xmath10 . if the problem is ill - posed , then @xmath11 does not depend continuously on the data @xmath6 .
hence if instead of the exact data @xmath6 , only an approximation @xmath12 is available , with @xmath13 , where @xmath14 is the noise level or observation error , then it is possible that @xmath15 does not exist or , if it exists , then it will not necessarily be a good approximation of @xmath11 , even if @xmath16 is very small .
this instability becomes evident when trying to approximate @xmath11 by standard numerical methods and procedures .
thus , for instance , except under rather restrictive conditions ( @xcite , @xcite ) , the application of the standard ls approximations procedure on a sequence @xmath17 of finite dimensional subspaces of @xmath18 , whose union is dense in @xmath18 , will result in a sequence @xmath19 of ls approximating solutions which does not converge to @xmath11 ( see @xcite ) . moreover
, this divergence can occur with arbitrarily large speed ( see @xcite ) .
ill - posed problems must be regularized before pretending to successfully attack the problem of numerically approximating their solutions .
regularizing an ill - posed problem such as ( [ eq:0 ] ) essentially means approximating the operator @xmath5 by a parametric family of continuous operators @xmath20 , where @xmath21 is called the regularization parameter .
more precisely , for @xmath22 with @xmath23 $ ] , let @xmath24 be a continuous ( not necessarily linear ) operator .
the set @xmath25 is said to be a `` family of regularization operators '' ( fro ) for @xmath5 , if for every @xmath26 , there exists a parameter choice rule @xmath27 such that @xmath28 here the parameter choice rule @xmath29 is such that @xmath30 if @xmath26 , then @xmath11 satisfies the normal equation @xmath31 and @xmath11 can be written as @xmath32 where @xmath33 is the spectral family associated to the self - adjoint operator @xmath34 ( see @xcite , @xcite ) . however , since we are assuming that @xmath2 is not closed ( and therefore @xmath35 ) , if @xmath36 then the integral in ( [ eq : int - inv ] ) does not exist since in that case @xmath37 and @xmath38 has a pole at 0 .
moreover in this case , the operator @xmath5 defined in ( [ eq : int - inv ] ) for @xmath39 , is not bounded .
for that reason , many regularization methods are based on spectral theory and consist on defining @xmath40 where @xmath41 is a family of functions appropriately defined such that for every @xmath42 $ ] there holds @xmath43 .
let @xmath44 be a parametric family of functions @xmath45 defined for all @xmath22 .
we shall say that @xmath44 is a `` spectral regularization method '' ( srm ) , if it satisfies the following hypotheses : _ h1_. for every fixed @xmath46 @xmath47 is piecewise continuous with respect to @xmath48 , for @xmath49 ; _ h2_. there exists a constant @xmath50 ( independent of @xmath21 ) such that @xmath51 for every @xmath49 ; _
h3_. for every @xmath52 , @xmath53 it can be shown that if @xmath44 is a srm then the family of operators @xmath54 defined by @xmath55 is a fro for @xmath5 ( @xcite , theorem 4.1 ) . in this case
we shall say that @xmath56 is a
spectral regularization family " for @xmath5 .
the use of this terminology has to do with the fact that each one of its elements is defined in terms of an integral with respect to the spectral family @xmath57 associated to the operator @xmath34 .
note that given the operator @xmath3 , it is sufficient that @xmath47 be defined for @xmath58 $ ] , since @xmath59 is constant " outside that interval .
it is well known that for ill - posed problems it is not possible to reconstruct the exact solution @xmath11 with any degree of accuracy unless additional _ a - priori _ information about @xmath11 is available ( @xcite , @xcite proposition 3.11 ) . on the other hand ,
given certain _ a - priori _ information about @xmath11 , it could be desirable to know the best order of convergence ( of the regularization error @xmath60 as a function of the regularization parameter @xmath21 , or of the total error @xmath61 as a function of the noise level @xmath16 ) , that can be achieved with a regularization method under those _ a - priori _ assumptions .
conversely , given an order of convergence , one could be interested in determining the possible existence of source sets " on which a certain regularization method reaches that order of convergence . in this case
it could further be of interest to determine maximal source sets " .
all these problems are strongly related to the concepts of qualification and saturation of a regularization method ( @xcite , @xcite , @xcite , @xcite , @xcite , @xcite ) . in @xcite
the notion of qualification of a regularization method was introduced for the first time and the decisive role of this concept in relation to the order of convergence of the regularization error was shown . in the sequel , we shall simply denote with @xmath41 the srm @xmath62 .
we now recall the definition of classical qualification for srms ( see @xcite ) .
[ def : calif - clasica ] let @xmath41 be a srm and denote with @xmath63 the set @xmath64 and let @xmath65 .
if @xmath66 , we say that @xmath41 has classical qualification and in that case the number @xmath67 is called order " of the classical qualification . note that @xmath68 by virtue of _ h2 _ and therefore @xmath63 is always nonempty . in @xcite math and pereverzev
first introduced the following definition of qualification for a spectral regularization method , formalizing and extending the classical notion of the concept .
[ def : calif - mathe ] let @xmath69\rightarrow ( 0,\infty)$ ] be an increasing function .
it is said that the regularization method @xmath41 has qualification @xmath70 if there exists a constant @xmath71 such that @xmath72}{\sup}{\left\vert1-\lambda g_\alpha(\lambda)\right\vert}\rho(\lambda)\leq \gamma \,\rho(\alpha)\quad \forall \;\alpha \in ( 0,a].\ ] ] in this article we generalize the previous concept , mainly by allowing the function @xmath73 appearing in the left hand side of ( [ eq : calif - mathe ] ) to be substituted by a general function @xmath74 with similar properties .
it is important to point out that in @xcite the classical qualification " of a method was defined to be the number @xmath67 in definition [ def : calif - clasica ] ( even in the case @xmath75 ) . however , from our point of view the generalized qualification " of a method will not be a number but rather a function of the regularization parameter @xmath21 as an order of convergence in the sense of definition [ def : calif - mathe ] . in the case of srms with classical qualification of positive finite order @xmath67 , the corresponding generalized qualification will be shown to be the function @xmath76 , coinciding with the classical approach . since in the extreme cases
@xmath77 and @xmath75 that function does not define an order of convergence , we have preferred to exclude them from the definition of classical qualification ( definition [ def : calif - clasica ] ) and , accordingly , we shall say that the method does not have classical qualification .
the organization of this article is as follows . in section 2 the concepts of weak and strong source - order pair and of order - source pair
are defined and three qualification levels for srm are introduced : weak , strong and optimal . a sufficient condition for the existence of weak qualification is provided and necessary and sufficient conditions for an order of convergence to be strong or optimal qualification are given .
in section 3 , examples of all qualification levels are provided and the relationships between them and with the classical qualification and the qualification introduced in @xcite are shown .
in particular , srms having qualification in each one of the three levels and not having classical qualification are presented .
finally several implications of this theory in the context of orders of convergence , converse results and maximal source sets for inverse ill - posed problems are shown in section 4 .
it is well known that there exist srms for which the corresponding @xmath67 given in definition [ def : calif - clasica ] is infinity , e.g.truncated singular value decomposition ( tsvd ) , landweber s method and showalter s method .
however , a careful analysis leads to observe that the concept of qualification as optimal order of convergence of the regularization error remains alive underlying most of these and many other methods . in this section
we generalize the definition of qualification introduced by math - pereverzev in @xcite and thereby the notion of classical qualification of a srm .
also three different levels of qualification are introduced : weak , strong and optimal .
these levels introduce natural hierarchical categories for the srms and we show that the generalized qualification corresponds to the lowest of these levels . moreover , a sufficient condition which guarantees that a srm possesses qualification in the sense of this generalization is provided and necessary and sufficient conditions for a given order of convergence to be strong or optimal qualification are found .
we denote with @xmath78 the set of all non decreasing functions @xmath79 such that @xmath80 and with @xmath81 the set of all continuous functions @xmath82 satisfying @xmath83 and such that @xmath84 for every @xmath85 if moreover @xmath86 is increasing , then it is an _
index function _ in the sense of math - pereverzev ( @xcite ) .
let @xmath87 .
we say that `` @xmath70 precedes @xmath88 at the origin '' and we denote it with @xmath89 , if there exist positive constants @xmath90 and @xmath91 such that @xmath92 for every @xmath93 . [
def : order ] let @xmath87 .
we say that `` @xmath70 and @xmath88 are equivalent at the origin '' and we denote it with @xmath94 , if they precede each other at the origin , that is , if there exist constants @xmath95 , @xmath96 , @xmath97 such that @xmath98 for every @xmath93 . clearly ,
@xmath99 " introduces an order of equivalence in @xmath78 .
analogous definitions and notation will be used for @xmath100 .
let @xmath41 be a srm , @xmath101 , @xmath102 and @xmath103 * i ) * we say that @xmath104 is a `` weak source - order pair for @xmath41 '' if it satisfies @xmath105 * ii ) * we say that @xmath106 is a `` strong source - order pair for @xmath41 '' if it is a weak source - order pair and there is no @xmath107 for which @xmath108 in ( [ eq : o ] ) can be replaced by @xmath109 .
that is , if ( [ eq : o ] ) holds and also @xmath110 * iii ) * we say that @xmath111 is an `` order - source pair for @xmath41 '' if there exist a constant @xmath112 and a function @xmath113 with @xmath114 , such that @xmath115 in the previous definitions we shall refer to the function @xmath70 as the order of convergence " and to @xmath86 as the source function " .
the reason for using this terminology will become clear in section 4 when we shall see applications of these concepts in the context of direct and converse results for regularization methods .
the following observations follow immediately from the definitions . 1 .
if @xmath104 is a weak source - order pair for @xmath41 which is not a strong source - order pair , then there exists @xmath116 such that @xmath117 and therefore @xmath118 can not be an order - source pair for @xmath41 . thus if @xmath111 is an order - source pair and
@xmath104 is a weak source - order pair , then @xmath104 is further a strong source - order pair in the sense of * _ ii_)*. 2 .
let @xmath87 .
if @xmath104 is a weak source - order pair for @xmath41 and @xmath89 then @xmath119 is also a weak source - order pair for @xmath41 .
if @xmath104 is a weak source - order pair for @xmath41 and @xmath120 is such that there exists @xmath121 for which @xmath122 for every @xmath107 , then @xmath123 is also a weak source - order pair for @xmath41 . in the following definition
we introduce the concept of generalized qualification and three different levels of it .
[ def : calif-3 ] let @xmath41 be a srm . * i ) * we say that @xmath70 is `` weak or generalized qualification of @xmath41 '' if there exists a function @xmath86 such that @xmath104 is a weak source - order pair for @xmath41 .
* ii ) * we say that @xmath70 is `` strong qualification of @xmath41 '' if there exists a function @xmath86 such that @xmath104 is a strong source - order pair for @xmath41 .
* iii ) * we say that @xmath70 is `` optimal qualification of @xmath41 '' if there exists a function @xmath86 such that @xmath104 is a strong source - order pair for @xmath41 ( it is sufficient that @xmath104 be a weak source - order pair ) and @xmath111 is an order - source pair for @xmath41 .
it is important to observe that weak qualification generalizes the concept of qualification introduced by math and pereverzev in @xcite and therefore , the notion of classical qualification .
in fact , if @xmath41 has continuous qualification @xmath124 in the sense of definition [ def : calif - mathe ] and @xmath125 , then the function @xmath126 is weak qualification of @xmath41 . however ,
these two notions are not equivalent .
we shall see later on that it is possible for a function to be weak qualification of a srm and not be qualification according to definition [ def : calif - mathe ] ( see comments at the end of section 3 ) .
it is timely to note here that if @xmath41 has classical qualification of order @xmath67 , then @xmath127 is weak qualification of @xmath41 and moreover @xmath128 is a weak source - order pair for @xmath41 for every @xmath129 $ ] .
conversely , if for @xmath130 , @xmath128 is a weak source - order pair for @xmath41 , then this method has classical qualification ( of order @xmath131 ) provided that @xmath132 \{@xmath133 is a weak source - order pair for @xmath41}@xmath134 .
the following result provides a sufficient condition for the existence of weak qualification of a srm .
[ teo : cond - caldebil ] let @xmath41 be a srm such that for every fixed @xmath107 , @xmath47 is decreasing in @xmath21 , for @xmath135 . *
a ) * if there exist an increasing function @xmath113 with @xmath114 , @xmath136 and @xmath96 such that for every @xmath137 @xmath138 then @xmath41 has weak qualification and in that case @xmath139 is weak qualification of the method . *
b ) * if for every @xmath22 , @xmath140 is positive and monotone decreasing for @xmath52 , then it is always possible to find @xmath141 and @xmath139 as in * a ) * satisfying for all @xmath22 . *
a ) * let @xmath113 be an increasing function with @xmath142 , @xmath136 and @xmath96 such that for every @xmath93 condition ( [ eq : cond - califdebil ] ) holds .
: there exists @xmath143 such that @xmath144 .
since @xmath145 is increasing , it follows that @xmath146 for every @xmath147 $ ] .
let @xmath116 .
then for every @xmath147 $ ] , @xmath148 since @xmath142 , there exists @xmath149 such that @xmath150 for every @xmath151,$ ] from which it follows that for every @xmath151 $ ] , @xmath152 > from ( [ eq:12 ] ) and ( [ eq:11 ] ) it follows that for every @xmath116 @xmath153 then , for any bounded @xmath154 the pair @xmath155 satisfies ( [ eq : o ] ) , i.e. , it is a weak source - order pair for @xmath41 .
thus we have proved that @xmath139 is weak qualification of @xmath41 .
: @xmath156 for every @xmath93 .
let @xmath116 .
since @xmath114 , there exists @xmath157 such that @xmath158 for every @xmath151.$ ] then @xmath159 for every @xmath160 , from what it follows that @xmath161 .
then , for any @xmath154 , @xmath162 therefore , @xmath155 is a weak source - order pair for @xmath41 , which implies that @xmath139 is weak qualification of @xmath41 .
( note that in this case any @xmath136 is weak qualification of @xmath41 . ) * b ) * let @xmath41 be a srm such that for every @xmath22 , @xmath140 is positive and monotone decreasing for @xmath52 .
for @xmath107 we define @xmath163 , where @xmath164 since for every @xmath107 , @xmath165 , it follows that given @xmath107 there exists @xmath166 such that @xmath167 for every @xmath168
. then @xmath169 , moreover @xmath170 $ ] for every @xmath107 and therefore , @xmath171 for every @xmath107 . on the other hand ,
since for every @xmath22 , @xmath140 is decreasing for @xmath107 , it follows immediately that @xmath172 is strictly increasing .
furthermore , since @xmath172 is bounded , it has countably many jump discontinuity points .
therefore , it is possible to assume , without loss of generality , that @xmath172 is continuous ( since , if it is not , we can redefine it in such a way that it be continuous , by subtracting the jumps at the discontinuity points ) .
thus @xmath173 is continuous , strictly increasing with @xmath174 .
therefore , its inverse function @xmath175 exists over the range of @xmath172 and it is strictly increasing and continuous with @xmath176 .
it is possible to extend @xmath175 to @xmath177 in such a way that it preserves all these properties .
we shall denote with @xmath141 this extension . for @xmath22
, we define @xmath178 since for every @xmath22 , @xmath140 is positive for all @xmath107 , it follows that @xmath179 is also positive .
since for every @xmath107 , @xmath180 , the definition of @xmath181 implies that @xmath182 for every @xmath107 , or equivalently , @xmath183 for every @xmath22 .
then @xmath184 for every @xmath22 and the fact that @xmath142 implies that @xmath185 .
if further @xmath186 is a non decreasing function , then @xmath187 and it suffices to define @xmath188 . on the contrary , since @xmath186 is bounded and positive with @xmath189 , there always exists a function @xmath136 such that @xmath190 for every @xmath22 , as we wanted to show .
> from the previous theorem , it follows that the srms @xmath41 such that for every @xmath107 , @xmath47 is decreasing for @xmath22 and for every @xmath22 , @xmath140 is positive and decreasing for @xmath107 , do possess weak qualification .
it is important to observe that most of the usual srms do in fact satisfy these conditions . in particular
this is so for landweber s and showalter s methods .
now given the srm @xmath41 and @xmath102 , we define @xmath191 note that @xmath192 in the next three results we will see that the characteristics of a given function @xmath193 , as a possible strong or optimal qualification of a srm , can be determined from properties of that function @xmath194 . [ teo : cond - calif](necessary and sufficient condition for strong qualification . )
a function @xmath195 such that @xmath196 is strong qualification of @xmath41 if and only if @xmath197 suppose that @xmath70 is strong qualification of @xmath41 .
then there exists a function @xmath154 such that @xmath104 is a strong source - order pair for @xmath41 .
then , for every @xmath107 , @xmath198 thus ( [ eq : cond - calif ] ) follows from ( [ eq : o ] ) and ( [ eq : no - o ] ) .
conversely , suppose now that @xmath199 for every @xmath107 .
we will show that @xmath70 is strong qualification of @xmath41 .
for that let us see that @xmath200 is a strong source - order pair for @xmath41 . since @xmath199 for every @xmath107 , it follows that @xmath201 then , @xmath194 verifies ( [ eq : o ] ) and ( [ eq : no - o ] ) , which , together with the fact that @xmath196 , implies that @xmath202 is a strong source - order pair and thus @xmath70 is strong qualification of @xmath41 .
[ teo : nu < nu - rho ] let @xmath193 be strong qualification of @xmath41 and @xmath154 . then @xmath104 is a strong source - order pair for @xmath41 if and only if there exists @xmath203 such that @xmath204 for every @xmath107 .
since @xmath70 is strong qualification , by proposition [ teo : cond - calif ] it follows that @xmath205 for every @xmath107 .
suppose now that @xmath104 is a strong source - order pair for @xmath41 .
then there exist positive constants @xmath206 and @xmath91 such that @xmath207 for every @xmath107 , @xmath208 .
then , for every @xmath107 @xmath209 and therefore @xmath210 for every @xmath107 .
conversely , suppose that there exists @xmath203 such that @xmath211 for every @xmath212 .
since @xmath205 , it then follows that @xmath213 that is , @xmath104 is a weak source - order pair for @xmath41 .
moreover since @xmath74 and @xmath214 are positive for all @xmath107 , it follows that @xmath74 verifies ( [ eq : no - o ] ) and therefore @xmath106 is , furthermore , a strong source - order pair for @xmath41 .
[ teo : cond - calopt ] ( necessary and sufficient condition for optimal qualification . )
a function @xmath193 such that @xmath196 is optimal qualification of @xmath41 if and only
if @xmath194 verifies and . suppose that @xmath70 is optimal qualification .
then @xmath70 is strong qualification and it follows from proposition [ teo : cond - calif ] that @xmath194 verifies ( [ eq : cond - calif ] ) .
moreover since @xmath70 is optimal qualification , there exists @xmath215 such that @xmath106 is a strong source - order pair and @xmath111 is an order - source pair . from the latter
it follows that there exist a constant @xmath112 and a function @xmath113 with @xmath114 , such that @xmath216 on the other hand , since @xmath106 is a strong source - order pair for @xmath41 , it follows from proposition [ teo : nu < nu - rho ] that there exists @xmath203 such that @xmath217 > from ( [ eq:2 ] ) and ( [ eq:3 ] ) it follows that @xmath218 that is , @xmath194 satisfies ( [ eq:4.50 ] ) as we wanted to show .
conversely , suppose that @xmath196 verifies ( [ eq:4.50 ] ) and ( [ eq : cond - calif ] ) . by proposition [ teo : cond - calif ]
we have that @xmath200 is a strong source - order pair for @xmath41 and ( [ eq:4.50 ] ) implies that @xmath219 is an order - source pair .
then , @xmath70 is optimal qualification of @xmath41 .
next we will show the uniqueness of the source function .
[ teo : unica nu ] if @xmath70 is optimal qualification of @xmath41 then there exists at most one function @xmath86 ( in the sense of the equivalence classes induced by definition [ def : order ] ) such that @xmath104 is a strong source - order pair and @xmath118 is an order - source pair for @xmath41 .
moreover if @xmath220 , then @xmath194 is such a unique function . given that @xmath70 is optimal qualification of @xmath41 , there exists at least one function @xmath86 such that @xmath106 is a strong source - order pair and @xmath111 is an order - source pair for @xmath41 .
suppose now that there exist @xmath221 and @xmath222 such that @xmath223 and @xmath224 are strong source - order pairs and @xmath225 and @xmath226 are order - source pairs for @xmath41 .
then there exist @xmath112 and a function @xmath227 with @xmath114 , such that @xmath228 for every @xmath229 .
then , @xmath230 on the other hand , since @xmath223 is a strong source - order pair , there exist positive constants @xmath206 and @xmath91 such that @xmath231 from ( [ eq:4 ] ) and ( [ eq:5 ] ) it follows that @xmath232 since @xmath142 we have that @xmath233 for every @xmath107 .
analogously , by interchanging @xmath221 and @xmath222 it follows that there exists @xmath234 such that @xmath235 for every @xmath107 and therefore , @xmath236 .
suppose now that @xmath196 .
since @xmath70 is optimal qualification of @xmath41 it follows from theorem [ teo : cond - calopt ] that @xmath194 verifies ( [ eq:4.50 ] ) and ( [ eq : cond - calif ] ) .
then , @xmath194 is the unique function such that @xmath202 is a strong source - order pair and @xmath219 is an order - source pair for @xmath41 .
the following is a result about the uniqueness of the order .
if @xmath237 and @xmath238 are strong source - order pairs for @xmath41 and there exists @xmath239 , then @xmath240 .
suppose that @xmath237 and @xmath238 are strong source - order pairs for @xmath41 .
we will first show that @xmath241 suppose that @xmath242 since @xmath237 is a strong source - order pair we have that @xmath243 and @xmath244 it follows from ( [ eq:6 ] ) and ( [ eq:7 ] ) that the @xmath245 on the right - hand side of the previous expression must be equal to zero , which is a contradiction .
then , @xmath241 similarly , it is shown that @xmath246 since there exists @xmath239 , we then have that @xmath247 and @xmath248 . then , @xmath249 and @xmath250 , that is , @xmath251 , as we wanted to show .
in this section we present several examples which illustrate the different qualification levels previously introduced as well as the relationships between them and with the concept of classical qualification and the qualification introduced in @xcite .
although some of these examples are only of academic interest and nature , they do serve to show the existence of regularization methods possessing qualification in each one of the levels introduced in this article .
* example 1 .
* tikhonov - phillips regularization method @xmath41 , where @xmath252 has classical qualification of order @xmath253 ( @xcite ) .
we will see that @xmath254 is optimal qualification in the sense of definition [ def : calif-3 ] * _ iii)_*. in fact , for @xmath212 , @xmath255 and if @xmath254 then @xmath256 , that is , @xmath194 verifies ( [ eq : cond - calif ] ) . also since @xmath257 we have that @xmath194 verifies ( [ eq:4.50 ] ) . from theorem [ teo : cond - calopt ] it then follows that @xmath254 is optimal qualification of @xmath41 .
* example 2 .
* let @xmath41 be the family of functions associated to the truncated singular value decomposition ( tsvd ) , @xmath258 it follows that @xmath259 , where @xmath67 is as in definition [ def : calif - clasica ] .
therefore , tsvd does not have classical qualification . in this case
we have that @xmath260 let @xmath261 and @xmath193 .
then @xmath262 then , it follows from theorem [ teo : cond - caldebil].a ) that any function @xmath193 is weak qualification of the method .
however , tsvd does not have strong qualification .
in fact , for any function @xmath193 we have that @xmath263 for every @xmath264 .
proposition [ teo : cond - calif ] implies then that @xmath70 is not strong qualification of the method . in @xcite
it was observed that tsvd has arbitrary qualification in the sense of definition [ def : calif - mathe ] .
* example 3 . * for @xmath22
we define @xmath265 it can be immediately verified that @xmath41 satisfies the hypotheses _ h1-h3 _ and therefore is a srm . since @xmath266 for all @xmath49
, it follows that for every @xmath130 , @xmath267 then , @xmath41 does not have classical qualification ( more precisely @xmath259 , where @xmath67 is as in definition [ def : calif - clasica ] ) .
we will now show that @xmath268 is optimal qualification of @xmath41 .
since @xmath269 for every @xmath264 , it follows from proposition [ teo : cond - calif ] that @xmath70 is strong qualification of @xmath41 .
moreover since @xmath270 it follows that @xmath194 verifies ( [ eq:4.50 ] ) .
theorem [ teo : cond - calopt ] then implies that @xmath268 is optimal qualification of @xmath41 .
* example 4 . * for @xmath22 with @xmath271 ,
define @xmath272 clearly , @xmath41 satisfies hypotheses _ h1-h3 _ and therefore is a srm .
since @xmath273 for all @xmath274 , it follows that for every @xmath130 , @xmath275 then , @xmath77 and therefore @xmath41 does not have classical qualification .
however , we will show that @xmath276 is optimal qualification of @xmath41 .
in fact , since @xmath277 for every @xmath107 and @xmath278 it follows from theorem [ teo : cond - calopt ] that @xmath279 is optimal qualification of @xmath41 .
* example 5 .
* let @xmath41 be the tikhonov - phillips regularization method , which , as previously mentioned , it has classical qualification of order @xmath253 .
in example 1 we saw that @xmath254 is optimal qualification of this method and therefore it is also weak qualification of it . since @xmath280 it follows from definition [ def : calif-3].i ) and observation 2.a ) that @xmath281 is also weak qualification .
however , @xmath282 is not strong qualification of the method .
in fact , for any @xmath154 , we have that @xmath283 * example 6 .
* let @xmath41 be the srm defined in example 4 .
this method does not have classical qualification since @xmath77 .
we proved that @xmath284 is optimal qualification and therefore , it is also weak qualification . since @xmath285 , just like in the previous example , it follows immediately that @xmath286 is weak qualification .
let us show now that @xmath70 is not strong qualification of the method .
for any @xmath154 , we have that @xmath287 it is important to observe that if @xmath127 is strong qualification of a srm then it follows immediately from the definition of strong source - order pair that the method has classical qualification of order @xmath288 .
the converse , however , is not true as the next example shows .
hence it is the weak and not the strong qualification what generalizes the classical notion of this concept .
* example 7 .
* for @xmath22 with @xmath289 define @xmath290 and @xmath291 in this case , @xmath292 one can immediately show that @xmath41 is a srm with classical qualification of order @xmath253 .
however , @xmath254 is not strong qualification of the method .
in fact , for any @xmath154 , we can see that @xmath293 and therefore condition ( [ eq : no - o ] ) is not satisfied .
srms possessing strong but not optimal qualification , have very peculiar properties .
thus for instance , it is possible to show that if @xmath70 is strong qualification which is not optimal , then @xmath294 , the function @xmath295 it is not of bounded variation as a function of @xmath21 in any neighborhood of @xmath296 .
even so , the following three examples show the existence of srm having strong but not optimal qualification and they show that strong qualification in no case implies optimal qualification .
* example 8 . * given @xmath297 , for @xmath298 define @xmath299 so that @xmath300 it can be immediately checked that @xmath301 is a srm with classical qualification of order @xmath206 .
with @xmath302 we have that @xmath303 , @xmath304 since @xmath305 , from proposition [ teo : cond - calif ] it follows that @xmath202 is a strong source - order pair and @xmath302 is strong qualification of the method .
however , for every @xmath107 , @xmath306\;=\;0.\end{aligned}\ ] ] therefore equation ( [ eq:4.50 ] ) does not hold and @xmath302 is not optimal qualification of the method . * example 9 . * for @xmath298 define @xmath47 as follows : @xmath307 so that @xmath308 it can be immediately verified that @xmath41 is a srm which does not have classical qualification ( @xmath75 ) .
however , with @xmath309 we have that @xmath310}\\ & = & \lambda^{\frac12}.\end{aligned}\ ] ] since @xmath305 , by proposition [ teo : cond - calif ] @xmath202 is a strong source - order pair and @xmath311 is strong qualification of the method .
however , @xmath303 we have that @xmath312 and therefore ( [ eq:4.50 ] ) does not hold and @xmath313 is not optimal qualification of the method .
* example 10 .
* for @xmath314 and @xmath107 define @xmath315 so that @xmath316 just like in examples 8 and 9 it can be easily checked that @xmath317 is a srm which does not have classical qualification ( @xmath77 ) , that @xmath318 is strong but not optimal qualification of the method and that @xmath202 is a strong source - order pair with @xmath319 . note that examples 2 , 3 , 4 , 6 , 9 and 10 correspond to srms which do not have classical qualification but , however , they do have generalized qualification , falling in some of its three different levels . also landweber s method and showalter s method , which as previously pointed out do not have classical qualification ( in both cases @xmath75 ) , are srms defined by @xmath320 ( where @xmath321 and @xmath322 , respectively .
it can be easily proved , by using theorem [ teo : cond - caldebil ] , that @xmath323 is weak qualification of landweber s method and @xmath313 is weak qualification of showalter s method . however , in this last case it can be easily shown that @xmath313 does not satisfy condition ( [ eq : calif - mathe ] ) and therefore @xmath124 is not qualification in the sense of definition [ def : calif - mathe ] . the different qualification levels introduced in this article and the relationships between them are visualized in figure 1 .
the generalization of the concept of qualification of a srm introduced in the previous sections is strongly related with and it has a broad spectrum of applications in the context of orders of convergence , converse results and maximal source sets for inverse ill - posed problems .
we present next some results in this direction .
however , we point out that this is not the main objective of the present article . for that reason ,
some of this results will be stated without proof .
more detailed results in this regard will appear in a forthcoming article .
let @xmath0 be infinite dimensional hilbert spaces and @xmath1 a bounded , linear invertible operator such that @xmath2 is not closed . for @xmath324 , the set @xmath325 , will be referred to as the source set associated to the function @xmath86 and the operator @xmath3 " . in all that follows ,
the hypothesis @xmath326 can be replaced by @xmath86 continuous on @xmath327 and @xmath328 , where @xmath329 is the set of all functions @xmath330 which are measurable with respect to the measures @xmath331 for every @xmath332 .
the following direct result , whose proof follows immediately from the concept of weak source - order pair , states that if the exact solution @xmath11 of the problem @xmath333 belongs to the source set @xmath325 and @xmath104 is a weak source - order pair for @xmath41 , then the regularization error @xmath334 has order of convergence @xmath124 . for brevity reasons we do not give the proof here .
[ teo : gen-4.3 ] let @xmath335 be weak qualification of @xmath41 and @xmath154 such that @xmath104 is a weak source - order pair for @xmath41 .
if @xmath336 then @xmath337 for @xmath338 .
it is important to note here that the previous result can be viewed as a generalization of theorem 4.3 in @xcite , to the case of srm with weak qualification and general source sets .
in fact , that result corresponds to the particular case in which @xmath41 has classical qualification of order @xmath288 .
the following converse result states that if the regularization error has order of convergence @xmath124 and @xmath118 is an order - source pair , then the exact solution belongs to the source set given by the range of the operator @xmath339 .
[ teo : gen-411 ] if @xmath118 is an order - source pair for @xmath41 and @xmath337 for @xmath340 , then @xmath341 the proof follows immediately from the definition of order - source pair for the srm @xmath41 .
it is interesting to note that theorem [ teo : gen-411 ] can also be viewed as a generalization of theorem 4.11 in @xcite .
in fact , this corresponds to the particular case in which @xmath342 y @xmath343 .
if moreover @xmath70 is optimal qualification then the reciprocal of theorem [ teo : gen-411 ] also holds .
this is proved in the following theorem . if @xmath70 is optimal qualification of @xmath41 and @xmath344 , then @xmath337 for @xmath345 if and only if @xmath346 let @xmath70 be optimal qualification of @xmath41 and @xmath344 .
then by theorem [ teo : unica nu ] , @xmath219 is an order - source pair for @xmath41 and since @xmath337 for @xmath345 , it follows from theorem [ teo : gen-411 ] that @xmath347 conversely , if @xmath348 , since by virtue of theorem [ teo : unica nu ] @xmath202 is a strong source - order pair , theorem [ teo : gen-4.3 ] implies that @xmath337 for @xmath340 .
an important result regarding existence and maximality of source sets is the following : if @xmath70 is strong qualification of a srm and @xmath196 it follows from proposition [ teo : nu < nu - rho ] that @xmath349 is a maximal source set where @xmath70 is order of convergence of the regularization error .
more precisely we have the following result .
[ teo : rango ] let @xmath193 be strong qualification of @xmath41 such that @xmath196 and @xmath350 .
if @xmath104 is a strong source - order pair for @xmath41 and @xmath351 then @xmath352 .
under the hypotheses of the proposition [ teo : nu < nu - rho ] , there exists @xmath203 such that @xmath210 for every @xmath107 , which implies that @xmath353 .
if moreover @xmath70 is optimal qualification the following stronger result is obtained . if @xmath193 is optimal qualification of @xmath41 and @xmath196 , then @xmath349 is the only source set where @xmath70 is order of convergence of the regularization error of @xmath41 .
this result follows immediately from theorem [ teo : unica nu ] .
* examples : * \1 . for the tikhonov - phillips regularization method the only source set where @xmath254 is optimal qualification is @xmath354 , since in this case @xmath355 . \2 . in example 3 of section 3 we saw that @xmath268 is optimal qualification of @xmath41 and @xmath356 .
since @xmath357 it follows that @xmath354 is the only source set where @xmath70 is order of convergence of the regularization error .
in example 8 of the previous section , for @xmath302 we have that @xmath358 .
since @xmath70 is strong qualification of this srm , it follows that @xmath359 is a maximal source set where @xmath124 is order of convergence of the regularization error .
as pointed out at the end of section 3 , @xmath360 is weak qualification of showalter s method .
it can be easily shown that for every @xmath154 , @xmath104 is a weak source - order pair for the method .
therefore , it follows from theorem [ teo : gen-4.3 ] that the regularization error @xmath334 has order of convergence @xmath361 whenever @xmath362 .
. happens with landweber s method and @xmath363 .
in this article we have extended the definition of qualification for spectral regularization methods introduced by math and pereverzev in @xcite .
this extension was constructed bearing in mind the concept of qualification as the optimal order of convergence of the regularization error that a method can achieve ( @xcite , @xcite , @xcite , @xcite ) .
three different levels of generalized qualification were introduced : weak , strong and optimal .
in particular , the first of these levels extends the definition introduced in @xcite and a srm having weak qualification which is not qualification in the sense of definition [ def : calif - mathe ] was shown .
sufficient conditions for a srm to have weak qualification were provided , as well as necessary and sufficient conditions for a given order of convergence to be strong or optimal qualification .
examples of all three qualification levels were provided and the relationships between them as well as with the classical concept of qualification and the qualification introduced in @xcite were shown .
several srms having generalized qualification in each one of the three levels and not having classical qualification were presented . in particular , it was shown that the well known tsvd , showalter s and landweber s methods do have weak qualification .
finally several implications of this theory in the context of orders of convergence , converse results and maximal source sets for inverse ill - posed problems , were briefly shown .
more detailed results on these implications will appear in a forthcoming article . | the concept of qualification for spectral regularization methods ( srm ) for inverse ill - posed problems is strongly associated to the optimal order of convergence of the regularization error ( @xcite , @xcite , @xcite , @xcite ) . in this article , the definition of qualification is extended and three different levels are introduced : weak , strong and optimal . it is shown that the weak qualification extends the definition introduced by math and pereverzev ( @xcite ) , mainly in the sense that the functions associated to orders of convergence and source sets need not be the same .
it is shown that certain methods possessing infinite classical qualification , e.g. truncated singular value decomposition ( tsvd ) , landweber s method and showalter s method , also have generalized qualification leading to an optimal order of convergence of the regularization error .
sufficient conditions for a srm to have weak qualification are provided and necessary and sufficient conditions for a given order of convergence to be strong or optimal qualification are found .
examples of all three qualification levels are provided and the relationships between them as well as with the classical concept of qualification and the qualification introduced in @xcite are shown .
in particular , srms having extended qualification in each one of the three levels and having zero or infinite classical qualification are presented .
finally several implications of this theory in the context of orders of convergence , converse results and maximal source sets for inverse ill - posed problems , are shown .
* keywords .
* qualification , regularization method , inverse ill - posed problem . |
constant - energy contour maps of scattering intensity as a function of momentum transfer were measured using macs covering a wide @xmath29 range in the @xmath37 and @xmath38 scattering planes to study the splitting of the excitations around the magnetic zone centres . figs .
[ figs1](a c ) [ [ figs1](d f ) ] show the folded contour maps at zero energy , 2.0 ( 2.5 ) mev , and 5.0 mev in the @xmath38 [ @xmath37 ] plane . at @xmath870 mev [ figs . [ figs1](a ) and [ figs1](d ) ] , the magnetic bragg reflections , which are on top of the nuclear bragg peaks , can be observed as a very intense single spot at all expected magnetic zone centres . at finite energy transfers ,
the contour maps show the splitting of the minimum point of the magnon dispersion around the zone centres along the @xmath103 direction .
this splitting can be observed above the gap energy at @xmath872.0 mev for the @xmath38 plane [ fig .
[ figs1](b ) ] and 2.5 mev for the @xmath37 plane [ fig .
[ figs1](e ) ] at all magnetic zone centres . at higher energies ,
the excitations become dispersive and form two ellipses which start to overlap at @xmath39 mev .
the ellipses become larger and more elongated along @xmath103 with increasing energy transfer as shown in figs . [
figs1](c ) and [ figs1](f ) for @xmath104 mev . for comparison , the calculated scattering intensity maps plotted in figs .
[ figs1](g j ) show good agreement with the measured contour maps at the corresponding energies .
the calculated intensity is integrated using an energy width of @xmath105 mev with the mid - point at the labelled energy .
the discrepancy between the measured and calculated results , i.e. , the filled ovals for the measured data versus open circles for the calculated results , is due to the fact that the instrumental - resolution effect was ignored in the calculations .
-plane at ( a ) zero energy , ( b ) 2.0 mev , and ( c ) 5.0 mev .
another set of contour maps were measured in the @xmath37-plane at ( d ) zero energy , ( e ) 2.5 mev , and ( f ) 5.0 mev .
the data shown in ( b ) and ( c ) are background subtracted by the background that was measured at 80 k. the data in the @xmath38-plane ( @xmath37-plane ) are folded with respect to @xmath106 $ ] and @xmath107 $ ] ( @xmath107 $ ] and @xmath108 $ ] ) .
contour maps of the calculated scattering intensity are shown for the @xmath38 plane at ( g ) 2.0 mev and ( h ) 5.0 mev , and for the @xmath37 plane at ( i ) 2.5 mev and ( j ) 5.0 mev.,title="fig:",height=377 ] -plane at ( a ) zero energy , ( b ) 2.0 mev , and ( c ) 5.0 mev .
another set of contour maps were measured in the @xmath37-plane at ( d ) zero energy , ( e ) 2.5 mev , and ( f ) 5.0 mev .
the data shown in ( b ) and ( c ) are background subtracted by the background that was measured at 80 k. the data in the @xmath38-plane ( @xmath37-plane ) are folded with respect to @xmath106 $ ] and @xmath107 $ ] ( @xmath107 $ ] and @xmath108 $ ] ) .
contour maps of the calculated scattering intensity are shown for the @xmath38 plane at ( g ) 2.0 mev and ( h ) 5.0 mev , and for the @xmath37 plane at ( i ) 2.5 mev and ( j ) 5.0 mev.,title="fig:",height=264 ]
all constant-@xmath29 scans , which were used to construct the contour maps in fig . [ fig1](a ) , were fitted using an empirical dispersion convoluted with the instrumental resolution function in order to extract the excitation energies as a function of @xmath29 .
representative energy scans at several @xmath29 s around @xmath30 measured at spins are shown with the convoluted fit in fig .
[ figs2 ] .
the fitting results including those obtained from fitting the bt7 data are plotted in fig .
[ fig1](c ) .
scans measured on spins were fitted to the empirical dispersion relation convoluted with the resolution function .
the line - shape of the observed peak is governed by the convolution with the instrumental resolution , indicative of the resolution - limited peak width.,height=566 ]
the constant-@xmath29 scans at @xmath41 were performed at several temperatures below @xmath42 to study the evolution of the gap energy @xmath109 as the temperature approaches @xmath42 .
the data were fitted and convoluted with the instrumental resolution function as shown by solid lines in figs .
[ figs5](a d ) .
the obtained fitted gap energies were then plotted as a function of temperature in fig .
[ figs3](e ) .
the results exhibit a decrease of the gap energy , which correlates with the decline of the order parameter as @xmath42 is approached .
the fit to the power law @xmath110 yields @xmath111 k , in agreement with @xmath42 obtained from the order parameter fit , and @xmath112 , consistent with the exponent describing the ordered moment @xmath113 , which scales with the square root of the magnetic bragg scattering intensity @xmath114 .
this result confirms that the magnetic excitations are due to the fluctuations of the ordered moments .
furthermore , as the temperature increases toward @xmath42 , the peak width becomes broader , indicative of a shorter life - time of the excitations typical for the spin - waves . at @xmath41 .
( a d ) show constant-@xmath29 scans at four temperatures that are fitted to the empirical dispersion convoluted with the resolution function .
( e ) displays the gap energy @xmath109 ( closed symbols ) and peak width ( open symbols ) as a function of temperature .
the solid line denotes a fit the power law described in the text .
the dashed line represents the instrumental resolution.,height=566 ] and ( b ) @xmath44 .
solid lines show the high intensity modes , which were used to fit the data .
( c ) and ( d ) show the calculated intensity contour maps depicting the high - intensity modes.,height=453 ] t. all spin hamiltonian parameters are fixed to the values obtained from the global fit.,height=226 ]
to illustrate the origin of the nonreciprocal magnons in @xmath0-cu@xmath1v@xmath1o@xmath2 , we consider a simple model of an antiferromagnetic spin chain , where spin - waves can be analytically calculated .
we hope that this calculation will provide the reader with some insight , which is probably lost in the complicated spin network of @xmath0-cu@xmath1v@xmath1o@xmath2 and the numerical calculations , into the underlying mechanism giving rise to the nonreciprocal magnons .
we note that when only the nearest - neighbour spins connected by @xmath15 are considered , cu@xmath5 spins form a uniform zig - zag chain along the @xmath32 $ ] and @xmath33 $ ] directions ( fig .
[ figs9 ] ) , and hence the spin - chain model should qualitatively and sufficiently describe the magnon dispersion in @xmath0-cu@xmath1v@xmath1o@xmath2 .
the dzyaloshinskii - moriya ( dm ) interactions and anisotropic exchange interactions are introduced to the two - sublattice heisenberg antiferromagnetic spin - chain model in order to study their effects on the elementary excitations of the system .
the calculations show that the gap in the spin - wave spectrum is the result of the anisotropic exchange interactions and the nonreciprocity of the spin - wave spectrum is due to the dm interactions . as with the hamiltonian for @xmath0-cu@xmath1v@xmath1o@xmath2 , the hamiltonian for this spin - chain model
is given by @xmath115 where the summation @xmath13 between spin operators is taken over the nearest - neighbor bonds ( fig .
[ figs6 ] ) .
the first and second terms represent the isotropic exchange interactions @xmath116 and anisotropic exchange interactions @xmath117 , respectively .
the third term is the antisymmetric dm interactions , and the last term represents the interactions between the spins and the external magnetic field @xmath118 .
we note that the one - dimensional lattice spacing is equal to one .
that point along the @xmath119 direction , exchange interaction @xmath116 , and anisotropic exchange interaction interaction @xmath117 between the nearest neighbors .
a unit cell with two spin sublattices labelled by 1 and 2 is enclosed by the orange dashed rectangle.,height=113 ] we transform coordinates from the spin local coordinates ( primed coordinates ) , in which the positive @xmath120-axis coincides with the spin direction , to the crystal coordinates ( unprimed coordinates ) using the following transformation matrices @xmath121 using the holstein - primakoff transformation , the spin operators are transformed to the boson operators @xmath8 and @xmath122 via @xmath123 where @xmath124 for the two sublattices .
the spin hamiltonian can then be written in terms of @xmath8 and @xmath122 which now can be solved by introducing the fourier transform @xmath125 where the summation runs from @xmath126 to @xmath127 . using @xmath128 as a basis and noting that @xmath129 for a _ noncentrosymmetric _ system , we can write the hamiltonian as @xmath130 we then proceed to solve for the eigenvalues and eigenvectors of this matrix .
the eigenvalues of this matrix are @xmath131 with the corresponding unnormalized eigenvectors @xmath132 hence the following pairs of the operators ( @xmath133 , @xmath134 ) , ( @xmath135 , @xmath136 ) , ( @xmath136 , @xmath135 ) , and ( @xmath134 , @xmath133 ) are mixed , and we can define the superpositions of these pairs as new boson operators @xmath137 , @xmath138 , @xmath139 , and @xmath140 ; this procedure is equivalent to the bogoliubov transformation .
the eigenvalues are physically meaningful if they are real , i.e. , the term inside the square root of @xmath141 is positive .
the dispersion as a function of momentum transfer along @xmath26 is shown in fig .
[ figs7](a ) .
we note that the eigenvalues are linearly dependent on the magnetic field @xmath142 , the effect of which is shown in fig .
[ figs7](b ) . and ( b ) @xmath143 t.,height=188 ] when the parameters @xmath117 and @xmath144 are zero , we obtain @xmath145 which corresponds to the normal antiferromagnetic spin - wave spectrum , where the clockwise and anticlockwise modes are degenerate . adding @xmath117 will open the gap ( which can be seen from @xmath146 and @xmath147 in @xmath141 ) but does not lift the degeneracy : @xmath148 and when @xmath144 is present , the degeneracy is lifted and the spin - wave spectrum will be shifted , in this case , along @xmath149 as shown in fig . [ figs7 ]
. we can calculate the spin component of the magnon excitations in the local spin coordinates for a particular wavevector @xmath127 . from eqs .
[ eqs3 ] and [ eqs4 ] , we obtain @xmath150 where @xmath151 denotes magnon modes , and @xmath152 for the spin - chain model and @xmath153 for @xmath0-cu@xmath1v@xmath1o@xmath2 . using the values of @xmath8 and @xmath122 obtained from the eigenvectors
, we can calculate and draw dynamical spin structures of the magnon excitations .
we found that for the spin - chain model the @xmath35 mode is the anticlockwise spin excitation , and the @xmath34 mode is the clockwise excitations .
figure [ figs8 ] shows the anticlockwise and clockwise spin rotations corresponding to the @xmath35 mode and the @xmath34 mode , respectively , for @xmath154 r.l.u .
, @xmath155 , @xmath156 , and @xmath157 .
when the magnetic field is present , the spins in the @xmath158 mode rotate in the same ( opposite ) sense as the precession of an electron spin under a perpendicular magnetic field and thus require less ( more ) energy to excite one magnon [ fig .
[ figs7 ] ( b ) ] .
mode , where red and magenta arrows denote the sublattice 1 and sublattice 2 , respectively .
bottom , the structure for the @xmath34 mode , where blue and green arrows denote sublattice 1 and sublattice 2 , respectively .
spins for the @xmath35 mode rotate anticlockwise whereas the spins for the @xmath34 mode rotate clockwise .
we note that the chain direction is horizontal from left to right ( @xmath159 direction ) , and that only the fluctuating spin component in the @xmath160-plane is drawn.,height=302 ]
to obtain the spin dynamics of the magnons in @xmath0-cu@xmath1v@xmath1o@xmath2 , we follow the same procedure described in the spin - chain model . out of the 16 magnon modes , we chose to draw the spin structure belonging to the two most intense magnon modes , i.e. , the @xmath35 and @xmath34 modes as shown in fig .
[ fig1 ] . in @xmath0-cu@xmath1v@xmath1o@xmath2 , spins along the zig - zag chain rotate clockwise for the @xmath35 mode , and anticlockwise for the @xmath34 mode .
we note that the @xmath161 modes refer to the shifts of the dispersion and do not imply the spin rotation . because of the difference in the chain direction between @xmath0-cu@xmath1v@xmath1o@xmath2 and the spin - chain model , the shift of the anticlockwise and clockwise modes in @xmath0-cu@xmath1v@xmath1o@xmath2 is not the same as that in the spin - chain model .
figure [ figs9 ] shows the spin dynamics of @xmath0-cu@xmath1v@xmath1o@xmath2 calculated using the fitted hamiltonian parameters at @xmath162 ( left column ) and @xmath163 ( right column ) .
when the magnetic field is applied along the @xmath164 direction , the spin rotation of the @xmath35 mode is in the opposite sense to the spin precession in the @xmath25-plane so it requires more energy to excite this mode , whereas for the @xmath34 mode the spins rotate in the same sense as the spin precession .
hence the @xmath35 mode shifts upward while the @xmath34 mode shifts downward consistent with the measured ( fig .
[ fig2 ] ) and calculated ( fig .
[ figs5 ] ) magnon dispersion .
the crystallographic @xmath36-axis is horizontal and positive to the right .
the @xmath8 axis and applied field ( positive field ) point out of the page .
the left ( right ) panels show the spin dynamics of the @xmath35 ( @xmath34 ) mode , where the spin rotation is clockwise ( anticlockwise ) along the spin chain but anticlockwise ( clockwise ) along the @xmath36-axis , which is why the shift is to the right ( left ) of the zone centre .
we only draw the @xmath25-plane component of the fluctuating spins ; the majority of the spin component is anti - aligned and points along @xmath165 .
the red and blue arrows denote the spins with the @xmath8-component along @xmath166 and @xmath164 , respectively .
the spin dynamics on each of the 4 layers in the unit cell , which is enclosed by the dotted line , is drawn separately to provide clearer images of the spin rotation .
the spins in 4 unit cells along @xmath36 are plotted as because of our choice of @xmath167 , the dynamic spin structure repeats every 4 unit cells along @xmath36 .
black arrows denote the chain directions , which are along @xmath32 $ ] for layers @xmath168 and @xmath169 , and @xmath33 $ ] for layers
@xmath170 and @xmath8.,height=491 ] -cu@xmath1v@xmath1o@xmath2 is taken from ref . .
@xmath62 and @xmath171 represent the spin - flop and spin - flip transitions , respectively , whereas @xmath172 represents the transition from the spin - flop state to the intermediate state . the green and blown asterisks denote the positions , at which the contour map in fig .
[ fig4](a ) and the constant-@xmath29 scans in figs .
[ fig4](b ) and [ fig4](c ) were measured , respectively.,height=377 ] | magnons , the spin - wave quanta , are disturbances that embody a wave propagating through a background medium formed by ordered magnetic moments . in an isotropic heisenberg system , these disturbances vary in a continuous manner around an ordered spin structure , thus requiring infinitesimal energy as a wavevector approaches a magnetic zone centre . however , competing anisotropic interactions arising from broken symmetry can favour a distinct static and dynamic spin state causing a shift of the minimum point of the magnon dispersion to a nonreciprocal wavevector @xcite . here
we report the first direct evidence of these nonreciprocal magnons in an antiferromagnet . in the antiferromagnet we investigated , namely , noncentrosymmetric @xmath0-cu@xmath1v@xmath1o@xmath2 , they are caused by the incompatibility between anisotropic exchange and antisymmetric dzyaloshinskii - moriya interactions resulting in competing collinear and helical spin structures , respectively .
the nonreciprocity introduces the difference in the phase velocity of the counter - rotating modes , causing the opposite spontaneous magnonic faraday rotation of the left- and right - propagating spin - waves .
the breaking of spatial inversion and time reversal symmetry is revealed as a magnetic - field - induced asymmetric energy shift , which provides a test for the detailed balance relation .
while symmetry plays a central role in imposing uniformity on the fundamental laws of nature @xcite , symmetry breaking introduces `` the texture of the world''@xcite by adding layers of complexity to the physical laws . in condensed matter systems , symmetry and a lack of it
determine the underlying interactions of the governing hamiltonian .
in particular , the absence of spatial inversion symmetry in magnetic systems causes the relativistic spin - orbit coupling , which gives rise to many intriguing phenomena such as the spin hall effect @xcite , topological insulators @xcite , multiferroics @xcite , and noncentrosymmetric superconductors @xcite , to acquire antisymmetric dzyaloshinskii - moriya ( dm ) interactions @xcite . for noncentrosymmetric @xmath0-cu@xmath1v@xmath1o@xmath2 ,
the crystal structure breaks spatial inversion symmetry @xcite , and the antiferromagnetic ordering below @xmath3 k @xcite breaks time reversal symmetry .
the simultaneous breaking of both symmetries sets the stage for the intertwining electric and magnetic properties @xcite and for the existence of toroidal moments @xcite . in the magnetically ordered state ,
@xmath4 cu@xmath5 spins in @xmath0-cu@xmath6v@xmath6o@xmath7 align antiparallel along the crystallographic @xmath8-axis forming a collinear structure . in the presence of a magnetic field along the @xmath9-axis ,
weak ferromagnetism resulting from the dm - interaction - induced canted moments was observed .
when the magnetic field is applied along the @xmath8-axis , two magnetic transitions appear : one transition at @xmath106.5 t is characterized as the spin - flop transition whereas the other at @xmath1118.0 t is a result of the spin - flip @xcite .
combined density functional theory ( dft ) calculations and quantum monte carlo simulations suggest a complex spin - network shown in fig .
[ fig1](d ) @xcite . to the first approximation
, the spin hamiltonian for @xmath0-cu@xmath1v@xmath1o@xmath2 can be described by @xcite @xmath12 where the summation @xmath13 @xmath14 is taken over the nearest , second - nearest , and third - nearest neighbours ( nearest neighbours ) .
the first term represents the isotropic exchange interactions , where @xmath15 , @xmath16 , and @xmath17 depicted in fig .
[ fig1](d ) are all antiferromagnetic with @xmath18 @xcite .
the second term represents the anisotropic exchange interaction @xmath19 , which arises from the multiorbital correlation effect caused by the relativistic spin - orbit coupling and multiorbital hybridization @xcite .
the third term denotes the antisymmetric dzyaloshinskii - moriya interactions @xmath20 , which results from the absence of the inversion centre between the nearest - neighbour spins @xcite .
the last term represents spins in an external magnetic field , where @xmath21 is the electron spin @xmath22-factor , @xmath23 is the bohr magneton , and @xmath24 is the magnetic field applied parallel to the crystallographic @xmath8-axis .
the anisotropic exchange interaction stabilizes the collinear antiferromagnetic spin structure and introduces an energy gap to the magnon excitations , whereas the @xmath8-component of the dm vector favours a helical spin structure in the @xmath25-plane and determines the incommensurate wavevector of the helical modulation . due to the competition between these two terms
, spin fluctuations of the dynamic state may not be around the static spin structure , giving rise to distinct wavevectors for the ordered structure and the minimum of the dynamical spin - waves . 6 , and 10 t , respectively .
dashed lines denote the most intense branches resulting from the spin - wave calculations .
( d ) shows the magnetic bragg peak measured along @xmath26 below and above the spin - flop transition ( similar data measured at different fields are reported in ) .
arrows denote the magnetic bragg peaks , and asterisks ( @xmath27 ) indicate the second harmonic reflections . ] spin dynamics of @xmath0-cu@xmath1v@xmath1o@xmath2 was investigated using single - crystal inelastic neutron scattering as described in methods . an intensity contour map as a function of energy transfer @xmath28 and momentum transfer @xmath29 measured at the base temperature along @xmath26 shows two branches of spin - wave excitations symmetrically centred on both sides of the magnetic zone centre @xmath30 , where a magnetic bragg reflection is observed [ fig .
[ fig1](a ) ] .
the two magnon modes correspond to the clockwise and anticlockwise rotation of spins bonded by @xmath15 with @xmath19 and @xmath31 along the @xmath32 $ ] and @xmath33 $ ] uniform zig - zag chains ( fig .
[ figs9 ] ) .
the degeneracy of these two modes is lifted by the dm interaction , which causes the observed symmetric shift of the magnon modes to the @xmath34 ( left ) and @xmath35 ( right ) side of the zone centre [ fig .
[ fig1](b ) ] .
the @xmath35 ( @xmath34 ) mode corresponds to the anticlockwise rotation along the @xmath36-axis but clockwise ( anticlockwise ) along the spin - chain directions ( see supplementary material ) .
constant - energy contour maps of scattering intensity ( fig . [ figs1 ] ) covering a wider range in the @xmath37 and @xmath38 scattering planes confirm that the nonreciprocity of the magnon dispersion , which was not observed in the recent inelastic neutrons scattering on a powder sample @xcite , is only along @xmath26 , which is consistent with linear spin - wave calculations [ figs .
[ figs1](g j ) ] , as will be discussed later .
the energy of both modes increases steadily with roughly the same slope up to about 10 mev , and they cross at the zone centre at @xmath39 mev .
the absence of crossing avoidance suggests that the counter - rotating excitations are decoupled . the energy scan measured at the high - resolution cold neutron spectrometer spins yields the gap energy @xmath40 mev at @xmath41 [ fig .
[ figs2](d ) ] .
the gap energy as a function of temperature correlates with the decrease of the order parameter and the peak width becomes broader , which is indicative of shorter life - time as temperature increases toward @xmath42 ( fig .
[ figs3 ] ) .
these results confirm that the excitations are due to the fluctuations of the ordered magnetic moments .
the nonreciprocal magnons were recently observed in the field - induced ferromagnetic phase of noncentrosymmetric mnsi , where the shift of the single , non - degenerate magnon mode is asymmetric depending on the field direction @xcite , in contrast to the symmetric shift in the antiferromagnet at zero field observed in this study . to quantitatively describe the observed magnon dispersion in @xmath0-cu@xmath1v@xmath1o@xmath2 , we employed linear spin - wave calculations , which are described in methods .
the calculated magnons consist of 16 modes , denoted by the grey lines in figs .
[ fig1](c ) , [ figs4](a ) , and [ figs4](b ) .
however , only two modes , which were experimentally observed in fig . [ fig1](a ) and theoretically confirmed in figs .
[ fig1](b ) , [ figs4](c ) , and [ figs4](d ) , were selected to fit the measured dispersion .
the data points in fig .
[ fig1](c ) shows the measured magnon dispersion along @xmath26 obtained from the constant-@xmath29 scans ( fig .
[ figs2 ] ) . the dispersion along @xmath43 and @xmath44 ( fig .
[ figs4 ] ) as well as the field - dependence of the energy gap [ fig .
[ fig4](d ) ] were also measured and used in the global fit to obtain the relevant hamiltonian parameters in eq .
[ spinh ] .
the ratio @xmath45 was fixed to the result obtained from the dft calculations of @xmath46 @xcite . the minimal model , which includes three isotropic exchange interactions , the anisotropic exchange interaction @xmath19 , and the uniform dm vector @xmath47 , is able to capture the magnon nonreciprocity along @xmath26 and the dispersion as shown by the solid lines in figs .
[ fig1](c ) , [ figs4](a ) , and [ figs4](b ) .
the obtained fitted parameters are @xmath48 mev , @xmath49 mev , @xmath50 mev , @xmath51 mev , and @xmath52 mev .
the value of the dm parameter with @xmath53 is much higher than that measured in other @xmath4 cu@xmath5 spin systems @xcite , suggesting the exceptionally strong spin - orbit coupling .
scans were measured at @xmath41 and @xmath54 at ( a ) 1 t and ( b ) 2 t. the dashed and dotted lines denote positions of the energy gap . in ( c ) , the scan at ( 0,1.75,0 ) and @xmath55 t is compared with that at ( 0,2.25,0 ) and @xmath56 t , which shows the same shift in energy . all ( 0,1.75,0 ) data are shifted for clarity .
the gap energy as a function of field at @xmath41 and @xmath54 is shown in ( d ) .
the critical field is labelled by @xmath57 , at which the energy gap vanishes . ]
the symmetry between the anticlockwise and clockwise modes is broken in the presence of the applied magnetic field as the electron spins precess under the perpendicular field , and if the spin rotation is in the same ( opposite ) sense as ( to ) the spin precession , or anticlockwise ( clockwise ) rotation , the excitation energy becomes lower ( higher ) as illustrated for the spin - chain model in supplementary material .
experimentally , for @xmath0-cu@xmath1v@xmath1o@xmath2 when the applied field is positive along the @xmath8-axis ( @xmath58 t ) , the @xmath35 mode ( clockwise rotation along the zig - zag chain ) is shifted upward whereas the @xmath34 mode ( anticlockwise rotation ) is shifted downward [ fig .
[ fig2](a ) ] ; on the other hand , if the field is negative ( @xmath59 t ) , the shift of the dispersion reverses [ fig .
[ fig2](b ) ] .
spin - wave calculations confirm the energy shift in the presence of the field as shown in fig .
[ figs5 ] .
the energy scans at @xmath60 t ( @xmath61 t ) shown in figs .
[ fig3](a ) [ [ fig3](b ) ] display the asymmetric shift of the gap energy at @xmath41 and @xmath54 .
figure [ fig3](c ) depicts the same energy shift of the @xmath35 and @xmath34 modes when the field of the same magnitude is oppositely aligned .
the gap energy as a function of field [ fig .
[ fig3](d ) ] shows a linear relation , consistent with the spin - chain model ( supplementary material ) , with a negative slope for @xmath41 and positive slope for @xmath54 . extrapolating the linear relation to intersect the horizontal axis yields the critical field @xmath62 of @xmath636.61(2 ) t , at which the energy gaps at @xmath64 close and the spin - flop transition occurs @xcite .
when the applied field is increased from @xmath65 t to @xmath66 t , the magnetic bragg reflection at @xmath30 , indicated by the arrow in fig .
[ fig2](a ) , moves to the incommensurate wavevectors @xmath67 where @xmath68 , as denoted by the pair of arrows in figs .
[ fig2](c ) and [ fig2](d ) . the transfer of the bragg intensity documented in fig .
[ fig2](d ) occurs at the spin - flop transition reported at @xmath69 t @xcite , and is consistent with the transition from the collinear spin structure to the helical spin structure with the majority of the spin component being in the @xmath25-plane .
the harmonic reflections at roughly @xmath70 indicated by the asterisks in fig .
[ fig2](d ) substantiate the incommensurate modulation of the helical structure .
the spin - wave excitations in the spin - flop state show the minimum at @xmath30 [ fig .
[ fig2](c ) ] indicative of the reciprocal magnons . hence
, while the collinear spin structure below @xmath62 hosts the nonreciprocal magnons , the helical spin structure above @xmath62 gives rise to the reciprocal magnons with the polarisation most likely along the @xmath8-axis , highlighting the competitive nature of the anisotropic exchange and antisymmetric dm interactions .
scans , which were performed to investigate the negative - energy - transfer and positive - energy - transfer excitations for ( b ) @xmath71 and @xmath72 t , ( c ) @xmath73 and @xmath72 t , and ( d ) @xmath71 and @xmath74 t. ] the asymmetry between the @xmath34 and @xmath35 modes provides a great opportunity to verify the well - known detailed balance relation for the dynamical structure factor , which can be described by @xcite @xmath75 where @xmath76 is the boltzmann constant , and the momentum transfer @xmath77 is measured from the zone centre .
the breaking of inversion symmetry in the @xmath0-cu@xmath1v@xmath1o@xmath2 crystal structure and of time - reversal symmetry due to the applied field introduces the asymmetry between the clockwise and anticlockwise magnon modes , with the result that @xmath78 .
the energy - momentum contour map was measured at 27 k and 10 t , where the system is in the collinear antiferromagnetic states as the critical field @xmath62 increases to @xmath79 t at @xmath80 k ( fig .
[ figs10 ] ) .
the elevated temperature is necessary to populate magnons and facilitate the neutron - energy - gain ( negative energy transfer ) scattering process .
figure [ fig4](a ) illustrates the asymmetry upon the sign reversal of @xmath77 and @xmath28 separately , reflecting broken inversion and time reversal symmetry , respectively .
the detailed measurements of the constant-@xmath29 scans extended to negative energy shown in figs .
[ fig4](b d ) display the asymmetry between the neutron - energy - gain and neutron - energy - loss scattering intensity satisfying the detailed balance relation in eq . [ bdrel ] ; the asymmetry is reversed when the momentum transfer changes from @xmath81 to @xmath82 and vice versa [ figs .
[ fig4](b ) and [ fig4](c ) ] , and when the field direction is flipped [ figs .
[ fig4](b ) and [ fig4](d ) ] .
the magnon nonreciprocity due to the dm interaction causes the clockwise and anticlockwise modes to acquire different phase velocities resulting in the rotation of the magnon polarisation , called the spontaneous magnonic faraday effect .
in contrast to a ferromagnet @xcite , the magnonic faraday rotation in the antiferromagnet is opposite for the left- and right - propagating spin - waves , thus giving rise to the net nonreciprocal phase flow in thermal equilibrium .
taking the role of the dm interaction , an external electric field , which breaks inversion symmetry and lifts the degeneracy of the two counter - rotating modes , can cause magnon nonreciprocity and lead to the electric - field - induced magnonic faraday effect , which may find applications in the spin - wave field - effect transistor @xcite . + * methods * + single crystals of @xmath0-cu@xmath1v@xmath1o@xmath2 were grown using the method described in ref . .
inelastic neutron scattering measurements were conducted to study magnetic excitations in @xmath0-cu@xmath1v@xmath1o@xmath2 using the thermal - neutron triple - axis spectrometer bt7 @xcite , the cold - neutron triple - axis spectrometer spins , the multi axis crystal spectrometer ( macs ) @xcite , and the disk chopper time - of - flight spectrometer ( dcs ) @xcite , all of which are located at the nist center for neutron research ( ncnr ) , gaithersburg , md , usa . for the bt7 measurements , a single crystal of mass 1.39 g was aligned so that @xmath83 was in the scattering plane .
the final energy was fixed at 14.7 mev with horizontal collimations of open
80@xmath84 80@xmath84 120@xmath84 .
a pg filter was placed after the sample to eliminate the higher order contamination .
the crystal was cooled to the base temperature of 2 k using a @xmath85he closed - cycle cryostat .
a series of constant-@xmath29 scans was collected at the base temperature and at 50 k for background subtraction .
the measurements at spins were performed using neutrons with a fixed final energy of 5 mev and horizontal collimations of open
80@xmath84 80@xmath84 open , yielding the energy resolution of @xmath86 mev . a cold be filter was placed in the scattered beam to cut off the neutron energy higher than 5 mev .
a series of constant-@xmath29 scans was measured in the @xmath37 and @xmath38 planes .
all scans were measured at the base temperature of 2.5 k. at macs , contour maps at constant energies were measured in the @xmath37 and @xmath38 planes .
the fixed final energy of 5 mev was selected using the vertically focusing , horizontally flat analyser with cold be filters in the scattered beam .
the base temperature of 1.7 k was achieved using a @xmath85he cryostat . in the @xmath37 plane
, the constant energy cuts were measured between 0 and 12 mev and the background was measured at 40 k and 80 k for @xmath872.5 and 5.0 mev , whereas in the @xmath38 plane , the constant energy cuts were measured between 0 and 6 mev and the background was measured at 80 k for @xmath872.0 and 5.0 mev . for the dcs experiment , the single crystal was measured in magnetic fields up to 10 t using neutrons with incident energy @xmath88 mev , yielding the instrumental energy resolution of @xmath89 mev at the elastic position .
the crystal was mounted inside a vertical - field 10-t magnet and aligned so that the crystallographic @xmath8-axis was perpendicular to the scattering plane and parallel to the applied field .
the magnet was equipped with a closed - cycle @xmath85he cryostat , which can reach the base temperature of 1.5 k. linear spin - wave calculations were performed based on the collinear , antiferromagnetic spin structure and on the spin hamiltonian in eq .
[ spinh ] using the standard holstein - primakoff transformation and fourier transform implemented using a symbolic algebraic method .
we choose @xmath90 , @xmath91 , and @xmath92 to be along the crystallographic @xmath8- , @xmath36- , and @xmath9-axis , respectively .
there are a total of 16 spins in the unit cell and hence 16 spin - wave modes as shown by the grey lines in figs .
[ fig1](c ) , [ figs4](a ) , and [ figs4](b ) .
the spin - wave energies are non - negative eigenvalues obtained from diagonalising a @xmath93 matrix . the spin - wave intensity for each mode
was calculated using @xcite @xmath94 where @xmath95 barn @xmath96 , and @xmath97 is the magnetic form factor for cu@xmath5 ions .
@xmath98 is the geometric factor , where @xmath0 and @xmath99 are vector components @xmath90 , @xmath91 , or @xmath92 .
@xmath100 is the dynamical structure factor , which is the spatial and temporal fourier transform of the spin - pair correlation function @xmath101 that can be calculated from the eigenvectors .
+ * acknowledgements * + we thank y. motome for fruitful discussion .
work at mahidol university was supported in part by the thailand research fund grant number rsa5880037 and the thailand centre of excellence in physics .
work at ncnr was supported in part by the national science foundation under agreement no .
dmr-1508249 .
work at imram was partly supported by a grant - in - aid for scientific research ( 24224009 ) from the japan society for the promotion of science ( jsps ) , and by the research program
dynamic alliance for open innovation bridging human , environment and materials " .
+ * author contributions * + g.g . , y.z . , y.q
, l.w.h .
, n.p.b . , and
k.m . performed the neutron scattering experiments .
p.p .
, t.j.s . , and
k.m . performed the linear spin - wave calculations .
g.g .
, p.p . , and k.m .
wrote the manuscript .
all authors commented on the final version of the manuscript . +
* competing interests * the authors declare that they have no competing financial interests .
+ * correspondence * correspondence and requests for materials should be addressed to k. matan ( [email protected] ) .
+ 10 url # 1`#1`urlprefix[2]#2 [ 2][]#2 , & .
_ _ * * , ( ) . , & _
_ , vol .
( , , ) . . _ _ ( )
. _ _ * * , ( ) . .
_ _ * * , ( ) . .
_ _ * * , ( ) . , , , & .
_ _ * * , ( ) . & .
_ _ * * , ( ) . , , & .
_ _ * * , ( ) .
_ et al . _ . _
_ * * , ( ) . & .
_ _ * * , ( ) . ,
& .
_ _ * * , ( ) .
_ et al . _ . _
_ * * , ( ) .
, , , & .
_ _ * * , ( ) .
, , , & .
_ _ * * , ( ) . , & .
_ _ * * , ( ) . , & .
_ _ * * , ( ) . & .
_ _ * * , ( ) .
_ et al . _
( ) . .
_ et al . _ . _
_ * * , ( ) . , & .
_ _ * * , ( ) .
_ et al . _ . _
_ * * , ( ) . & .
_ _ * * , ( ) .
_ _ ( , ) . , , , & . _ _ * * , ( ) .
_ et al .
_ .
_ _ * * , ( ) . , , & .
_ _ * * , ( ) .
_ et al . _ . _
_ * * , ( ) .
_ et al . _ . _
_ * * , ( ) . & .
_ _ * * , ( ) .
* supplementary material + nonreciprocal magnons and symmetry - breaking in the noncentrosymmetric antiferromagnet * |
the standard model ( sm ) of the strong and electroweak interactions is the present paradigm of particle physics .
its validity has been tested to a level better than one per mille at particle accelerators @xcite .
nevertheless , there are arguments against the sm being the fundamental model of particle interactions @xcite , giving rise to the investigation of competing alternative or extended models , which can be tested at high - energy colliders , such as the large hadron collider ( lhc ) , or a @xmath0 , @xmath1 international linear collider ( ilc ) .
one of the most promising possibilities for physics beyond the sm is the incorporation of supersymmetry ( susy ) , which leads to a renormalizable field theory with precisely calculable predictions to be tested in present and future experiments .
the simplest supersymmetric extension of the sm is the minimal supersymmetric standard model ( mssm ) . among the most important phenomenological consequences of susy models ,
is the prediction of new particles .
there is much excitement for the possibility of discovering these new particles at the recently built lhc @xcite , and their properties will need to be precisely measured to confirm ( or refute ) that they belong to a susy model .
this last effort might be better suited for the ilc , currently being projected .
this job needs the performance of precision measurements , but also of precision computations which are well suited for experimental comparisons . in the present work
we will focus on the properties of the susy partners of the sm quarks the squarks .
once produced , squarks will decay in a way dependent on the model parameters ( see e.g. @xcite ) .
if gluinos ( the fermionic susy partners of gluons ) are light enough , squarks will mainly decay into gluinos and quarks ( @xmath2 ) @xcite , which proceeds trough a coupling constant of strong strength .
if the mass difference among different squarks is large enough , some squarks can decay via a bosonic channel into an electroweak gauge boson and another squark ( @xmath3 ) , and if higgs bosons are light enough , also the scalar decay channels are available ( @xmath4 ) @xcite , which can be dominant for third generation squarks due to the large yukawa couplings .
otherwise , the main decay channels of squarks are their partial decays into charginos / neutralinos ( the fermionic susy partners of the electroweak gauge and higgs bosons ) and quarks ( @xmath5 ) .
some of those channels are expected to be always open , given the large mass difference between quarks and squarks , and that the charginos / neutralinos are expected to be lighter than most of squarks in the majority of susy - breaking models .
in the few cases in which these channels are closed , the squarks will decay through flavour changing neutral channels @xcite , or through three- or four - body decay channels involving a non - resonant susy particle @xcite . here
we will concentrate on the squark decay channels involving charginos and neutralinos .
their partial decay widths were computed some time ago , including the radiative corrections due to the strong ( qcd ) @xcite , and the electroweak ( ew)@xcite sectors of the theory .
these radiative corrections are large in certain regions of the parameter space@xcite , and their complicated expressions are not suitable for their introduction in the monte - carlo programs used for experimental analyses . in this work we present approximations for the partial decay widths of squarks into charginos and neutralinos , including the qcd corrections , and compare these approximations against the fixed - order one - loop corrected partial widths . in section [ sec :
notation ] we introduce our notation and conventions for particles and couplings , and set up the numerical values that we will use in our analysis , section [ sec : qcd - corrections ] presents the qcd one - loop computation of the partial decay widths and shows some numerical examples , in section [ sec : rge ] we perform a renormalization group analysis of the partial decay widths , in section [ sec : numerical ] we perform a numerical comparison of the one - loop and renormalization group computations , and finally section [ sec : conclusions ] shows our conclusions .
to describe the computation of the partial decay widths , we will follow the conventions of ref . @xcite . throughout this work
we will use a third - generation notation to describe quarks and squarks , but the analytic results and conclusions are completely general , and can be used for quarks - squarks of any generation .
we will show numerical results only for third generation quarks / squarks ( top @xmath6/stop @xmath7/bottom @xmath8/sbottom @xmath9 ) , since their decay widths are the ones that present the most interesting properties .
we will study the partial decay widths of sfermions into fermions and charginos / neutralinos , @xmath10 we denote the two sfermion - mass eigenvalues by @xmath11 , with @xmath12 .
the sfermion - mixing angle @xmath13 is defined by the transformation relating the weak - interaction ( @xmath14 ) and the mass eigenstate ( @xmath15 ) sfermion bases : @xmath16 by this basis transformation , the sfermion mass matrix , @xmath17 becomes diagonal : @xmath18 .
@xmath19 is the soft - susy - breaking mass parameter of the @xmath20 doublet due to @xmath20 gauge invariance . ] , whereas @xmath21 is the soft - susy - breaking mass parameter of the singlet . @xmath22 and @xmath23 are the usual third component of the isospin and the electric charge respectively , @xmath24 is the corresponding fermion mass , and @xmath25 is the sinus of the weak mixing angle . , @xmath26 , @xmath27 , etc . ]
the mixing parameters in the non - diagonal entries read @xmath28 @xmath29 are the trilinear soft - susy - breaking couplings , @xmath30 is the higgsino mass parameter , and @xmath31 is the ratio between the vacuum expectation values of the two higgs doublets @xmath32 .
the input parameters in the sfermion sector are then : @xmath33 for each sfermion doublet . from them
, we can derive the masses and mixing angles : @xmath34 for the trilinear couplings , we require the approximate ( necessary ) condition @xmath35 where @xmath36 is of the order of the average squark masses for @xmath37 , to avoid colour - breaking minima in the mssm higgs potential .
although the tree - level chargino ( @xmath38)-neutralino ( @xmath39 ) sector is well known , we give here a short description , in order to set our conventions .
we start by constructing the following set of weyl spinors : @xmath40 the mass lagrangian in this basis reads @xmath41 where we have defined @xmath42 with @xmath43 and @xmath44 the @xmath20 and @xmath45 soft - susy - breaking gaugino masses .
the four - component mass - eigenstate fields are related to the ones in ( [ eq : inosweak ] ) by @xmath46 where @xmath47 , @xmath48 and @xmath49 are in general complex matrices that diagonalize the mass - matrices ( [ eq : massacplusneut ] ) : @xmath50 using this notation , the tree - level interaction lagrangian between fermion - sfermion-(chargino or neutralino ) reads @xcite @xmath51 here we have adopted a compact notation , where @xmath52 is either @xmath53 or its @xmath20 partner for @xmath54 being a neutralino or a chargino , respectively .
roman characters @xmath55 are reserved for sfermion indices and @xmath56 for chargino indices ; greek indices @xmath57 denote neutralinos ; roman indices @xmath58 indicate either a chargino or a neutralino .
for example , the top - squark interactions with charginos are obtained by replacing @xmath59 , @xmath60 , @xmath61 , @xmath62 .
the coupling matrices that encode the dynamics are given by @xmath63 with @xmath64 and @xmath65 the weak hypercharges of the left - handed @xmath20 doublet and right - handed singlet fermion , and @xmath66 and @xmath67 are the yukawa couplings normalized to the @xmath20 gauge coupling constant @xmath68 .
note the following , each coupling is formed by two parts : the gaugino part , formed exclusively by gauge couplings , and the higgsino part , which contains factors of the quark masses , each of these parts will receive different kinds of corrections ( see below ) . using these definitions ,
the tree - level partial decay widths read @xmath69\,\ , , \label{eq : treleevelgamma}\end{aligned}\ ] ] with @xmath70[x^2-(y+z)^2]}$ ] . for the numerical analysis and plots
we will use fixed values for the susy parameters , and make plots by changing one parameter at a time .
for the central values of the parameters we take : @xmath71 where we have introduced a parameter @xmath72 as a shortcut for all the susy mass parameters which are not explicitly given .
we use the gut relation @xmath73 for the bino mass parameter . for the sm parameters we use @xmath74 , @xmath75 , @xmath76 , @xmath77 , @xmath78 , @xmath79 .
the renormalization scale @xmath23 is taken to be the physical mass of the decaying squark .
the value of the trilinear couplings @xmath29 is given by the algebraic expression , the given numerical value corresponds to the default values of the other parameters , this numerical value will change in the plots , the chosen expression allows to show plots with a significant parameter variation avoiding colour - breaking - vacuum conditions ( [ eq : necessary ] ) . with these input parameters , the central values for the physical susy
particle masses are : @xmath80 it is illustrative to identify the largest ew - basis component in each physical state .
of course , we have performed our computation using the full numerical mixing among the ew - basis and the physical - basis components , but this identification will help us to analyze the numerical results . the lightest squarks ( @xmath81 ) are predominantly left - handed , the lightest chargino and neutralinos ( @xmath82 ) are predominantly gaugino - like , whereas the heaviest ones ( @xmath83 ) are predominantly higgsino - like . of course ,
the parameters in eq .
( [ eq : susyparams ] ) are just and example for illustrative purposes , we have checked that our conclusions hold for a wide range of the parameter space .
following this setup , we have computed the full one - loop qcd corrections to the squark partial decay widths into charginos and neutralinos ( [ eq : gammadef ] ) .
the renormalization prescriptions follow that of ref.@xcite .
the qcd corrections include contributions from gluon loops , gluino loops , and gluon bremsstrahlung .
the full one - loop corrections have been performed using the feynarts / formcalc / looptools packages .
we have used dimensional reduction for the regularization of ultraviolet ( uv ) divergences , and a small gluon mass to regularize the infrared ( ir ) divergences .
the three - body phase - space integration of the real gluon emission is performed analytically over the full energy range , and the dependence on the gluon mass is seen to cancel between the virtual and the real corrections .
we have found full agreement with previous works , and will not repeat the full lengthy formulae here .
the corrections are seen to be numerically large , specially in certain regions of the parameter space @xcite , specially those involving processes with a bottom - squark in the initial state , and in a regime of large @xmath31 values .
we follow the hints from higgs - boson physics @xcite , and define effective yukawa couplings which should encode the leading part of the corrections@xcite : @xmath84 where @xmath85 is the running quark mass and @xmath86 is the finite threshold correction .
the susy - qcd contributions to @xmath86 are : @xmath87 where the function @xmath88 is the scalar three - point function at zero momentum transfer , and reads : @xmath89 the effective description of squark decays consists in replacing the tree - level quark masses in the couplings ( [ v1apm ] ) by the effective yukawa couplings of eq .
( [ eq : hbhbteff1 ] ) , and use this lagrangian to compute the partial decay width , schematically : @xmath90 this expression contains the large one - loop corrections from the finite threshold corrections ( [ eq : hbhbteff1 ] ) , but it also contains higher order corrections . at this point
we can make the following : make a computation that combines the higher order effects ( which ignore the effects of external momenta ) and the fixed one - loop ( which ignore the higher order effects ) . at the same time , this will allow us to quantify the degree of accuracy obtained by the effective description .
we define a _
yukawa - improved decay width computation _ : @xmath91 where @xmath92 here @xmath93 is the one - loop fixed order prediction for the partial decay width , @xmath94 is the one - loop expansion of the prediction using effective couplings , and therefore , the _ remainder _ contribution ( @xmath95 ) is the part of the one - loop contribution that can not be described by the yukawa effective couplings , it quantifies the approximation done by the effective description .
the one - loop effective prediction @xmath94 is computed by taking the computation using effective couplings ( [ eq : effdefin1 ] ) , expanding it in series , and keeping only the one - loop terms . specifically : @xmath96\ \ , \ ] ] and @xmath97 therefore , the part of the one - loop effective mass is : @xmath98\ \ , \ ] ] this is the mass that will be used in the effective yukawa couplings to compute @xmath99 , and @xmath100 is the running quark mass at the quark mass scale .
finally , we define a _
yukawa - improved _ correction factor in the following way : @xmath101 [ cols="^,^ " , ] finally fig .
[ fig : msq ] shows the evolution of the corrections to the partial decay widths as a function of the @xmath20 squark mass parameter @xmath102 .
the abrupt change which is seen at the middle of the plots corresponds with the situation in which @xmath103 , and the physical states suffer an abrupt change between left and right chirality .
this explains the difference in value and behaviour of the corrections in the regions of @xmath102 below and above that point .
this is also the situation , shown in fig .
[ fig : stopdeccormsq1 ] , where the partial decay widths can become zero , and the one - loop corrections can become non - perturbative . in all situations
the description including the log - terms provides a better description of the radiative corrections with a @xmath104 ( [ eq : defremainder ] ) much smaller than with the effective yukawa couplings ( [ eq : hbhbteff1 ] ) alone , and a much softer variation , meaning that the description of eq .
( [ eq : logterms ] ) is accurate for all values of the squark mass .
let us remember , that by changing the susy parameter @xmath102 , the physical squark masses also change , and that since the renormalization scale is taken to be the decaying squark mass , the log - terms of eq .
( [ eq : logterms ] ) effectively run also with @xmath102 as @xmath105 ( for mostly - left - handed squarks ) a contribution that can not be described with the effective yukawa couplings ( [ eq : hbhbteff1 ] ) .
we have proposed and analyzed an effective description of squark interactions with charginos and neutralinos in the mssm .
we have applied it to the partial decay widths of squarks into charginos and neutralinos .
we have compared it with the full one - loop corrections , and have proposed a way to combine the effective description ( which includes higher order terms ) with the complete one - loop description ( which includes all kinetic and mass - effects factors ) , providing an _ improved _ computation , @xmath106 ( [ improved ] ) . the difference between the effective description and the _
improved _ computation is encoded in the _ remainder _ contribution , @xmath104 eq . ( [ eq : defremainder ] ) , which gives a measure of the precision of the effective description .
the effective description includes the effective yukawa couplings ( [ eq : hbhbteff1 ] ) , which take into account the resummation effects @xcite .
note that the computation of the threshold corrections ( [ eq : deltamq ] ) includes only the ( @xmath31 , @xmath107 ) proportional terms , since the terms that would be proportional to the trilinear couplings ( @xmath108 , @xmath109 ) are actually subleading @xcite .
this description produces large _ remainder _ @xmath95 corrections , and does not reproduce the behaviour of the one - loop corrections as a function of several parameters notably it is missing a @xmath110 term .
therefore , we conclude that it does not reproduce satisfactorily the one - loop corrections , and it is not a good approximation . we have computed the missing @xmath110 terms with the help of the renormalization group eq .
( [ eq : logterms ] ) , and found agreement between the renormalization group analysis and the large mass expansion of the one - loop result ( [ eq : logterm1loop ] ) . after including also the @xmath110-terms of eq .
( [ eq : logterms ] ) , the effective description produces a reasonable approximation to the radiative - corrected partial decay widths of squarks into charginos and neutralinos , as shown by a small absolute value of the _ remainder _ contributions @xmath104 , and by a nearly - flat behaviour of the corrections as a function of different parameters .
the origin of the logarithmic terms can be explained in different ( complementary ) ways , depending on the kind of approximation that we take .
first of all , from the fundamental point of view , they are _ non - decoupling _ terms that appear due to the supersymmetry - breaking .
since we are testing susy relations ( equality of the gauge / yukawa couplings to the gaugino / higgsino couplings ) , and susy is broken ( by the term @xmath111 among others ) , we have to find some effect that tells us about the breaking of susy at that scale
e.g. a @xmath110-term .
second , from a fixed - order one - loop description point of view , the susy relation appears because the uv - divergences of the loops containing gluinos cancel with the uv - divergences of the loops containing gluons , and the log - terms that accompany those loops combine between them producing a @xmath110-term . and
third , from a renormalization group ( and effective theory ) point of view , the gaugino / higgsino couplings run different than the gauge / yukawa couplings in the region where susy is broken , that is , for scales below @xmath111 .
the running of the gaugino / higgsino couplings from the scale @xmath111 to the chosen renormalization scale produces @xmath112-terms .
since the yukawa couplings already include some qcd running , whereas the ew - gauge couplings do not , the relation between the gauge / gaugino and yukawa / higgsino couplings has some differences ( [ eq : logterms ] ) .
the presence of the non - decoupling @xmath110-terms implies a deviation of the equality between the higgsino / gaugino and higgs / gauge couplings predicted by exact susy .
this deviation is important , and has to be taken into account in the experimental measurement of susy relations . at the same time
, it gives us access to information about heavy particles that can not be directly produced at the lhc / ilc . for these reasons it is important to include these effects in the computation for the predictions of squark observables at the lhc and the ilc .
the effective description of squark / chargino / neutralino couplings given by eqs .
( [ eq : hbhbteff1 ] ) , ( [ eq : deltamq ] ) , ( [ eq : logterms ] ) , is simple to write , and to introduce in computer codes , it costs little computational power , and provides a reasonable description for squark decays into charginos and neutralinos , so it can be used in monte - carlo generators and other computer programs that provide predictions for the lhc and the ilc to improve their accuracy at a minimum cost .
j.g . and r.s.f .
have been supported in part by mec and feder under project fpa2007 - 66665c02 - 02 , j.g . also by dursi generalitat de catalunya under project 2005sgr00564 ; s.p . by a _
ramn y cajal _ contract from mec ( spain ) ( pdryc-2006 - 000930 ) and partially by cicyt ( fpa2006 - 2315 ) and dgiid - dga ( 2008-e24/2 ) ; r.s.f . also by a mec fpi grant ( bes-2005 - 8861 ) .
this work has also been supported by the spanish consolider - ingenio 2010 programme cpan ( csd2007 - 00042 ) .
j.g . wishes to thank the hospitality of the universidad de zaragoza , and s.p . wishes to thank the hospitality of the universitat de barcelona , where part of this work was done .
h. e. haber , http://arxiv.org/abs/hep-ph/9308209[`arxiv:hep-ph/9308209 ` ] , proceedings of _ recent advances in the superworld_,woodlands , usa , 13 - 16 apr 1993 , pp .
37 - 51 , eds .
j.l . lopez and d.v .
nanopoulos . j. a. aguilar - saavedra et al .
( ecfa / desy lc physics working group collaboration ) , http://arxiv.org/abs/hep-ph/0106315[`arxiv:hep-ph/0106315 ` ] , _ tesla technical design report part iii : physics at an @xmath1 linear collider _ , r. heuer , d.j .
miller , f. richard , p.m. zerwas editors .
de jong , http://arxiv.org/abs/0809.3708[`arxiv:0809.3708 [ hep - ex ] ` ] , to appear in the proceedings of _ the 16th international conference on supersymmetry and the unification of fundamental interactions _ ( susy08 ) , june 16 - 21 , 2008 , seoul , korea .
a. freitas et al . ,
suppl . * 117 * , supplement 1 , 807810 ( 2003 ) , http://arxiv.org/abs/hep-ph/0211108[`arxiv:hep-ph/0211108 ` ] , proceedings of _ 31st international conference on high energy physics ( ichep 2002 ) _ ,
amsterdam , the netherlands , 24 - 31 jul 2002 , p.807 and proceedings of _ international workshop on linear colliders ( lcws 2002 ) _ , jeju island , korea , 26 - 30 aug 2002 , p.189 .
f. del aguila et al .
j. * c57 * , 183308 ( 2008 ) , http://arxiv.org/abs/0801.1800[`arxiv:0801.1800 [ hep - ph ] ` ] , report of working group 1 of the _ cern workshop flavor in the era of the lhc : a workshop on the interplay of flavor and collider physics _ ,
geneva , switzerland , november 2005 - march 2007 .
a. j. buras , http://arxiv.org/abs/hep-ph/9806471[`arxiv:hep-ph/9806471 ` ] , in `` probing the standard model of particle interactions '' ( les houches lectures session lxviii ) , r. gupta , a. morel , e. de rafael , f. david , eds . , north holland ( 1999 ) . | we propose an effective description of squark interactions with charginos / neutralinos .
we recompute the strong corrections to squark partial decay widths , and compare the full one - loop computation with the effective description .
the effective description includes the effective yukawa couplings , and another logarithmic term which encodes the supersymmetry - breaking .
the proposed effective couplings reproduce correctly the radiative - corrected partial decay widths of the squark decays into charginos and neutralinos in all relevant regions of the parameter space . |
the single - flavour quark condensate @xmath10is a fundamental parameter of @xmath9 , determining the relative size of mass and momentum terms in the expansion .
since it can not be predicted theoretically , its value must be determined experimentally , e.g. by measuring the @xmath6scattering lengths , whose values are predicted very precisely within the framework of @xmath9 , assuming a big quark condensate @xcite , or of generalised @xmath9 , where the quark condensate is a free parameter @xcite .
the @xmath1decay is a very clean environment for the measurement of @xmath6scattering lengths , since the two pions are the only hadrons and they are produced close to threshold .
the only theoretical uncertainty enters through the constraint @xcite between the scattering lengths @xmath8and @xmath7 . in the @xmath4decay a cusp - like structure
can be observed at @xmath11 , due to re - scattering from @xmath12 .
the scattering lengths can be extracted from a fit of the @xmath5distribution around the discontinuity .
simultaneous @xmath13 and @xmath14 beams were produced by 400 gev energy protons from the cern sps , impinging on a beryllium target .
the kaons were deflected in a front - end achromat in order to select the momentum band of @xmath15 gev/@xmath16 and focused at the beginning of the detector , about 200 m downstream .
for the measurements presented here , the most important detector components are the magnet spectrometer , consisting of two drift chambers before and two after a dipole magnet and the quasi - homogeneous liquid krypton electromagnetic calorimeter .
the momentum of the charged particles and the energy of the photons are measured with a relative uncertainty of 1% at 20 gev .
a detailed description of the na48/2 detector can be found in ref .
analysing part of the 2003 data , @xmath17 @xmath1events were selected with a background contamination of 0.5% .
the background level was estimated from data , using the so - called `` wrong sign '' events , i.e. with the signature @xmath18 , that , at the present statistical level , can only be background , since the corresponding kaon decay violates the @xmath19 rule and is therefore strongly suppressed @xcite .
the main background contributions are due to @xmath12events with @xmath20 or a pion mis - identified as an electron .
the background estimate from data was cross - checked using monte carlo simulation ( mc ) .
the form factors of the @xmath1decay are parametrised as a function of five kinematic variables @xcite ( see fig .
[ cmvar ] ) : the invariant masses @xmath21 and @xmath22 and the angles @xmath23 , @xmath24 and @xmath25 .
topology of the @xmath26decay.,width=377 ] the matrix element @xmath27 contains a hadronic part , that can be described using two axial ( @xmath28 and @xmath29 ) and one vector ( @xmath30 ) form factors @xcite . after expanding them into partial waves and into a taylor series in @xmath31 ,
the following parametrisation was used to determine the form factors from the experimental data @xcite : @xmath32 in a first step , ten independent five - parameter fits were performed for each bin in @xmath21 , comparing data and mc in four - dimensional histograms in @xmath22 , @xmath33 , @xmath34 and @xmath25 , with 1500 equal population bins each .
the second step consisted in a fit of the distributions in @xmath21 ( see figs .
[ result],[delta ] ) , to extract the ( constant ) form factor parameters .
@xmath35 distribution as a function of @xmath21 .
the points represent the results of the first - step fits , the line is fitted in the second step.,width=453 ] @xmath28 , @xmath29 and @xmath30 dependence on @xmath21 .
the points represent the results of the first - step fits , the lines are fitted in the second step.,title="fig:",width=302 ] + @xmath28 , @xmath29 and @xmath30 dependence on @xmath21 .
the points represent the results of the first - step fits , the lines are fitted in the second step.,title="fig:",width=302 ] @xmath28 , @xmath29 and @xmath30 dependence on @xmath21 .
the points represent the results of the first - step fits , the lines are fitted in the second step.,title="fig:",width=302 ] + @xmath28 , @xmath29 and @xmath30 dependence on @xmath21 .
the points represent the results of the first - step fits , the lines are fitted in the second step.,title="fig:",width=302 ] the @xmath35 distribution was fitted with a one - parameter function given by the numerical solution of the roy equations @xcite , in order to determine @xmath7 , while @xmath8was constrained to lie on the centre of the universal band .
the following preliminary result was obtained : @xmath36 where the systematic uncertainty was determined by comparing two independent analyses and taking into account the effect of reconstruction method , acceptance , fit method , uncertainty on background estimate , electron - id efficiency , radiative corrections and bias due to the neglected @xmath22 dependence .
the form factors are measured relative to @xmath37 , which is related to the decay rate .
the obtained value for @xmath7is compatible with the @xmath9 prediction @xmath38 @xcite and with previous measurements @xcite .
about 10,000 @xmath3events were selected from the 2003 data and about 30,000 from the 2004 data with a background contamination of 3% and 2% , respectively .
the background level was estimated from data by reversing some of the selection criteria and was found to be mainly due to @xmath4events with a pion mis - identified as an electron ( see fig .
[ ke400 ] ) .
invariant mass distribution in logarithmic scale of the @xmath3events selected from the 2003 data ( crosses ) compared to the signal mc ( red ) plus physical ( yellow ) and accidental ( blue ) background.,width=340 ] the branching fraction was measured , as a preliminary result from the 2003 data only , normalised to @xmath4 : @xmath39 where the systematic uncertainty takes into account the effect of acceptance , trigger efficiency and energy measurement of the calorimeter , while the external uncertainty is due to the uncertainty on the @xmath4branching fraction .
this result is about eight times more precise than the best previous measurement @xcite . for the form factors the same formalism
is used as in @xmath1 , but , due to the symmetry of the @xmath40 system , the @xmath41-wave is missing and only two parameters are left : @xmath42 and @xmath43 . using the full data sample ,
the following preliminary result was obtained : @xmath44 which is compatible with the @xmath1result .
from 2003 data , about 23 million @xmath4events were selected , with negligible background .
the squared invariant mass of the @xmath40 system ( @xmath5 ) was computed imposing the mean vertex of the @xmath45s , in order to improve its resolution close to threshold . at @xmath11
, the distribution shows evidence for a cusp - like structure ( see fig .
[ cusp ] ) due to @xmath6re - scattering . left : @xmath5of the selection @xmath4data events .
the arrow indicates the position of the cusp .
right : angle between the @xmath46 and the @xmath45 in the @xmath40 centre of mass system .
the points represent the data , the three curves , the mc distribution for different values of @xmath47,width=283 ] left : @xmath5of the selection @xmath4data events .
the arrow indicates the position of the cusp .
right : angle between the @xmath46 and the @xmath45 in the @xmath40 centre of mass system .
the points represent the data , the three curves , the mc distribution for different values of @xmath47,width=283 ] in a further analysis , evidence was found for a non - zero value of @xmath47 ( see fig .
[ cusp ] ) : @xmath52 where the systematic uncertainty takes into account the effect of acceptance and trigger efficiency .
99 g. colangelo _ aip conf .
_ * 756 * , 60 ( 2005 ) .
m. knecht _ et al .
_ _ nucl .
phys . b _ * 457 * , 513 ( 1995 ) .
b. ananthanarayan _ et al .
_ _ phys .
_ * 353 * , 207 ( 2001 ) .
j. r. batley _ et al .
_ _ phys .
b _ * 634 * , 474 ( 2006 ) .
p. bloch _ et al .
_ _ phys .
b _ * 60 * , 393 ( 1976 ) .
n. cabibbo and a. maksymowicz _ phys .
_ * 137 * , b438 ( 1965 ) ; _ ibid . _
* 168 * , 1926 ( 1968 ) .
j. bijnens _ et al . _ _
2nd da@xmath53ne phisics handbook _ , 315 ( 1995 ) .
a. pais and s. b. treiman _ phys
_ * 168 * , 1858 ( 1968 ) .
g. amoros and j. bijnens
_ j. phys .
g _ * 25 * , 1607 ( 1999 ) .
g. colangelo _ et al .
_ _ nucl .
b _ * 603 * , 125 ( 2001 ) .
l. rosselet _ et al .
_ _ phys .
d _ * 15 * , 574 ( 1977 ) .
s. pislak _ et al .
_ _ phys .
d _ * 67 * , 072004 ( 2003 ) .
s. shimizu _ et al .
_ _ phys .
d _ * 70 * , 037101 ( 2004 ) .
n. cabibbo and g. isidori _ jhep _ * 0503 * , 021 ( 2005 ) .
j. r. batley _ et al .
_ _ phys .
b _ * 633 * , 173 ( 2006 ) .
s. eidelman _ et al .
_ _ phys .
b _ * 592 * , 1 ( 2004 ) . | the na48/2 experiment at the cern sps collected in 2003 and 2004 large samples of the decays @xmath0(@xmath1 ) , @xmath2(@xmath3 ) and @xmath4 . from the @xmath1form factors and from the cusp in the @xmath5distribution of the @xmath4events , the @xmath6scattering lengths
@xmath7and @xmath8could be extracted .
this measurement is a fundamental test of chiral perturbation theory ( @xmath9 ) .
the branching fraction and form factors of the @xmath3decay were precisely measured , using a much larger data sample than in previous experiments . |
observations of cosmic microwave background radiation provides a wealth of information on the cosmic reionization history ( e.g. , * ? ? ?
* ; * ? ? ?
for example , @xcite showed that the reionization occurred at @xmath8 assuming an instantaneous reionization scenario .
however , the detailed history of reionization and the nature of ionizing sources are not yet fully understood . since the uv background ( uvb )
radiation can heat up the interstellar medium ( ism ) to @xmath9k and disturb star formation , uvb coupled with the ionization history of the universe significantly influences the galaxy formation ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
therefore it is very important to study the uvb intensity and the nature of ionizing sources .
@xcite pointed out that the uvb is dominated by quasars at @xmath10 . using the sdss sample
, @xcite showed that the bright - end slope of the quasar luminosity function at @xmath11 are considerably steeper than that at lower redshifts , and concluded that the quasars can not maintain the ionization of igm at @xmath11 .
subsequently , much argument have been focused on the possibility that the igm is ionized mainly by the uv radiation from high - redshift ( hereafter high-@xmath12 ) star - forming galaxies ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the key quantity in determining the igm ionization rate is the escape fraction of ionizing photons ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , which is the number ratio of photons escaping from a galaxy to the intrinsically radiated photons by stars .
this parameter controls the contribution to the uvb intensity from star - forming galaxies . in this work
, we examine the values of @xmath0 in high-@xmath12 star - forming galaxies .
there are several observational constraints on @xmath0 at @xmath13 .
@xcite found @xmath14 from the composite spectrum of 29 lyman break galaxies ( lbgs ) at @xmath15 , where @xmath16 is the _ relative _ fraction of escaping lyman continuum ( 900 ) photons relative to the fraction of escaping non - ionizing uv ( 1500 ) photons .
it is usually defined as @xmath17 where @xmath18 , @xmath19 and @xmath20 represent the observed 1500/900 flux density ratio , the intrinsic 1500/900 luminosity density ratio , and the line - of - sight opacity of the igm for 900 photons , respectively .
equation ( [ eq : fesc_rel ] ) compares the observed flux density ratio ( corrected for the igm opacity ) with the models of uv spectral energy distribution of star - forming galaxies .
@xcite and @xcite estimated the upper limit of @xmath21 for some lbgs at @xmath15 .
@xcite directly detected the escaped ionizing photons from 2 lbgs in the ssa22 field at @xmath22 , and estimated the average value of @xmath23 .
moreover , @xcite successfully detected the lyman continuum emission from 10 ly-@xmath24 emitters ( laes ) and 7 lbgs within a sample of 198 laes and lbgs in the ssa22 field .
they showed that the mean value of @xmath16 for the 7 lbgs is 0.11 after correcting for dust extinction , and 0.20 if the igm absorption is taken into account . in the early theoretical works , some authors studied the @xmath0 with ideally modelled galaxies .
for example , @xcite estimated the @xmath0 of milky way type galaxy using a semi - analytic method , and reported @xmath25 .
@xcite investigated the dependence of @xmath0 on various physical quantities , such as the collapse redshift and star formation efficiency using a semi - analytic method . @xcite and @xcite studied the effect of inhomogeneous structure of gas on @xmath0 , and showed that @xmath0 increases in clumpy systems by a factor of @xmath26 than in a homogeneous gas distribution .
@xcite investigated the influence of bubbles made by supernovae on @xmath0 using a semi - analytic method . using numerical simulations
, @xcite studied the effect of supernovae feedback , and reported a high @xmath0 ( @xmath27 ) for a disk galaxy with @xmath28 .
theoretical studies in a more fully cosmological environment can be performed by combining cosmological hydrodynamic simulations of galaxy formation and a three - dimensional radiative transfer calculation .
for example , ( * ? ? ?
* hereafter y09 ) post - processed the eulerian hydrodynamic simulation of @xcite with rt , and showed that the galaxies in an isolated halo of @xmath29 can have relatively large values of @xmath30 .
moreover they found that @xmath0 decreases gradually as a function of time owing to the dust pollution and the shifting star formation sites . on the other hand
, @xcite examined the escape fractions of two galaxies in a cosmological sph simulation from @xmath31 to 2.4 , which later become milky way type disk galaxies at @xmath32 .
they found small values of @xmath33 , in disagreement with y09 .
however they also reported that @xmath0 decreases with redshift from @xmath31 to 2.4 , in qualitative agreement with y09 .
@xcite further examined the @xmath0 of star - forming galaxies in a wide mass range ( @xmath34 ) at @xmath35 , and found that @xmath0 decreases steeply as the halo mass increases in their cosmological sph simulations , in contrast to the work by @xcite and @xcite . using cosmological amr simulations
, @xcite reported that haloes with @xmath36 have @xmath37 , and much lower @xmath0 for lower mass haloes with @xmath38 .
their results suggest that @xmath0 increases with halo mass , at least in the range of @xmath38 .
@xcite extracted 10 haloes with masses @xmath39 at @xmath40 from cosmological amr radiation hydrodynamic simulations , and examined the escape fraction of ionizing photons .
they found that @xmath0 fluctuates rapidly on a time - scale of a few to 10 myrs depending on the star formation rates , and varies widely from almost zero to nearly unity .
they found @xmath41 for a normal imf for the haloes with @xmath42 , but @xmath43 for lower mass haloes , disregarding the effect of dust .
although the halo mass dependence of @xmath0 is very important for the study of cosmic reionization , there are significant differences in the theoretical estimates from cosmological hydrodynamic simulations as described above .
in particular , in these previous works , the number of studied haloes has been very small ( @xmath44 ) , therefore it has been difficult to gauge the halo mass dependence of @xmath0 . in the present paper , we calculate the values of @xmath0 for a much larger number of haloes ( several hundreds ) in cosmological volumes of comoving 10@xmath45100@xmath46mpc , and examine its halo mass dependence .
in addition , we study the effects of interstellar dust and uvb radiation on @xmath0 .
the outline of the paper is as follows . in [ sec : model ] , the models and numerical methods are described .
we present the results on escape fractions in
[ sec : result ] , and discuss the dust effect and the contribution of star - forming galaxies to the reionization of the universe in [ sec : discussion ] .
we then summarise in [ sec : summary ] .
ccccccc series & box - size & @xmath47 & @xmath48 & @xmath49 & @xmath50 & @xmath51 + n144l10 & 10.00 & @xmath52 & @xmath53 & @xmath54 & 2.78 & 2.75 + n216l10 & 10.0 & @xmath55 & @xmath56 & @xmath57 & 1.85 & 2.75 n400l100 & 100.0 & @xmath58 & @xmath59 & @xmath60 & 6.45 & 0.0 we use an updated and modified version of the tree - particle - mesh ( treepm ) smoothed particle hydrodynamics ( sph ) code gadget-3 ( originally described in * ? ? ? * ) .
the sph calculation is performed based on the entropy conservative formulation @xcite .
our fiducial code includes radiative cooling by h , he , and metals @xcite , star formation , supernova feedback , a phenomenological model for galactic winds @xcite , and a sub - resolution model of multiphase ism @xcite .
we also include the heating by a uniform uvb , which we will discuss more in [ sec : uvb ] . in this multiphase ism model ,
high - density ism is pictured to be a two - phase fluid consisting of cold clouds in pressure equilibrium with a hot ambient phase .
cold clouds grow by radiative cooling out of the hot medium , and this material forms the reservoir of baryons available for star formation .
the star formation rate ( sfr ) is estimated for each gas particle that have densities above the threshold density , and the star particles are spawned statistically based on the sfr . for the star formation model ,
the `` pressure model '' described in @xcite is being used .
this model estimates the sfr based on the local gas pressure rather than the gas density , and implicitly considers the effect of h@xmath61 formation .
the simulations used in this paper also uses the new `` multicomponent variable velocity '' wind model developed by @xcite .
this new wind model is based on both the energy - driven wind and the momentum - driven wind model discussed by @xcite , and the wind speed in this model depends on the galaxy stellar mass and sfr .
it gives more favorable results , and agrees better with the observations of , e.g. , cosmic civ mass density and igm temperature than the previous model with a constant wind speed . to enable this new wind model
, @xcite implemented a on - the - fly group - finder into gadget-3 to compute the galaxy masses and sfrs while the simulation is running .
the group - finder , which is a simplified variant of the algorithm developed by @xcite , identifies the isolated groups of star and gas particles ( i.e. , galaxies ) based on the baryonic density field .
the detailed procedure of this galaxy grouping is described in @xcite .
the outer baryonic density threshold for a galaxy is 0.01@xmath62 , where @xmath62 is the threshold density for star formation .
the parameters of the simulations are summarised in table [ table : sim ] . due to the computational load of the rt calculation
, we primarily use the n144l10 series for our main results . the other two series ( n216l10 and n400l100 )
are used for the resolution tests , with a fewer number of simulated galaxies postprocessed with rt .
the adopted cosmological parameters are consistent with the wmap results : @xmath63kms@xmath64mpc@xmath64 ( @xmath65 ) , @xmath66 , @xmath67 , @xmath68 , @xmath69 , and @xmath70 . to estimate @xmath0 , we compute the stellar radiation transfer and the ionization structure of gas in each dark matter halo by post - processing the simulation output .
first we set up a uniform grid around each dark matter halo with a grid cell size equal to the gravitational softening length of the simulation , and translate the sph gas information into a gridded data .
the grid typically have @xmath71300@xmath72 cells for high - mass haloes ( @xmath73 ) and @xmath7170@xmath72 cells for low - mass haloes ( @xmath74 ) .
the rt scheme used in this paper is the authentic radiation transfer ( art ) method originally developed by @xcite , and the treatment is basically the same as in y09 .
the performance of our rt code has already been reported as part of the rt code comparison study by @xcite , and it can calculate the ionization structure precisely ( see figure 6 in * ? ? ?
usually the _ short - characteristic _ method is computationally cheaper than the _ long - characteristic _ method by a factor of @xmath75 , where @xmath75 is the grid size of rt calculation . in the short - characteristic method ,
the amount of calculation is reduced by the interpolation of optical depth from the nearest grids .
however the short - characteristic method suffers from an artificial photon diffusion effect .
our art method is based on the long - characteristic method , and it is devised to reduce the calculation amount to a similar level as the short - characteristic method . hence the art method is suitable for the calculation of @xmath0 for a large number of galaxies . as for the scattering of photons , we employ the on - the - spot approximation @xcite , in which the scattered photons are assumed to be absorbed immediately on the spot . in this work ,
the rt equation is solved along @xmath76 rays with uniform angular resolution from each source .
the number of ionizing photons emitted from the source stars is computed based on the theoretical spectral energy distribution ( sed ) given by the population synthesis code pgase v2.0 @xcite .
we take only the star particles that are younger than @xmath77 yrs as the sources of ionizing photons , and consider the effect of age and metallicty of the stellar population by interpolating the table generated from the result of pgase .
we shoot the radiation rays in a radial fashion from each star particles .
we assume the @xcite initial mass function with the mass range of @xmath78 .
typically the postprocessing rt calculation takes about 100 hours for a large grid of 300@xmath72 , and 1 hour for a small grid of 70@xmath72 on a single cpu .
our art code is parallelized by mpi , and has a high parallelization efficiency .
we process each star particle in parallel , and each cpu calculates the radiation field from each star particle . in practice , we use @xmath79 cpus simultaneously to process one halo with rt .
we also include the effect of dust attenuation by distributing the interstellar dust proportionally to the metallicity , with a size distribution of @xmath80 @xcite , where @xmath81 is the radius of a dust grain .
we adopt the dust grain size range of @xmath82 m as our fiducial model .
the dust mass is calculated as @xmath83 @xcite , where @xmath84 , @xmath85 , and @xmath86 are the dust mass , gas mass , and metallicity in a grid .
the density of a dust grain is assumed to be 3g@xmath87 like silicates .
the dust opacity is given by @xmath88 , where @xmath89 , and @xmath90 are the absorption efficiency factor , dust size , number density of dust grains , and path length , respectively . since the assumed range of dust size is larger than the wavelength of lyman limit , we assume @xmath91 for ionizing photons @xcite
. the baryonic gas in galaxies can be ionized by the uvb , and heated up to @xmath92k .
it would be ideal to compute the rt of uvb as well as the stellar radiation , but in practice it is a very expensive calculation .
let us briefly explain why the uvb rt calculation is much heavier than the stellar radiation transfer . for stellar radiation , using on - the - spot approximation , we only calculate the rt along the angular rays between stars and grids with the dilution factor by the distance , whereas for uvb we have to calculate the rt of all angular rays .
therefore in the art method , the number of rays that we have to calculate is @xmath93 for stellar radiation , and @xmath94 for uvb . if @xmath95 , the calculation amount for uvb is larger than the stellar radiation . in practice , @xmath76 is greater than @xmath96 by @xmath97 orders of magnitude . due to this difficulty of uvb rt , usually a uniform uvb radiation field with an optically thin approximation
is assumed across the simulation box as a simple approximation .
our fiducial simulation also includes a uniform uvb with a modified @xcite spectrum ( see * ? ? ?
* ) , where the reionization takes place at @xmath98 as suggested by the quasar observations ( e.g. , * ? ? ?
* ) and stellar radiative transfer calculations ( e.g. , * ? ? ?
however , the optically thin approximation is very crude , and the effects of different uvb has not been explored very much .
here we use following four n144l10 simulations with different treatment of uvb to examine the effects of uvb : 1 .
* fiducial * : a uniform uvb radiation field with an optically thin approximation is assumed as stated above .
* mh0.5 * ( modified haardt 0.5 ) : the ism is optically thin to the same uvb , however the intensity of uvb is reduced to the half of the fiducial run .
* otuv * ( optically thick uv ) : the ism is optically thin to the uvb in the lower density regions with @xmath99@xmath87 , but completely optically thick in higher density region ( @xmath100 ) , where @xmath62 is the threshold density , above which star formation is allowed .
the value of @xmath62 was determined by @xcite based on the observed sf cut - off column density of the kennicutt law .
the otuv method implicitly assumes that the uvb can not penetrate into the high density regions by self - shielding .
we find that this treatment reproduces the observed hi column density distribution function very well @xcite , and more detailed analyses using rt calculation supports this self - shielding density ( yajima et al .
2010 , in preparation ) .
no - uvb * : uvb does not exist at all .
we compute the ionization structure in each halo by solving the equation of ionization equilibrium as follows : @xmath101 where @xmath102 and @xmath103 are the photoionization rate by uvb and stellar radiation ; @xmath104 is the collisional ionization rate ; @xmath105 and @xmath106 are the number density of free - electron , neutral and ionized hydrogen , respectively ; @xmath107 is the total recombination coefficient to all bound excitation levels .
the value of @xmath103 is estimated by the full rt calculation , but @xmath102 is computed with the optically thin approximation .
figure [ fig : ionize ] shows the ionization structure of gas in a high - mass halo ( halo `` a '' ) and a low - mass halo ( halo `` b '' ) in the n144l10 fiducial uvb run . the gas in halo - a shows very complex and clumpy structure , going through continuous merging processes . the young star clusters are born in dense , neutral clumpy regions and irradiate the ambient ism .
however , the dense neutral gas clumps survive owing to high recombination rates .
on the other hand , the halo - b shows more or less spherical gas distribution before we process it with rt .
star clusters are born near the central high - density region .
once the halo is processed with rt , most of the low density gas on the right - hand - side is ionized by the uvb and stellar radiation . in particular ,
when the location of a young star cluster is slightly off - center , it can ionize one side of the halo preferentially , and creates a conical region of highly ionized region .
the high density gas on the left - hand - side of the star cluster remains neutral , and the ionizing photons can not escape to the left - hand - side .
the value of angle - averaged @xmath0 of each halo is basically determined by the covering fraction of these highly ionized region .
we estimate the average value of @xmath0 for each dark matter halo as follows . for each light ray from each star particle in the halo , we count up the number of escaped ionizing photons by integrating the transferred spectrum as a function of wavelength , and then divide it by the total number of intrinsically radiated ionizing photons . then the values of @xmath0 are averaged over all the rays coming out from the halo at the surface of the grid that was set up around the halo .
hereafter @xmath0 denotes the angle - averaged value for each halo most of the time . for the n144l10 fiducial uvb run .
different colors are used for different redshifts ( red : @xmath108 , blue : @xmath109 , green : @xmath110 ) .
the triangles in the bottom right panel show the mean values in each mass bin with 1-@xmath111 error bars .
the data points with @xmath112 are shown at @xmath113 for plotting purposes . ] as a function of redshift .
different symbols indicate different halo mass ranges ( filled circles : @xmath114 , stars : @xmath115 , filled triangles : @xmath116 , open squares : @xmath117 , crosses : @xmath118 , open triangles : @xmath119 , and open circle : @xmath120 ) .
the green open circles are for a relatively massive halo ( @xmath121 at @xmath108 ) examined by @xcite , which is to be compared with our open triangles .
the red open circles are for a similarly massive halo ( @xmath5 ) in the s33 run of @xcite , which is to be compared with our crosses . ] at @xmath108 in the n144l10 fiducial uvb run .
the top panel shows the pdf of all star particles in haloes with @xmath122 , and the bottom panel is for haloes with @xmath123 .
lower mass haloes contain more star particles with larger @xmath0 . ]
figure [ fig : redshift ] shows the @xmath0 as a function of halo mass at @xmath1 for the n144l10 fiducial uvb run .
the solid triangles are the average values of @xmath0 in each mass bin with 1-@xmath111 error bars .
we first take the average in the linear scale , and then take the logarithm for the data points ( i.e. , @xmath124 ) .
at all redshifts , we find a clear qualitative trend that the mean @xmath0 declines with increasing halo mass .
our results are similar to the trend reported by @xcite , although our @xmath0 values are lower than theirs .
our results show the opposite trend to @xcite , although for high - mass galaxies , our @xmath0 is similar to theirs . in @xcite ,
most of the simulated galaxies show a disk - like structure . in their scenario , for low mass galaxies , stars are born after the disk is formed , and most young stars are embedded deep inside the disk . as a result ,
most of the ionizing photons are absorbed in the disk , and @xmath0 is low . however in more massive galaxies , some star - clusters can form near the edge of dense disk , and they can be exposed by the mergers of galaxies . as a result of this effect , they argued that more ionizing photons can escape from higher mass haloes .
therefore the geometry of simulated galaxies may cause the difference in the halo mass dependence of @xmath0 .
the galaxy sample size in @xcite and @xcite is @xmath125 , and it is somewhat small to discuss the systematic trend of @xmath0 as a function of halo mass . the spatial resolution of @xcite , which is adaptively refined depending on the gas density , is from @xmath126kpc to @xmath127pc .
although this resolution of maximum refinement level is better than that of ours , their rt scheme ( otvet ) is coarser in estimating the ionization structure @xcite .
these differences in the accuracy and resolution of fluid and rt calculations may have caused the difference in @xmath0 .
we will further discuss the possible resolution effects in section [ sec : summary ] .
figure [ fig : mass_redshift ] shows the redshift evolution of mean @xmath0 in each halo mass bin . in @xcite ,
the @xmath0 of high - mass haloes with @xmath128 clearly decreases with redshift ( blue open circles ) , and that of the low - mass haloes does not change largely . on the other hand , our results and @xcite indicate that @xmath0 of high - mass haloes with @xmath128 does not change largely with redshift . for low - mass haloes with @xmath129 , it seems that @xmath0 is increasing slightly with decreasing redshift in our simulations .
this might be due to the increasing cosmic sfr density and increasing uvb intensity from @xmath110 to @xmath108 . indeed , if we calculate the radiative transfer without the contribution of uvb in eq.(2 ) for the fiducial run at @xmath108 with the same gas and stellar distribution , @xmath0 decreases by @xmath130 per cent .
in addition , the mass fraction of gas with @xmath131 within haloes increases with increasing redshift , which leads to lower escape fraction due to higher recombination rate .
figure [ fig : pdf - mass - divided ] shows the probability distribution function ( pdf ) of star particles as a function of @xmath0 in haloes with @xmath122 ( top panel ) and @xmath123 ( bottom panel ) .
the probability is defined by @xmath132 , where @xmath96 is the number of star particles that have the value of @xmath0 , @xmath133 is the total number of source star particles , and @xmath134 is the bin width .
the figure shows that the lower mass haloes have a longer tail towards higher values of @xmath0 .
since the ionization structure in low - mass haloes shows conical regions of highly ionized gas , ionizing photons can escape easily through these ionized cones , but not through other angular directions covered by highly neutral gas .
this allows for some star particles in lower mass haloes to have high @xmath0 . on the other hand ,
the higher mass haloes show very complex and clumpy distribution of highly neutral gas , therefore it is more difficult for the ionizing photons to escape , and there are no star particles with @xmath135 .
thus the pdf for higher mass haloes is concentrated at @xmath33 .
we also find that there is a large dispersion in @xmath0 for the low - mass haloes with @xmath122 .
this result may explain some of the recent observations .
for example , @xcite and @xcite detected ionizing radiation from high-@xmath12 galaxies , with a detection rate of about 10 per cent .
the detected galaxies show extremely high @xmath0 ( @xmath136 per cent ) .
our results do not show such high values of @xmath0 , however the @xmath0 derived by @xcite and @xcite are estimated from the flux ratio at the lyman limit and uv continuum .
recently @xcite pointed out that the nebulae emission lines can boost the above flux ratio , leading to a very high @xmath0 with an assumption that the @xmath0 of nebular and stellar emission is a few tens of per cent . in our simulation sample ,
about 10 per cent show high @xmath0 ( @xmath137 ) .
these galaxies may corresponding to the recently observed objects with very high @xmath0 .
furthermore , @xcite showed that @xmath0 decreases with increasing uv flux . in our simulations ,
sfr is positively correlated with halo mass , therefore our result of decreasing @xmath0 with increasing halo mass is consistent with that of @xcite .
figure [ fig : uvall ] shows @xmath0 as a function of halo masses for different uvb model runs at @xmath108 . similarly to the fiducial uvb model run , @xmath0 decreases as the halo mass increases in all uvb models . since the ionization fraction of gas should increase with the increasing uvb intensity
, one would naively expect a higher @xmath0 for the runs with stronger uvb intensities .
however , when the uvb intensity increases , the stronger gas heating results in less efficient cooling of gas and hence less star formation .
this reduces the number of ionizing photons , and decreases the ionization fraction around the star - forming regions .
therefore these two competing effects counteract each other and self - regulate the ionization fraction , resulting in no significant differences in @xmath0 between different uvb runs , as shown in the bottom right panel of figure [ fig : uvall ] .
we also considered the possibility that the sites of star formation might be different depending on the uvb intensity .
figure [ fig : distance ] shows the probability distribution function ( pdf ) of star particles as a function of relative distance `` @xmath138 '' between each star particle and the highest gas density peak in the halo , normalized by the maximum radius of the halo ( @xmath139 ) . to focus on the scatter of @xmath0 that we see for the lower mass haloes , here we selected only the low - mass haloes that include only one young star particle
( there could be other star particles that are older in the same halo . )
this figure shows that , when the uvb intensity becomes weaker , star - forming regions are more spread out to larger radii .
therefore the weaker uvb runs have more extended tails to larger radii compared to the fiducial uvb model , progressively in the order of mh0.5 , otuv , and no - uvb run , although the probability in the extended tail is very small .
if a star cluster is farther away from the gas density peak , the value of @xmath0 would increase because the solid angle subtended by the high - density neutral clouds would become smaller . ) for the four uvb models at @xmath108 .
different colors indicate different uvb models ( red : fiducial , blue : mh0.5 , green : otuv , and magenta : no - uvb ) .
the triangles show the mean values in each mass bin with 1-@xmath111 error bars .
the data points with @xmath112 are set to @xmath113 for plotting purposes .
there is not much differences between different uvb model runs due to self - regulation effect described in the text . ] in different uvb models at @xmath108 , where @xmath138 is the relative distance between each star particle and the highest gas density peak in the halo , normalized by the maximum radius of the halo .
to focus on the scatter of @xmath0 , here we used only the low - mass haloes that include only one young star particle .
the result of the fiducial uvb run is shown by the dashed histogram in other panels for comparison .
the runs with a weaker uvb have more young star particles at larger distances from the density peak . ] to further investigate the effect of star cluster locations on @xmath0 , we plot @xmath0 as a function of @xmath138 for each halo in different uvb models in figure [ fig : fesc_distance ] . here
again , we selected only the low - mass haloes that include only one young star particle to focus on the scatter in @xmath0 of the low - mass haloes .
the gas near the density peak has high recombination rate , and therefore optically thick .
the value of @xmath0 strongly depends on the size of viewing angle to optically thick cloud from the source . as the source deviates from the central density peak , the viewing angle towards optically thick clouds decreases , and on average , @xmath0 increases with increasing @xmath138 .
moreover the mean value of @xmath0 does not depend on the uvb models very much , although the scatter is somewhat larger at larger @xmath138 values owing to the small sampling . at the lower values of @xmath138 , the scatter among different uvb models are smaller , because uvb can not penetrate into the high - density cloud by the self - shielding effect .
the results shown in figures [ fig : distance ] and [ fig : fesc_distance ] suggest that the variation in @xmath138 is one of the key factors that determines the scatter in @xmath0 for low - mass haloes . .
similarly to figure [ fig : distance ] , only the low - mass haloes that include only one young star cluster are used .
different colors indicate different uvb models ( red : fiducial , blue : mh0.5 , green : otuv , and magenta : no - uvb ) .
the triangles show the mean values in each @xmath138 bin with 1-@xmath111 error bars .
the data points with @xmath140 are set to @xmath141 for plotting purposes . ]
furthermore @xmath0 may depend on the hydrogen number density at the location of star clusters , because the neutral hydrogen gas near the star clusters can effectively absorb ionizing photons .
figure [ fig : fenh ] shows the @xmath0 of low - mass haloes as a function of hydrogen number density @xmath142 at the location of star clusters in different uvb models . in this figure we use only the low - mass haloes that include one young star particle to focus on the scatter of @xmath0 , similarly to figures [ fig : distance ] and [ fig : fesc_distance ] .
we find that @xmath0 steeply decreases with increasing local @xmath142 .
when @xmath143@xmath87 , more than 90 per cent of the emitted ionizing photons can not escape from the galaxy .
once young stars born at such a high density region , most of ionizing photons are absorbed on the spot because of high recombination rate
. therefore we find that the variation of the local hydrogen number density also causes the large scatter in @xmath0 for low - mass haloes .
this result is consistent with that shown in figure [ fig : fesc_distance ] , because we expect lower @xmath142 for grid cells at greater @xmath138 . .
similarly to figure [ fig : distance ] , only the low - mass haloes that include only one young star cluster are used .
different colors indicate different uvb models ( red : fiducial , blue : mh0.5 , green : otuv , and magenta : no - uvb ) .
the triangles show the mean values in each mass bin with 1-@xmath111 error bars .
the data points with @xmath112 are set to @xmath113 for plotting purposes . ]
we also examine the dependence of @xmath0 on other physical quantities , such as metallicity , sfr , and specific sfr ( @xmath144sfr / stellar mass ) in figure [ fig : other ] .
this figure includes all star - forming haloes , just like figure [ fig : redshift ] .
we find that @xmath0 decreases with increasing metallicity and sfr , and vice versa for specific sfr .
these correlations are expected , because the metallicity and sfr are both positively correlated with galaxy stellar mass and halo mass . however ,
if the metallicity increases , the dust attenuation could have an extra effect on @xmath0 , which we will discuss in the next section . .
the negative correlation between @xmath0 and metallicity is mainly caused by the positive correlation between halo mass and metallicity .
it is not the metals that directly controls @xmath0 .
the data points with @xmath112 are set to @xmath113 for plotting purposes . ]
interstellar dust can decrease the @xmath0 by absorbing the ionizing photons .
we evaluate the effect of dust attenuation on @xmath0 by comparing the rt result with and without the dust treatment , as described in [ sec : dust ] .
figure [ fig : dust ] shows the values of @xmath0 with and without the treatment of dust extinction , as a function of halo mass at @xmath108 .
the reduction rate of @xmath0 does not depend on the halo masses very much , and ranges from 0 to 20 per cent , with an average of 14 per cent .
if the dust - to - gas ratio is the same , the optical depth is roughly proportional to @xmath145 , where @xmath146 is typical dust size .
note that @xmath147 d\ell \propto a_{\rm d}^{-1}$ ] , where @xmath84 and @xmath148 are dust mass and dust density .
if we change the dust grain size distribution to a smaller size range of @xmath149 , the mean reduction rate increases to @xmath150 per cent at @xmath108 .
the weak dependence on halo mass is perhaps because most of the star - forming regions are enriched close to the solar metallicity , irrespective of the host halo masses .
our reduction rate is somewhat smaller than that reported by y09 , which is probably owing to the difference in the volume occupied by the metal rich gas . in y09 ,
the model galaxy was an isolated system , and the ism was globally mixed by the shock from supernova explosion , because there were no further infall of pristine gas from intergalactic space .
on the other hand , in the present work , pristine gas can accrete onto the haloes from intergalactic space , which reduces the volume fraction occupied by the metal rich gas around star - forming regions .
however , the reduction rate of @xmath0 reported in y09 are still in the 1-@xmath111 range of the present work . with and without dust extinction in the fiducial n144l10 uvb model at @xmath108 .
open black circles show @xmath0 without dust extinction , and the filled red circles show @xmath0 with dust extinction .
magenta crosses show the case when the dust grain size range is changed to @xmath151 m .
the triangles indicate the mean values in each mass bin with 1-@xmath111 error bars .
the data points with @xmath112 are set to @xmath113 for plotting purposes . _ lower panel _
: mean reduction rate of @xmath0 for each mass bin when including dust extinction .
different colors show different redshifts ( red : @xmath108 , blue : @xmath109 and green : @xmath110 ) .
the dashed magenta result is when the dust grain size range is changed to @xmath151 m .
the error bars are 1-@xmath111 . ]
@xcite and @xcite reported that the effect of dust attenuation on @xmath0 is very small .
in particular , @xcite argued that in their simulations , only the ionizing photons from stars in the outer disk can escape from the halo , and the escaping photons pass through only the low density regions with low metallicity and low dust content .
this may lead to smaller variations of @xmath0 in their simulations compared to other simulations that do not fully resolve the disk structure .
@xcite achieved a spatial resolution of @xmath152kpc in their sph simulations using a zoom - resimulation technique , but it is not clear if they had a sufficient resolution to resolve the disk structure of galaxies .
also , our sph simulations have lower resolution than those of @xcite , therefore we would expect a larger variation in @xmath0 compared to theirs , and our results are consistent with this expectation . in our sph simulations
, the escaping photons may also pass through moderately high density regions with higher metal and dust content , with stronger effects of dust attenuation .
furthermore , our dust model is different from those of @xcite and @xcite .
they adopted dust extinction curves of large and small magellanic clouds , whereas we use the dust size distribution of our galaxy derived by @xcite for a silicate - type dust .
despite the fact that the effective absorption cross section of our dust model is smaller than in their models , the scatter of @xmath0 in our simulations may be larger than theirs owing to the differences in both the resolution and dust treatment . in the future
, we plan to improve our dust model by including the treatments of formation and destruction of dust particles . as a result of our rt calculation presented in the earlier sections , we are able to estimate the total number of ionizing photons that escape from all the star - forming galaxies in the entire simulation box .
our calculation includes haloes with @xmath153 at @xmath108 , and those with @xmath154 at @xmath110 .
figure [ fig : reion ] compares the comoving emission rate density of ionizing photons @xmath155 ( i.e. , the number of ionizing photons emitted per unit time and per unit volume ) in our n144l10 fiducial uvb run to the required @xmath155 to reionize the universe @xcite , which is shown by the black solid curves for the clumping factors of @xmath156 and 30 .
the blue solid circles are for the intrinsically radiated photons from all star - forming galaxies in our simulation , and the red circles are for the escaped photons after the rt calculation . the clumping factor of igm at @xmath157 is still very uncertain , and the results from numerical simulations vary depending on the resolution and the treatment of physical processes such as star formation and radiation transfer . earlier ,
@xcite suggested @xmath158 at @xmath159 , however @xcite reported @xmath160 using higher resolution simulations with a rt treatment .
more recently , @xcite reported @xmath161 using a cosmological sph simulation with an optically thin approximation .
as a function of redshift .
blue filled circles are for the intrinsically radiated photons from all simulated galaxies , and the red filled circles are for the escaped ionizing photons after the rt calculation .
black solid lines show the required @xmath155 to reionize the universe , derived by @xcite for various clumping factors of igm .
black dotted line shows the contribution from qsos estimated by @xcite .
the filled squares are @xmath155 derived from the ly-@xmath24 opacity data of igm by @xcite . ]
our results show that @xmath155 of escaped ionizing photons after the rt calculation is greater than the required @xmath155 to ionize the universe at @xmath110 if @xmath160 , but below the required value if @xmath158 .
therefore our fiducial simulation suggests that the star - forming galaxies can ionize the igm as long as @xmath162 .
our results on @xmath155 is higher than those derived by @xcite , which was derived by using the results of cosmological sph simulations and observational data of ly@xmath24 forest .
although the error bars of their data points look very small , it represents only the dispersion of ly@xmath24 opacity data .
there are still significant uncertainties in the spectral shape and the mean free path of ionizing photons in @xcite . in their calculation , they use the distance between lyman limit systems as a mean free path of ionizing photons .
however many low density hi gas clouds can decrease the mean free path , and hence increase @xmath155 .
together with the uncertainties in our simulation such as the resolution and the details of star formation and feedback models , the differences between our results and that of @xcite can be accounted for .
we also study the fractional contribution to the ionizing photons by the haloes with different masses , as shown in figure [ fig : nifrac ] .
for the intrinsically radiated photons ( blue lines ) , there is no clear trend with the halo mass , and all the haloes contribute roughly equal number of ionizing photons .
however , the figure shows that most of the escaped ionizing photons ( red histograms ) come from the lower mass haloes ( @xmath163 ) .
this is because in our simulations , higher mass haloes have lower @xmath0 .
earlier in section [ sec : dependence ] , we discussed the large variation of @xmath0 derived by @xcite . based on a clustering analysis , @xcite reported that the sample in @xcite are hosted by haloes with @xmath123 .
when compared with our results in figure [ fig : nifrac ] , it suggests that the sample of @xcite might not be tracing the bulk of ionizing sources at @xmath108 , and only sampling the massive end of the distribution .
in addition , the detected sample in @xcite is brighter than @xmath164 mag in @xmath165 band . if we use the equation in @xcite and the relation between halo mass and sfr ( @xmath166 ) , the limiting magnitude of iwata s sample corresponds to @xmath167 ( sfr @xmath168yr@xmath64 ) . therefore they are also tracing only the massive end of the distribution , although one order of magnitude deeper in halo mass than @xcite .
observations of fainter sources are needed to capture the entire ionizing radiation from lower mass haloes .
in the sph simulations used for this work , haloes with @xmath169 are not resolved well .
for example , in the case of the fiducial n144l10 run , the halo with @xmath170 consists of 100 dark matter particles . in the future
, we plan to study @xmath0 of even lower mass haloes using higher resolution simulations .
if these low - mass haloes are the primary sources of ionizing photons at @xmath157 , the hii bubbles during the reionization epoch will be produced by numerous low mass haloes .
in addition , the large dispersion in @xmath0 that we found in this work suggests that the sizes of the hii bubbles may have a large variety in the early stages of reionization . in the present work
, we find that the value of @xmath0 depends strongly on the distribution of star particles with respect to the high density gas .
however , the clumpiness of ism around the star - forming regions could strongly depend on the resolution limit of simulations , therefore it is important to evaluate the resolution effects on @xmath0 .
in particular , the cosmological sph simulations have a difficulty in resolving the clumpy structure of ism when the number of sph particles is very small , and the values of @xmath0 for low - mass haloes may be more strongly affected by the limited resolution . when the clumpiness of ism is very high
, it may have both positive and negative effect on @xmath0 : the positive effect is that the ionizing photons may be able to escape through the void regions more easily , however those photons soon may be absorbed by the nearby high - density neutral clumps , which would be a negative effect .
to study the resolution effect , figure [ fig : resolution ] compares @xmath0 in the three runs with different resolution , as described in table [ table : sim ] . the lowest resolution run ( n400l100 ) with a larger box size shows
overall higher @xmath0 than the other runs . the highest resolution run ( n216l10 )
gives similar values of @xmath0 to the fiducial resolution run ( n144l10 ) , but somewhat lower @xmath0 for the lower mass haloes with @xmath171 .
given the this result , it is quite possible that our @xmath0 results have not converged yet , and we might still be overestimating @xmath0 for the low - mass haloes in the present work .
nevertheless , it is true in all the runs that @xmath0 has a large scatter for the low - mass haloes , and that @xmath0 decreases on average with increasing halo masses . in figure
[ fig : resolution ] , for the n216l10 run , we had to confine the sample to the low - mass haloes with @xmath172 , because of the heavy computational load . in the n216l10 run ,
we need to set up a large grid of @xmath173 for the high - mass halos with @xmath174 , and these grids take too long to process with rt when we want to process a large sample .
we will tackle the systematic study of haloes with higher resolution simulations using the next generation of supercomputers .
for the low - mass haloes , the @xmath0 of n216l10 run is smaller by a factor of @xmath175 than in the n144l10 run .
if we assume that @xmath0 of all haloes becomes one half , the resulting @xmath155 also becomes a half .
then the red circles in figure [ fig : reion ] would decrease by 0.3dex for the n216l10 run , and the threshold clumpiness factor for igm reionization changes to @xmath176 at @xmath177 . using three runs with different resolution , as described in table [ table :
the lower right panel compares the mean @xmath0 in each mass bins for the three runs with 1-@xmath111 error bars .
the points with @xmath112 are set to @xmath113 for plotting purposes . ]
we have performed three - dimensional radiation transfer calculations of stellar radiation for a large number of high-@xmath12 star - forming galaxies in cosmological sph simulations to explore the escape fraction of ionizing photons .
our major findings are as follows : * the value of @xmath0 decreases steeply with increasing halo mass , irrespective of numerical resolution .
+ * there is a large dispersion in @xmath0 for low - mass haloes with @xmath178 . +
* the values of @xmath0 do not vary much with redshift and different uvb models . + * the average reduction rate of @xmath0 owing to the dust attenuation effect is @xmath179 with a large dispersion . + * the results of our fiducial n144l10 run suggests that the star - forming galaxies can ionize the igm at @xmath1 , if the clumping factor is @xmath180 ( 10 ) at @xmath108 ( 6 ) .
if we use the results of the n216l10 run , we roughly estimate that the above threshold values would change to @xmath181 ( 3 ) at @xmath108 ( 6 ) .
our results suggest that the star - forming galaxies become the main contributor of igm ionization at @xmath182 . +
* the low mass haloes with @xmath183 are the main ionizing sources of igm in our simulations owing to their high @xmath0 .
the fraction of escaped ionizing photons coming from the haloes with @xmath184 at @xmath1 is 70 per cent for the fiducial n144l10 run . as we summarised in section
[ sec : intro ] , the current results on the escape fraction of ionizing photons are confusing , as different results are obtained from different simulations .
for example , @xcite argued that @xmath0 increases with increasing halo mass in the range of @xmath185 , and their values of @xmath0 were mostly less than a few per cent , much smaller than the other published work .
the trend found in our simulations ( decreasing @xmath0 with increasing halo mass ) is similar to that found by @xcite , but we find lower @xmath0 values despite of the fact that our simulations have lower resolution than their zoom - resimulations .
therefore the differences in @xmath0 between our work and @xcite can not be explained simply by the resolution effect .
we also note that @xcite obtained much higher values of @xmath0 ( @xmath710.4 ) than @xcite did , using the same amr method , but for a different halo mass range . considering these facts , the differences that we see now in the results of @xmath0 may have to do more with the different treatment of radiation transfer and the uv background radiation , rather than the resolution or numerical technique .
however , more detailed comparisons are needed to make more definite statements .
one of our main points is that the variation in @xmath0 is caused by the different geometry of ism distribution in the halo .
recently @xcite suggested that supernovae feedback and star formation efficiency can determine the geometry of a disk galaxy ( see also * ? ? ? * ) .
therefore the uncertainties in the treatment of star formation , feedback , and radiation transfer are all important for the calculations of @xmath0 , and we need to continue to improve these models through comparisons with future observations of high-@xmath12 galaxies . if we allow ourselves to speculate even further and combine all the current results mentioned above , it is possible that @xmath0 has a peak at @xmath186 as a function of halo mass at @xmath1 .
but this is highly speculative and by no means based on any definite physical arguments .
the strength of our current work is the large sample size of galaxies that we processed with rt .
our simulations also adopt a new galactic wind model which produces more favorable results on the cosmic star formation rate and the igm statistics such as civ mass density @xcite .
although it is difficult to address the exact effect of our wind model on @xmath0 unless we process simulations with different wind models with rt calculation , our test calculations showed that the effect is not so strong , and our main conclusions of this paper should remain unchanged even if we modify the wind model slightly .
we consider that the most significant results in the current work are the large scatter of @xmath0 for the low mass haloes , and its decline with the increasing halo mass .
the earlier works by other authors did not discuss the scatter among different haloes with a wide range of mass owing to their small sample size . at @xmath187 ,
even lower mass haloes with @xmath169 may become the main sources of igm ionization .
however in such low - mass systems , the uv radiation of massive stars may influence the gas dynamics significantly @xcite .
simulations with higher resolution than presented in this paper are needed to follow the star formation in such low - mass systems , and we need to solve the hydrodynamics and radiation transfer simultaneously to examine the effect of radiative feedback . in the future ,
we plan to couple the rt with hydrodynamics and study the effects of radiative feedback by young stars and agns .
hy thanks to m. umemura for valuable discussions and comments .
we are grateful to volker springel for providing us with the original version of gadget-3 , on which @xcite simulations are based .
this work is supported in part by the nsf grant ast-0807491 , national aeronautics and space administration under grant / cooperative agreement no .
nnx08ae57a issued by the nevada nasa epscor program , and the president s infrastructure award from unlv .
this research is also supported by the nsf through the teragrid resources provided by the texas advanced computing center ( tacc ) .
numerical simulations and analyses have been performed on the unlv cosmology cluster , the _ first _ simulator and _ t2k - tsukuba _ at center for computational sciences , university of tsukuba .
kn is grateful to the hospitality of the ipmu , university of tokyo , where part of this work was done . | combing the three - dimensional radiative transfer ( rt ) calculation and cosmological sph simulations , we study the escape fraction of ionizing photons ( @xmath0 ) of high - redshift galaxies at @xmath1 .
our simulations cover the halo mass range of @xmath2 .
we postprocess several hundred simulated galaxies with the authentic radiative transfer ( art ) code to study the halo mass dependence of @xmath0 . in this paper , we restrict ourselves to the transfer of stellar radiation from local stellar population in each dark matter halo .
we find that the average @xmath0 steeply decreases as the halo mass increases , with a large scatter for the lower mass haloes .
the low mass haloes with @xmath3 have large values of @xmath0 ( with an average of @xmath4 ) , whereas the massive haloes with @xmath5 show small values of @xmath0 ( with an average of @xmath6 ) .
this is because in our simulations , the massive haloes show more clumpy structure in gas distribution , and star - forming regions are embedded inside these clumps , making it more difficult for the ionizing photons to escape . on the other hand , in low mass haloes
, there are often conical regions of highly ionized gas due to the shifted location of young star clusters from the center of dark matter halo , which allows the ionizing photons to escape more easily than in the high - mass haloes . by counting the number of escaped ionizing photons ,
we show that the star - forming galaxies can ionize the intergalactic medium at @xmath1 .
the main contributor to the ionizing photons is the haloes with @xmath7 owing to their high @xmath0 .
the large dispersion in @xmath0 suggests that there may be various sizes of hii bubbles around the haloes even with the same mass in the early stages of reionization .
we also examine the effect of uv background radiation field on @xmath0 using simple , four different treatment of uv background .
[ firstpage ] radiative transfer ism : dust , extinction galaxies : evolution galaxies : formation galaxies : high - redshift methods : numerical |
[ [ sec : la ] ] finding the irreducible characters is a fundamental problem in representation theory . as a prototype of this problem , consider a complex semisimple lie algebra @xmath2 .
the problem is solved in two steps , following the historical development : 1 .
the category of finite dimensional @xmath2-modules is semisimple and the corresponding irreducible @xmath2-characters are given by weyl s character formula .
a general solution to the irreducible character problem in the bgg category @xmath0 was given much later by the kazhdan - lusztig ( kl ) polynomials ( theorems of beilinson - bernstein and brylinski - kashiwara ) @xcite .
[ [ sec : lsa ] ] the study of lie superalgebras and their representations was largely motivated by the notion of supersymmetry in physics .
a killing - cartan type classification of finite dimensional complex simple lie superalgebras was achieved by kac @xcite in 1977 .
the most important subclass of the simple lie superalgebras ( called basic classical ) , including two infinite series of types @xmath3 and @xmath4 , bears a close resemblance to the usual simple lie algebras , so we can make sense of root systems , dynkin diagrams , triangular decomposition , cartan and borel subalgebras , ( parabolic ) category @xmath0 , and so on .
however , the representation theory of lie superalgebras @xmath5 has encountered several substantial difficulties , as made clear by numerous works over the last three decades ( cf .
@xcite as a sample of earlier literature , and some more recent references which can be found in the next paragraph ) : 1 . there exist non - conjugate borel subalgebras for a given lie superalgebra @xmath5 .
the category @xmath6 of finite dimensional @xmath5-modules is in general not semisimple .
a uniform weyl type finite dimensional character formula does not exist .
3 . one has a notion of a weyl group associated to the even subalgebra of @xmath5 ; however , the linkage in the category @xmath0 ( or in @xmath6 ) of @xmath5-modules is not solely controlled by the weyl group .
4 . a block in the category @xmath0 ( or in @xmath6 )
may contain infinitely many simple objects .
the conventional wisdom of solving the irreducible character problem for lie superalgebras has been to follow closely the two steps for lie algebras in [ sec : la ] . as the problem is already very difficult in the category @xmath6 ,
there has been little attempt in understanding the category @xmath0 . for type @xmath3
lie superalgebra @xmath7 , there have been several different general approaches over the years .
serganova @xcite in 1996 developed a mixed geometric and algebraic approach to solving the irreducible character problem in the category @xmath6 .
brundan in 2003 @xcite developed a new elegant purely algebraic solution to the same problem in @xmath6 using lusztig - kashiwara canonical basis .
developing the idea of super duality @xcite ( which generalizes @xcite ) which connects the categories @xmath0 for lie superalgebras and lie algebras of type @xmath3 for the first time , two of the authors @xcite very recently established the super duality conjecture therein .
in particular they provided a complete solution to the irreducible character problem for a fairly general parabolic category @xmath0 ( including @xmath6 as a very special case ) in terms of kl polynomials of type @xmath3 .
independently , brundan and stroppel @xcite proved the super duality conjecture in @xcite , offering yet another solution of the irreducible character problem in @xmath6 .
[ [ section ] ] the goal of this paper is to formulate and establish a super duality which connects parabolic category @xmath0 for lie superalgebra of type @xmath4 with parabolic category @xmath0 for classical lie algebras of types @xmath1 , vastly generalizing the type @xmath3 case of @xcite . in particular
, it provides a complete solution of the irreducible character problem for @xmath4 in some suitable parabolic category @xmath0 , which includes all finite dimensional irreducibles , in terms of parabolic kl polynomials of @xmath1 types ( cf .
deodhar @xcite ) .
[ [ section-1 ] ] before launching on a detailed explanation of our main ideas below , it is helpful to keep in mind the analogy that the ring of symmetric functions ( or its super counterpart ) in infinitely many variables carries more symmetries than in finitely many variables , and a truncation process can easily recover finitely many variables .
the super duality can be morally thought as a categorification of the standard involution @xmath8 on the ring of symmetric functions and it only becomes manifest when the underlying lie ( super)algebras pass to infinite rank . then truncation functors can be used to recover the finite rank cases which we are originally interested in . even though the finite dimensional lie superalgebras of type @xmath4 depend on two integers @xmath9 and @xmath10 , our view is to fix one and let the other , say @xmath10 , vary , and so let us denote an @xmath4 lie superalgebra by @xmath11 . by choosing appropriately a borel and a levi subalgebra of @xmath11
, we formulate a suitable parabolic category @xmath12 of @xmath11-modules .
it turns out that there are four natural choices one can make here which correspond to the four dynkin diagrams @xmath13 in [ dynkin ] ( the type @xmath14 case has been treated in @xcite ) .
there is a natural sequence of inclusions of lie superalgebras : @xmath15 let @xmath16 .
a suitable category @xmath17 of @xmath5-modules can be identified with the inverse limit @xmath18 . on the other hand , we introduce truncation functors @xmath19 for @xmath20 , as analogues of the truncation functors studied in algebraic group setting ( cf .
donkin @xcite ) .
these truncation functors send parabolic verma modules to parabolic vermas or zero and irreducibles to irreducibles or zero . in particular , this allows us to derive the irreducible characters in @xmath12 once we know those in @xmath21 .
corresponding to each of the above choices of @xmath11 and @xmath22 , we have the lie algebra counterparts @xmath23 and parabolic categories @xmath24 for positive integers @xmath10 .
moreover , we have natural inclusions of lie algebras @xmath25 , for all @xmath10 , which allow us to define the lie algebra @xmath26 and the parabolic category @xmath27 of @xmath2-modules .
similarly , the category @xmath27 can be identified with the inverse limit @xmath28 . in the main body of the paper ,
we actually replace @xmath29 et cetera by their ( trivial ) central extensions .
the reason is that the truncation functors depend implicitly on a stabilization scalar , which is interpreted conceptually as a level of representations with respect to the central extensions . to establish a connection between @xmath17 and @xmath27
, we introduce another infinite rank lie superalgebra @xmath30 and its parabolic category @xmath31 .
the lie superalgebra @xmath30 contains @xmath2 and @xmath5 as natural subalgebras ( though not as levi subalgebras ) , and this enables us to introduce two natural functors @xmath32 and @xmath33 . using the technique of odd reflections among others , we establish in section [ sec : character ] a key property that @xmath34 and @xmath35 respect the parabolic verma and irreducible modules , respectively .
this result is already sufficient to provide a complete solution of irreducible @xmath4-characters in the category @xmath12 in the first half of the paper ( by the end of section [ sec : t ] ) .
we remark that the idea of introducing an auxiliary lie superalgebra @xmath30 and category @xmath36 has been used in the type @xmath3 superalgebra setting @xcite .
recall that for the usual category @xmath0 of lie algebras , the kl polynomials were interpreted by vogan @xcite in terms of kostant @xmath37-homology groups .
the @xmath38-homology groups make perfect sense for lie superalgebras , and we may take this interpretation as the definition for the otherwise undefined kl polynomials in category @xmath0 for lie superalgebras ( cf . section [ sec : lsa ] ( 3 ) ) , as serganova @xcite did in the category @xmath6 of @xmath39-modules . in section
[ sec : homology ] , we show that the functors @xmath34 and @xmath35 match the corresponding @xmath37-homology groups and hence the corresponding kl polynomials ( compare @xcite ) . actually the computation in @xcite of the @xmath37-homology groups with coefficients in the lie superalgebra oscillator modules via howe duality was the first direct supporting evidence for the super duality for @xmath4 as formulated in this paper .
section [ sec : category ] of the paper is devoted to proving that both @xmath34 and @xmath35 are indeed category equivalences . as a consequence
, we have established that the categories @xmath40 and @xmath21 are equivalent , which is called super duality .
a technical difference here from @xcite is that we need to deal with the fact that parabolic verma modules in @xmath21 may not have finite composition series .
an immediate corollary of the super duality is that any known bgg resolution in the category @xmath40 gives rise to a bgg type resolution in the category @xmath21 , and vice versa .
[ [ section-2 ] ] the finite dimensional irreducible @xmath4-modules are of highest weight , and they are classified in terms of the dynkin labels by kac @xcite .
we note that the finite dimensional irreducible modules of non - integral highest weights are typical and so their characters are known ( * ? ? ? * theorem 1 ) .
it turns out that a more natural labeling of the remaining finite dimensional irreducible @xmath4-modules ( of integral highest weights ) is given in terms of young diagrams just as for classical lie algebras ( see e.g. @xcite for such a formulation and a new proof using odd reflections ) .
as borel subalgebras are not conjugate to each other , it becomes a nontrivial problem to find the extremal weights , i.e. highest weights with respect to different borel , of a given finite dimensional irreducible @xmath4-module .
we provide an elegant and simple answer in terms of a combinatorial notion which we call _ block frobenius coordinates _ associated to young diagrams .
we observe that our solution of the irreducible character problem in @xmath22 includes solutions to all finite dimensional irreducible @xmath4-characters .
the category @xmath6 for a general @xmath4 ( with the exception of @xmath41 ) is not a highest weight category and does not admit an abstract kl theory in the sense of cline , parshall and scott @xcite , as indicated in the case of @xmath42 ( * ? ? ?
* section 2 ) .
for a completely independent and different approach to the finite dimensional irreducible @xmath4-characters in the category @xmath6 , see gruson and serganova @xcite .
the finite dimensional irreducible characters of @xmath41 were obtained in @xcite .
the finite dimensional irreducible characters of @xmath43 were also computed in @xcite .
[ [ section-3 ] ] in hindsight , here is how our super duality approach overcomes the difficulties as listed in [ sec : lsa ] . 1 .
the existence of non - conjugate borel subalgebras for a lie superalgebra is essential for establishing the properties of the functors @xmath34 and @xmath35 .
choices of suitable borel subalgebras are crucial for a formulation of the compatible sequence of categories @xmath22 for @xmath44 .
2 . the category @xmath6 of finite dimensional @xmath11-modules does not play any special role in our approach .
even the natural " @xmath45-modules @xmath46 do not correspond well with each other under truncation functors , as they are natural with respect to the
wrong " borel .
3 . in the @xmath47 limit
, the linkage in the category @xmath21 of @xmath5-modules is completely controlled by the weyl group of the corresponding lie algebra @xmath2 ( which contains the even subalgebra of @xmath5 as a subalgebra ) .
4 . in the @xmath47 limit
, it is no surprise for a block to contain infinitely many simple objects . in the extreme cases described in section [ sec : m=0 ] ,
we indeed obtain an equivalence of module categories between two classical ( non - super ! ) lie algebras of types @xmath48 and @xmath49 of infinite rank at opposite levels .
if one is willing to regard @xmath50 as classical ( recall that the finite dimensional @xmath51-module category is semisimple ) , there is another equivalence of categories which relates @xmath50 to the infinite rank lie algebra of type @xmath52 . in this sense ,
our super duality has a flavor of the langlands duality .
the super duality approach here can be further adapted to the setting of kac - moody superalgebras ( including affine superalgebras ) and this will shed new light on the irreducible character problem for these superalgebras . the details will appear elsewhere .
it is well known that the proof of kl conjectures involves deep geometric machinery and results on @xmath49-modules of flag manifolds .
the formulation of super duality suggests potential direct connections on the ( super ) geometric level behind the categories @xmath40 and @xmath21 , which will be very important to develop .
[ [ section-4 ] ] the paper is organized as follows . in section [ sec : superalgebras ] the lie superalgebras @xmath2 , @xmath5 and @xmath53 are defined , with their respective module categories @xmath27 , @xmath17 and @xmath31 introduced in section [ sec : o ] . in section [ sec : character ] , we provide a complete solution of the irreducible @xmath4 character problem in category @xmath22 for all @xmath10 , including all finite dimensional irreducible @xmath4-characters , in terms of the kl polynomials of classical type .
we establish in section [ sec : category ] equivalence of the categories @xmath40 and @xmath21 .
section [ finite : dim : repn ] offers a diagrammatic description of the extremal weights of the finite dimensional irreducible @xmath4-modules . throughout the paper the symbols @xmath54 , @xmath55 , and @xmath56
stand for the sets of all , positive and non - negative integers , respectively .
all vector spaces , algebras , tensor products , et cetera , are over the field of complex numbers @xmath57
. * acknowledgment . * the first author is partially supported by an nsc - grant and an academia sinica investigator grant , and he thanks ncts / tpe and the department of mathematics of university of virginia for hospitality and support .
the second author is partially supported by an nsc - grant and thanks ncts / south .
the third author is partially supported by nsf and nsa grants , and he thanks the institute of mathematics of academia sinica in taiwan for hospitality and support .
the results of the paper were announced by the first author in the ams meeting at raleigh in april 2009 , and they were presented by the third author in conferences at ottawa , canada and durham , uk in july 2009 .
in this section , we introduce infinite rank lie ( super)algebras @xmath58 , @xmath59 and @xmath60 associated to the @xmath61 dynkin diagrams in below , where @xmath62 denotes one of the four types @xmath13 .
let @xmath64 .
consider the free abelian group with basis @xmath65 , with a symmetric bilinear form @xmath66 given by @xmath67 we set @xmath68 for @xmath69 , we denote by @xmath70 the contragredient lie ( super)algebras ( ( * ? ? ?
* section 2.5 ) ) whose dynkin diagrams ( 23,0)@xmath71(-20,8)@xmath70 together with certain distinguished sets of simple roots @xmath72 are listed as follows : -3 cm ( 24,3 ) ( 8,2)(0,0)[c]@xmath73 ( 10.4,2)(0,0)[c]@xmath73 ( 14.85,2)(0,0)[c]@xmath73 ( 17.25,2)(0,0)[c]@xmath73 ( 19.4,2)(0,0)[c]@xmath73 ( 5.6,2)(0,0)[c]@xmath73 ( 8.4,2)(1,0)1.55 ( 10.82,2)(1,0)0.8 ( 13.2,2)(1,0)1.2 ( 15.28,2)(1,0)1.45 ( 17.7,2)(1,0)1.25 ( 6,2)(1,0)1.4 ( 12.5,1.95)(0,0)[c]@xmath74 ( -.5,2)(0,0)[c]@xmath75 : ( 5.5,1)(0,0)[c]@xmath76 ( 8,1)(0,0)[c]@xmath77 ( 17.2,1)(0,0)[c]@xmath78 ( 19.3,1)(0,0)[c]@xmath79 -3 cm ( 24,2 ) ( 8,2)(0,0)[c]@xmath73 ( 10.4,2)(0,0)[c]@xmath73 ( 14.85,2)(0,0)[c]@xmath73 ( 17.25,2)(0,0)[c]@xmath73 ( 19.4,2)(0,0)[c]@xmath73 ( 5.6,2)(0,0)[c]@xmath73 ( 8.4,2)(1,0)1.55 ( 10.82,2)(1,0)0.8 ( 13.2,2)(1,0)1.2 ( 15.28,2)(1,0)1.45 ( 17.7,2)(1,0)1.25 ( 6,1.8)@xmath80 ( 12.5,1.95)(0,0)[c]@xmath74 ( -.5,2)(0,0)[c]@xmath81 : ( 5.5,1)(0,0)[c]@xmath82 ( 8,1)(0,0)[c]@xmath76 ( 17.2,1)(0,0)[c]@xmath78 ( 19.3,1)(0,0)[c]@xmath79 -3 cm ( 24,2 ) ( 5.6,2 ) ( 8,2)(0,0)[c]@xmath73 ( 10.4,2)(0,0)[c]@xmath73 ( 14.85,2)(0,0)[c]@xmath73 ( 17.25,2)(0,0)[c]@xmath73 ( 19.4,2)(0,0)[c]@xmath73 ( 8.35,2)(1,0)1.5 ( 10.82,2)(1,0)0.8 ( 13.2,2)(1,0)1.2 ( 15.28,2)(1,0)1.45 ( 17.7,2)(1,0)1.25 ( 6.8,2)(0,0)[c]@xmath80 ( 12.5,1.95)(0,0)[c]@xmath74 ( -.2,2)(0,0)[c]@xmath83 : ( 5.5,1)(0,0)[c]@xmath82 ( 8,1)(0,0)[c]@xmath76 ( 17.2,1)(0,0)[c]@xmath78 ( 19.3,1)(0,0)[c]@xmath79 -3 cm ( 24,2 ) ( 5.7,2)(0,0)[c]@xmath73 ( 8,2)(0,0)[c]@xmath73 ( 10.4,2)(0,0)[c]@xmath73 ( 14.85,2)(0,0)[c]@xmath73 ( 17.25,2)(0,0)[c]@xmath73 ( 19.4,2)(0,0)[c]@xmath73 ( 6.8,2)(0,0)[c]@xmath84 ( 8.4,2)(1,0)1.55 ( 10.82,2)(1,0)0.8 ( 13.2,2)(1,0)1.2 ( 15.28,2)(1,0)1.45 ( 17.7,2)(1,0)1.25 ( 12.5,1.95)(0,0)[c]@xmath74 ( -.5,2)(0,0)[c]@xmath85 : ( 5.5,1)(0,0)[c]@xmath86 ( 8,1)(0,0)[c]@xmath76 ( 17.2,1)(0,0)[c]@xmath78 ( 19.3,1)(0,0)[c]@xmath79 -3 cm ( 24,3.5 ) ( 8,2)(0,0)[c]@xmath73 ( 10.4,2)(0,0)[c]@xmath73 ( 14.85,2)(0,0)[c]@xmath73 ( 17.25,2)(0,0)[c]@xmath73 ( 19.4,2)(0,0)[c]@xmath73 ( 6,3.8)(0,0)[c]@xmath73 ( 6,.3)(0,0)[c]@xmath73 ( 8.4,2)(1,0)1.55 ( 10.82,2)(1,0)0.8 ( 13.2,2)(1,0)1.2 ( 15.28,2)(1,0)1.45 ( 17.7,2)(1,0)1.25 ( 7.6,2.2)(-1,1)1.3 ( 7.6,1.8)(-1,-1)1.3 ( 12.5,1.95)(0,0)[c]@xmath74 ( -.5,2)(0,0)[c]@xmath87 : ( 3.3,0.3)(0,0)[c]@xmath88 ( 4.7,3.8)(0,0)[c]@xmath76 ( 8.2,1)(0,0)[c]@xmath77 ( 17.2,1)(0,0)[c]@xmath78 ( 19.3,1)(0,0)[c]@xmath79 according to ( * ? ? ? * proposition 2.5.6 ) these lie ( super)algebras are @xmath89 , @xmath90 , @xmath91 for @xmath92 and @xmath93 for @xmath94 , respectively
. we will use the same notation ( 23,0)@xmath71(-20,8)@xmath70 to denote the diagrams of all the degenerate cases for @xmath95 as well .
( see sections [ sec : realize ] and [ sec : m=0 ] below ) . we have used ( 15,5 ) to denote an odd non - isotropic simple root .
so @xmath90 is actually a lie superalgebra ( instead of lie algebra ) , but it is classical from the super duality viewpoint in this paper . for @xmath96 let ( 23,0)@xmath71(-20,8)@xmath97 , ( 23,0)@xmath98(-20,8)@xmath99 and ( 23,0)@xmath98(-20,8)@xmath100 denote the following dynkin diagrams , where @xmath101 denotes an odd isotropic simple root : -3 cm ( 24,3 ) ( 8,2)(0,0)[c]@xmath73 ( 10.4,2)(0,0)[c]@xmath73 ( 14.85,2)(0,0)[c]@xmath73 ( 17.25,2)(0,0)[c]@xmath73 ( 5.6,2)(0,0)[c]@xmath73 ( 8.4,2)(1,0)1.55 ( 10.82,2)(1,0)0.8 ( 13.2,2)(1,0)1.2 ( 15.28,2)(1,0)1.45 ( 6,2)(1,0)1.4 ( 12.5,1.95)(0,0)[c]@xmath74 ( 0,1.2)(1.6,1.2)@xmath97 ( 5.5,1)(0,0)[c]@xmath102 ( 8,1)(0,0)[c]@xmath103 ( 10.3,1)(0,0)[c]@xmath104 ( 15,1)(0,0)[c]@xmath105 ( 17.2,1)(0,0)[c]@xmath106 -3 cm ( 24,2 ) ( 8,2)(0,0)[c]@xmath73 ( 10.4,2)(0,0)[c]@xmath73 ( 14.85,2)(0,0)[c]@xmath73 ( 17.25,2)(0,0)[c]@xmath73 ( 5.6,2)(0,0)[c]@xmath101 ( 8.4,2)(1,0)1.55 ( 10.82,2)(1,0)0.8 ( 13.2,2)(1,0)1.2 ( 15.28,2)(1,0)1.45 ( 6,2)(1,0)1.4 ( 12.5,1.95)(0,0)[c]@xmath74 ( 0,1.2)(1.6,1.2)@xmath99 ( 5.5,1)(0,0)[c]@xmath107 ( 10.3,1)(0,0)[c]@xmath108 ( 8,1)(0,0)[c]@xmath109 ( 14.8,1)(0,0)[c]@xmath110 ( 17.5,1)(0,0)[c]@xmath111 -3 cm ( 24,2 ) ( 8,2)(0,0)[c]@xmath101 ( 10.4,2)(0,0)[c]@xmath101 ( 14.85,2)(0,0)[c]@xmath101 ( 17.25,2)(0,0)[c]@xmath101 ( 5.6,2)(0,0)[c]@xmath101 ( 8.4,2)(1,0)1.55 ( 10.82,2)(1,0)0.8 ( 13.2,2)(1,0)1.2 ( 15.28,2)(1,0)1.45 ( 6,2)(1,0)1.4 ( 12.5,1.95)(0,0)[c]@xmath74 ( 0,1.2)(1.6,1.2)@xmath100 ( 5.5,1)(0,0)[c]@xmath107 ( 8,1)(0,0)[c]@xmath112 ( 14.8,1)(0,0)[c]@xmath113 ( 17.2,1)(0,0)[c]@xmath114 ( 10.3,1)(0,0)[c]@xmath115 the lie superalgebras associated with these dynkin diagrams are @xmath116 , @xmath117 and @xmath118 , respectively . in the limit @xmath119 ,
the associated lie superalgebras are direct limits of these lie superalgebras , and we will simply drop @xmath120 to write ( 23,0)@xmath71(-20,8)@xmath121 = ( 23,0)@xmath71(-20,8)@xmath122 and so on .
any of the _ head _ diagrams ( 23,0)@xmath123(-20,8)@xmath70 may be connected with the _ tail _ diagrams ( 23,0)@xmath71(-20,8)@xmath97 , ( 23,0)@xmath98(-20,8)@xmath99 and ( 23,0)@xmath98(-20,8)@xmath100 to produce the following dynkin diagrams ( @xmath124 ) : @xmath125{{\ovalbox(1.6,1.2){$\mf{k}^\xx$ } } } } \put(5.8,0.5){\line(1,0){1.85 } } \put(8.5,0.5){\makebox(0,0)[c]{{\ovalbox(1.6,1.2){$\mf{t}_n$ } } } } \put(15,0.5){\makebox(0,0)[c]{{\ovalbox(1.6,1.2){$\mf{k}^\xx$ } } } } \put(15.8,0.5){\line(1,0){1.85 } } \put(18.5,0.5){\makebox(0,0)[c]{{\ovalbox(1.6,1.2){$\ov{\mf{t}}_n$ } } } } \put(25,0.5){\makebox(0,0)[c]{{\ovalbox(1.6,1.2){$\mf{k}^\xx$ } } } } \put(25.8,0.5){\line(1,0){1.85 } } \put(28.5,0.5){\makebox(0,0)[c]{{\ovalbox(1.6,1.2){$\wt{\mf{t}}_n$ } } } } \end{picture } \ ] ] we will denote the sets of simple roots of the above diagrams accordingly by @xmath126 , @xmath127 and @xmath128 . for @xmath129 , we also denote the sets of positive roots by @xmath130 , @xmath131 and @xmath132 , and the sets of roots by @xmath133 , @xmath134 and @xmath135 , respectively .
let us denote the @xmath61 dynkin diagrams of ( [ dynkin : combined ] ) at @xmath129 by ( 23,0)@xmath98(-20,8)@xmath58 , ( 23,0)@xmath98(-20,8)@xmath59 and ( 23,0)@xmath98(-20,8)@xmath136 .
we provide a realization for the corresponding lie superalgebras .
for @xmath64 consider the following totally ordered set @xmath137 @xmath138 for @xmath64 define the following subsets of @xmath139 : @xmath140 for a subset @xmath141 of @xmath137 , define @xmath142 for a homogeneous element @xmath143 in a super vector space @xmath144 we denote by @xmath145 its @xmath146-degree . for @xmath64 consider the infinite dimensional super space @xmath147 over @xmath57 with ordered basis @xmath148 .
we declare @xmath149 , if @xmath150 , and @xmath151 , if @xmath152 .
the parity of the vector @xmath153 is to be specified .
with respect to this basis a linear map on @xmath154 may be identified with a complex matrix @xmath155 .
the lie superalgebra @xmath156 is the lie subalgebra of linear transformations on @xmath154 consisting of @xmath157 with @xmath158 for all but finitely many @xmath159 s .
denote by @xmath160 the elementary matrix with @xmath161 at the @xmath162th row and @xmath163th column and zero elsewhere .
the vector spaces @xmath164 and @xmath165 are defined to be subspaces of @xmath154 with ordered basis @xmath166 indexed by @xmath167 and @xmath168 , respectively .
the corresponding subspaces of @xmath164 , @xmath165 and @xmath154 with basis vectors @xmath169 , with @xmath170 indexed by @xmath171 , @xmath172 and @xmath173 , respectively , are denoted by @xmath174 , @xmath175 and @xmath176 , respectively .
this gives rise to lie superalgebras @xmath177 , @xmath178 , @xmath179 , @xmath180 and @xmath181 .
let @xmath182 be one of the spaces @xmath183 or @xmath184 .
the standard cartan subalgebra of @xmath185 is spanned by the basis @xmath186 , with corresponding dual basis @xmath187 , where @xmath162 runs over the index sets @xmath137 , @xmath173 , @xmath188 , @xmath189 , @xmath168 , @xmath172 , respectively . in this subsection
we set @xmath190 . for @xmath64
define a non - degenerate skew - supersymmetric bilinear form @xmath66 on @xmath154 by @xmath191 restricting the form to @xmath176 , @xmath164 , @xmath174 , @xmath165 and @xmath175 gives rise to non - degenerate skew - supersymmetric bilinear forms that will again be denoted by @xmath66 .
let @xmath182 be as before .
the lie superalgebra @xmath192 is the subalgebra of @xmath185 preserving the bilinear form @xmath66 .
the standard cartan subalgebra of @xmath192 is spanned by the basis @xmath193 , with corresponding dual basis @xmath194 .
we have the following realizations of the corresponding lie superalgebras for @xmath195 and @xmath196 : . [ cols="^,^,^,^",options="header " , ] the sets @xmath197 , @xmath198 and @xmath199 give rise to the following sets of positive roots : @xmath200 again , the subsets of simple roots can be read off from the corresponding diagrams in .
the dynkin diagrams of @xmath192 with a distinguished set of simple roots , for @xmath202 are listed in order as follows ( see also remark [ rem:000 ] below ) : -3 cm ( 24,3 ) ( 8,2)(0,0)[c]@xmath101 ( 10.4,2)(0,0)[c]@xmath101 ( 14.85,2)(0,0)[c]@xmath101 ( 17.25,2)(0,0)[c]@xmath101 ( 19.4,2)(0,0)[c]@xmath101 ( 5.6,2)(0,0)[c]@xmath73 ( 8.4,2)(1,0)1.55 ( 10.82,2)(1,0)0.8 ( 13.2,2)(1,0)1.2 ( 15.28,2)(1,0)1.45 ( 17.7,2)(1,0)1.25 ( 19.8,2)(1,0)1.5 ( 6,1.8)@xmath80 ( 12.5,1.95)(0,0)[c]@xmath74 ( 22.6,1.95)(0,0)[c]@xmath74 ( 8,1)(0,0)[c]@xmath112 ( 10.5,1)(0,0)[c]@xmath115 ( 14.8,1)(0,0)[c]@xmath203 ( 17.15,1)(0,0)[c]@xmath204 ( 19.8,1)(0,0)[c]@xmath205 ( 5.4,1)(0,0)[c]@xmath206 ( 0,1.2)(1.6,1.2)@xmath207 -3 cm ( 24,4 ) ( 8,2)(0,0)[c]@xmath101 ( 10.4,2)(0,0)[c]@xmath101 ( 14.85,2)(0,0)[c]@xmath101 ( 17.25,2)(0,0)[c]@xmath101 ( 19.4,2)(0,0)[c]@xmath101 ( 6,3.8)(0,0)[c]@xmath101 ( 6,.3)(0,0)[c]@xmath101 ( 8.4,2)(1,0)1.55 ( 10.82,2)(1,0)0.8 ( 13.2,2)(1,0)1.2 ( 15.28,2)(1,0)1.45 ( 17.7,2)(1,0)1.25 ( 19.8,2)(1,0)1.5 ( 7.5,2.2)(-1,1)1.3 ( 7.6,1.8)(-1,-1)1.25 ( 6,.8)(0,1)2.6 ( 12.5,1.95)(0,0)[c]@xmath74 ( 22.6,1.95)(0,0)[c]@xmath74 ( 8.1,1)(0,0)[c]@xmath115 ( 10.9,1)(0,0)[c]@xmath208 ( 14.8,1)(0,0)[c]@xmath203 ( 17.15,1)(0,0)[c]@xmath204 ( 19.8,1)(0,0)[c]@xmath205 ( 4.5,3.8)(0,0)[c]@xmath112 ( 3.5,0.3)(0,0)[c]@xmath209 ( 0,1.2)(1.6,1.2)@xmath210 -3 cm ( 24,3 ) ( 8,2)(0,0)[c]@xmath73 ( 10.4,2)(0,0)[c]@xmath73 ( 14.85,2)(0,0)[c]@xmath73 ( 17.25,2)(0,0)[c]@xmath73 ( 19.4,2)(0,0)[c]@xmath73 ( 5.6,2 ) ( 8.4,2)(1,0)1.55 ( 10.82,2)(1,0)0.8 ( 13.2,2)(1,0)1.2 ( 15.28,2)(1,0)1.45 ( 17.7,2)(1,0)1.25 ( 19.8,2)(1,0)1.5 ( 6,1.8)@xmath80 ( 12.5,1.95)(0,0)[c]@xmath74 ( 22.6,1.95)(0,0)[c]@xmath74 ( 8,1)(0,0)[c]@xmath211 ( 10.5,1)(0,0)[c]@xmath104 ( 14.8,1)(0,0)[c]@xmath106 ( 17.15,1)(0,0)[c]@xmath212 ( 19.8,1)(0,0)[c]@xmath213 ( 5.4,1)(0,0)[c]@xmath214 ( -1,1.2)(1.6,1.2)@xmath215 ( 1,1.5)= ( 2,1.2)(1.6,1.2)@xmath216 -3 cm ( 24,3 ) ( 8,2)(0,0)[c]@xmath73 ( 10.4,2)(0,0)[c]@xmath73 ( 14.85,2)(0,0)[c]@xmath73 ( 17.25,2)(0,0)[c]@xmath73 ( 19.4,2)(0,0)[c]@xmath73 ( 5.6,2)(0,0)[c]@xmath73 ( 8.4,2)(1,0)1.55 ( 10.82,2)(1,0)0.8 ( 13.2,2)(1,0)1.2 ( 15.28,2)(1,0)1.45 ( 17.7,2)(1,0)1.25 ( 19.8,2)(1,0)1.5 ( 6,1.8)@xmath84 ( 12.5,1.95)(0,0)[c]@xmath74 ( 22.6,1.95)(0,0)[c]@xmath74 ( 8,1)(0,0)[c]@xmath211 ( 10.5,1)(0,0)[c]@xmath104 ( 14.8,1)(0,0)[c]@xmath106 ( 17.15,1)(0,0)[c]@xmath212 ( 19.8,1)(0,0)[c]@xmath213 ( 5.4,1)(0,0)[c]@xmath217 ( -1,1.2)(1.6,1.2)@xmath218 ( 1,1.5)= ( 2,1.2)(1.6,1.2)@xmath219 -3 cm ( 24,3 ) ( 8,2)(0,0)[c]@xmath73 ( 10.4,2)(0,0)[c]@xmath73 ( 14.85,2)(0,0)[c]@xmath73 ( 17.25,2)(0,0)[c]@xmath73 ( 19.4,2)(0,0)[c]@xmath73 ( 5.6,2)(0,0)[c]@xmath73 ( 8.4,2)(1,0)1.55 ( 10.82,2)(1,0)0.8 ( 13.2,2)(1,0)1.2 ( 15.28,2)(1,0)1.45 ( 17.7,2)(1,0)1.25 ( 19.8,2)(1,0)1.5 ( 6,1.8)@xmath80 ( 12.5,1.95)(0,0)[c]@xmath74 ( 22.6,1.95)(0,0)[c]@xmath74 ( 8,1)(0,0)[c]@xmath109 ( 10.5,1)(0,0)[c]@xmath108 ( 14.8,1)(0,0)[c]@xmath220 ( 17.15,1)(0,0)[c]@xmath221 ( 19.8,1)(0,0)[c]@xmath222 ( 5.4,1)(0,0)[c]@xmath206 ( -1,1.2)(1.6,1.2)@xmath223 ( 1,1.5)= ( 2,1.2)(1.6,1.2)@xmath224 -3 cm ( 24,4 ) ( 8,2)(0,0)[c]@xmath73 ( 10.4,2)(0,0)[c]@xmath73 ( 14.85,2)(0,0)[c]@xmath73 ( 17.25,2)(0,0)[c]@xmath73 ( 19.4,2)(0,0)[c]@xmath73 ( 6,3.8)(0,0)[c]@xmath73 ( 6,.3)(0,0)[c]@xmath73 ( 8.4,2)(1,0)1.55 ( 10.82,2)(1,0)0.8 ( 13.2,2)(1,0)1.2 ( 15.28,2)(1,0)1.45 ( 17.7,2)(1,0)1.25 ( 19.8,2)(1,0)1.5 ( 7.6,2.2)(-1,1)1.3 ( 7.6,1.8)(-1,-1)1.3 ( 12.5,1.95)(0,0)[c]@xmath74 ( 22.6,1.95)(0,0)[c]@xmath74 ( 8.1,1)(0,0)[c]@xmath108 ( 10.7,1)(0,0)[c]@xmath225 ( 14.8,1)(0,0)[c]@xmath220 ( 17.15,1)(0,0)[c]@xmath221 ( 19.8,1)(0,0)[c]@xmath222 ( 4.5,3.8)(0,0)[c]@xmath109 ( 3.5,0.3)(0,0)[c]@xmath226 ( -1,1.2)(1.6,1.2)@xmath227 ( 1,1.5)= ( 2,1.2)(1.6,1.2)@xmath228 for the sake of completeness we also list the corresponding sets of positive roots . @xmath229 [ rem:000 ] it is easy to see that we have the following isomorphisms of lie superalgebras with identical dynkin diagrams : @xmath230 , @xmath231 , @xmath232 , @xmath233 .
we will replace the above matrix realization of the lie superalgebras with dynkin diagrams ( 23,0)@xmath98(-20,8)@xmath58 , ( 23,0)@xmath98(-20,8)@xmath59 and ( 23,0)@xmath98(-20,8)@xmath136 by their central extensions , for @xmath234 .
these central extensions will be convenient and conceptual for later formulation of truncation functors and super duality . let @xmath64 . consider the central extension @xmath235 of @xmath156 by the one - dimensional center @xmath236 determined by the @xmath237-cocycle @xmath238b),\quad a , b\in\gl(\wt{v}_m),\end{aligned}\ ] ] where @xmath239 and @xmath240 denotes the supertrace .
observe that the cocycle @xmath241 is a coboundary . indeed , as a vector space , @xmath242 , and let us denote by @xmath243 for @xmath244 to indicate that it is in @xmath235 .
then the map from @xmath235 to the direct sum of lie superalgebras @xmath245 , which sends @xmath243 to @xmath246 is an algebra isomorphism , i.e. , @xmath247= { [ x , y ] } ' + \tau(x , y ) k$ ] . for @xmath248 the restrictions of @xmath241 to the subalgebras @xmath185 give rise to respective central extensions , which in turn induce central extensions on @xmath249 and @xmath192 .
we denote such a central extension of @xmath192 or @xmath249 by @xmath58 ( respectively , @xmath59 and @xmath136 ) when it corresponds to the dynkin diagram ( 23,0)@xmath250(-20,8)@xmath58 in tables [ table1 ] and [ table2 ] ( respectively , ( 23,0)@xmath250(-20,8)@xmath59 and ( 23,0)@xmath98(-20,8)@xmath136 ) .
we make a trivial yet crucial observation that @xmath58 and @xmath59 are naturally subalgebras of @xmath136 .
the standard cartan subalgebras of @xmath58 , @xmath59 and @xmath136 will be denoted by @xmath251 , @xmath252 and @xmath253 , respectively .
@xmath251 , @xmath252 or @xmath253 has a basis @xmath254 with dual basis @xmath255 in the restricted dual @xmath256 , @xmath257 or @xmath258 , where @xmath162 runs over the index sets @xmath259 , @xmath260 or @xmath261 , respectively . here
@xmath262 is defined by letting @xmath263 for all relevant @xmath162 in each case .
* in the remainder of the paper we shall drop the superscript @xmath62 . * for example , we write @xmath2 , @xmath5 and @xmath264 for @xmath58 , @xmath59 and @xmath60 , with associated dynkin diagrams ( 23,0)@xmath98(-20,8)@xmath2 , ( 23,0)@xmath98(-20,8)@xmath5 and ( 23,0)@xmath98(-20,8)@xmath53 , respectively , where @xmath62 denotes a fixed type among @xmath13 .
in this section , we first introduce the categories @xmath265 , @xmath266 and @xmath267 of @xmath2-modules , @xmath5-modules and @xmath264-modules , respectively .
then we study the truncation functors which relate @xmath5 to finite dimensional lie superalgebras of @xmath4 type .
let @xmath64 be fixed .
we fix an arbitrary subset @xmath268 of @xmath269 .
let @xmath270 , @xmath271 and @xmath272 be the union of @xmath268 and the subset of simple roots of ( 23,0)@xmath71(-20,8)@xmath273 , ( 23,0)@xmath71(-20,8)@xmath274 and ( 23,0)@xmath250(-20,8)@xmath275 , respectively , with the leftmost one removed .
we have @xmath276 for @xmath201 . as @xmath270 , @xmath271 and @xmath272
are fixed , we will make the convention of suppressing them from notations below .
set @xmath277 , @xmath278 and @xmath279 to be the standard levi subalgebras of @xmath2 , @xmath5 and @xmath53 corresponding to the subsets @xmath270 , @xmath271 and @xmath272 , respectively .
the borel subalgebras of @xmath2 , @xmath5 and @xmath53 , spanned by the central element @xmath280 and upper triangular matrices , are denoted by @xmath81 , @xmath281 and @xmath282 , respectively .
let @xmath283 , @xmath284 and @xmath285 be the corresponding parabolic subalgebras with nilradicals @xmath286 , @xmath287 and @xmath288 and opposite nilradicals @xmath289 , @xmath290 and @xmath291 , respectively . given a partition @xmath292 , we denote by @xmath293 the length of @xmath294 and by @xmath295 its conjugate partition .
we also denote by @xmath296 the modified frobenius coordinates of @xmath294 : @xmath297 where @xmath298 let @xmath299 and @xmath300 be a partition .
the tuple @xmath301 is said to satisfy a dominant condition if @xmath302 for all @xmath303 , where @xmath304 denotes the coroot of @xmath305 .
associated to such a dominant tuple and each @xmath306 , we define the weights ( which will be called _ dominant _ ) @xmath307 we denote by @xmath308 , @xmath309 and @xmath310 the sets of all dominant weights of the form ( [ weight : im ] ) , ( [ weight : ovim ] ) and ( [ weight : wtim ] ) for all @xmath311 , respectively . by definition
we have bijective maps @xmath312 for @xmath313 , let @xmath314 denote the highest weight irreducible @xmath277-module of highest weight @xmath294 .
we extend @xmath314 to a @xmath315-module by letting @xmath286 act trivially .
define as usual the parabolic verma module @xmath316 and its irreducible quotient @xmath317 over @xmath2 : @xmath318 similarly , for @xmath319 , we define the irreducible @xmath278-module @xmath320 , the parabolic verma @xmath5-module @xmath321 and its irreducible @xmath5-quotient @xmath322 , as well as the irreducible @xmath323-module @xmath324 , the parabolic verma @xmath53-module @xmath325 and its irreducible @xmath53-quotient @xmath326 .
[ lem : paraverma ] let @xmath313 . * the restrictions to @xmath277 of the @xmath2-modules @xmath316 and @xmath317 decompose into direct sums of @xmath328 for @xmath329 . *
the restrictions to @xmath278 of the @xmath5-modules @xmath321 and @xmath322 decompose into direct sums of @xmath330 for @xmath331 . *
the restrictions to @xmath323 of the @xmath53-modules @xmath325 and @xmath326 decompose into direct sums of @xmath332 for @xmath329 .
part ( i ) is clear .
the proofs of ( ii ) and ( iii ) are analogous , and so we shall only give the proof for ( ii ) .
the @xmath278-module @xmath333 is a direct sum of irreducible modules of the form @xmath330 .
now the category of @xmath278-modules that have an increasing composition series with composition factors isomorphic to @xmath330 , with @xmath329 , is a semi - simple tensor category ( * ? ? ?
* section 3.2 ) .
thus @xmath334 also decomposes into a direct sum of @xmath330 with @xmath329 , and so does its irreducible quotient @xmath322 .
let @xmath335 be the category of @xmath2-modules @xmath336 such that @xmath336 is a semisimple @xmath337-module with finite dimensional weight subspaces @xmath338 , @xmath339 , satisfying * @xmath336 decomposes over @xmath340 into a direct sum of @xmath341 for @xmath342 .
* there exist finitely many weights @xmath343 ( depending on @xmath336 ) such that if @xmath344 is a weight in @xmath336 , then @xmath345 , for some @xmath170 .
the parabolic verma modules @xmath316 and irreducible modules @xmath317 for @xmath313 lie in @xmath335 , by lemma [ lem : paraverma ] .
analogously we define the categories @xmath17 and @xmath31 of @xmath5- and @xmath53-modules , respectively .
they also contain suitable parabolic verma and irreducible modules .
the morphisms in @xmath335 , @xmath17 and @xmath31 are all ( not necessarily even ) @xmath2- , @xmath346- and @xmath347-homomorphisms , respectively . for @xmath96 ,
recall the sets @xmath349 of simple roots for the dynkin diagrams .
the associated lie superalgebras @xmath23 , @xmath11 and @xmath348 can be identified naturally with the subalgebras of @xmath2 , @xmath5 and @xmath53 generated by @xmath280 and the root vectors of the corresponding dynkin diagrams in ( [ dynkin : combined ] ) , and moreover , @xmath350 for all @xmath10 .
observe that the @xmath11 s ( modulo the trivial central extensions ) are exactly all the finite dimensional lie superalgebras of @xmath4 type .
since @xmath351 , the standard cartan subalgebra of @xmath11 equals @xmath352 .
similarly , we use the notation @xmath353 and @xmath354 for the standard cartan subalgebras of @xmath23 and @xmath348 , respectively . recall the notation @xmath355 , @xmath356 , and @xmath357 from ( [ weight : im ] ) , ( [ weight : ovim ] ) and
( [ weight : wtim ] ) . given @xmath358 with @xmath359 for @xmath360
, we may regard it as a weight @xmath361 in a natural way .
similarly , for @xmath358 with @xmath362 for @xmath360 , we regard @xmath356 as a weight @xmath363 . finally ,
for @xmath358 with @xmath364 for @xmath360 , we regard @xmath357 as a weight @xmath365 .
the subsets of such weights @xmath366 in @xmath367 , @xmath368 and @xmath369 will be denoted by @xmath370 , @xmath371 and @xmath372 , respectively .
the corresponding parabolic verma and irreducible @xmath23-modules are denoted by @xmath373 and @xmath374 , respectively , with @xmath375 , while the corresponding category of @xmath23-modules is denoted by @xmath376 .
similarly , we introduce the self - explanatory notations @xmath377 , @xmath378 , @xmath379 , and @xmath380 , @xmath381 , @xmath382 for @xmath11- and @xmath348-modules , respectively .
let @xmath383 .
for @xmath384 , we can write @xmath385 , where @xmath344 runs over @xmath386 .
the _ truncation functor _
@xmath387 is defined by sending @xmath388 to @xmath389 , summed over @xmath390 . when it is clear from the context we shall also write @xmath391 instead of @xmath392 . analogously , truncation functors @xmath393 and @xmath394 are defined .
[ lem : trunc ] let @xmath383 and @xmath395 . * for @xmath396 we have @xmath397 * for @xmath398 we have @xmath399 * for @xmath400 we have @xmath401 we will show ( i ) only .
the proofs of ( ii ) and ( iii ) are similar . since @xmath402
, it is enough to show ( i ) for @xmath403 .
suppose that @xmath404 for all @xmath360 .
let @xmath405 denote the standard levi subalgebra of @xmath2 corresponding to the removal of the vertex @xmath406 of the dynkin diagram of @xmath2 .
then @xmath407 .
now @xmath317 is the unique irreducible quotient of the @xmath2-module obtained via parabolic induction from the @xmath408-module @xmath374 ( where the @xmath23-module @xmath374 is extended to an @xmath408-module by a trivial action of @xmath409 ) .
our choice of the levi subalgebra and of the opposite nilradical assures that this parabolically induced module truncates to @xmath374 .
thus its irreducible quotient @xmath317 also truncates to @xmath374 .
the remaining case in ( i ) is clear .
the central extensions introduced in section [ sec : ext ] allow us to study , in a uniform fashion , modules whose weights stabilize at any @xmath410 ( not just at @xmath411 ) .
for example , if @xmath294 is a weight with @xmath412 , for @xmath413 , then , without central extensions , the usual truncation functors would always truncate an irreducible or parabolic verma of such a highest weight to zero .
a way around central extensions is to define truncation functors depending on each @xmath410 .
this approach , although equivalent , looks less elegant .
in this section we introduce two functors @xmath414 and @xmath415 , and then establish a fundamental property of these functors ( theorem [ matching : modules ] ) . as a consequence
, we obtain an irreducible @xmath4-character formula in a parabolic category @xmath0 in terms of the kl polynomials of @xmath1 types ( theorem [ character ] and remark [ rem : sol ] ) .
the kazhdan - lusztig polynomials in the categories @xmath40 , @xmath21 and @xmath36 in terms of kostant @xmath37-homology groups are shown to match perfectly with one another ( theorem [ matching : klpol ] ) .
let @xmath416 be a lie superalgebra with a borel subalgebra @xmath417 with corresponding sets of simple and positive roots @xmath418 and @xmath419 , respectively . as usual
, for a positive root @xmath420 , we let @xmath421 denote a root vector associated to root @xmath422 .
let @xmath305 be an isotropic odd simple root in @xmath418 and @xmath304 be its corresponding coroot .
the set @xmath423 forms a new set of positive roots whose corresponding set of simple roots is @xmath424 we shall denote by @xmath425 the corresponding new borel subalgebra .
the process of such a change of borel subalgebras is referred to as _ odd reflection _ with respect to @xmath305 @xcite .
the following simple and fundamental lemma for odd reflections has been used by many authors ( cf .
e.g. @xcite ) .
[ hwt odd ] let @xmath426 be a simple @xmath416-module of @xmath427-highest weight @xmath428 and let @xmath143 be a @xmath427-highest weight vector .
let @xmath305 be a simple isotropic odd root in @xmath418 . 1 .
if @xmath429 , then @xmath426 is a @xmath430-module of @xmath425-highest weight @xmath428 and @xmath143 is a @xmath425-highest weight vector .
2 . if @xmath431 , then @xmath426 is a @xmath416-module of @xmath425-highest weight @xmath432 and @xmath433 is a @xmath425-highest weight vector .
fix @xmath436 .
starting with the third dynkin diagram in associated to @xmath53 , we apply the following sequence of @xmath437 odd reflections
. first we apply one odd reflection corresponding to @xmath112 , then we apply two odd reflections corresponding to @xmath208 and @xmath438 . after
that we apply three odd reflections corresponding to @xmath439 , @xmath440 , and @xmath441 , et cetera , until finally we apply @xmath10 odd reflections corresponding to @xmath442 .
the resulting new borel subalgebra for @xmath53 will be denoted by @xmath434 and its corresponding simple roots are listed in the following dynkin diagram : -6 cm ( 24,4 ) ( 15.25,2)(0,0)[c]@xmath101 ( 17.4,2)(0,0)[c]@xmath73 ( 21.9,2)(0,0)[c]@xmath73 ( 24.4,2)(0,0)[c]@xmath101 ( 26.8,2)(0,0)[c]@xmath101 ( 13.2,2)(1,0)1.45 ( 15.7,2)(1,0)1.25 ( 17.8,2)(1,0)0.9 ( 20.1,2)(1,0)1.4 ( 22.35,2)(1,0)1.6 ( 24.9,2)(1,0)1.5 ( 27.3,2)(1,0)1.5 ( 19.5,1.95)(0,0)[c]@xmath74 ( 29.7,1.95)(0,0)[c]@xmath74 ( 15.2,3)(0,0)[c]@xmath443 ( 17.8,1)(0,0)[c]@xmath109 ( 22,1)(0,0)[c]@xmath444 ( 24.4,1)(0,0)[c]@xmath445 ( 27,1)(0,0)[c]@xmath446 ( 8.0,2)(0,0)[c](1.6,1.2)@xmath447 ( 8.8,2)(1,0)1.7 ( 11.8,2)(0,0)[c](2.6,1.2)@xmath448 the crucial point here is that the subdiagram to the left of the first @xmath101 is the dynkin diagram of @xmath449 . on the other hand , starting with the third dynkin diagram in associated to @xmath53 , we apply the following new sequence of @xmath437 odd reflections .
first we apply one odd reflection corresponding to @xmath115 , then we apply two odd reflections corresponding to @xmath450 and @xmath451 . after
that we apply three odd reflections corresponding to @xmath452 , @xmath453 , and @xmath454 , et cetera , until finally we apply @xmath10 odd reflections corresponding to @xmath455 .
the resulting new borel subalgebra for @xmath53 will be denoted by @xmath435 and its corresponding simple roots are listed in the following dynkin diagram : -6 cm ( 24,4 ) ( 15.25,2)(0,0)[c]@xmath101 ( 17.4,2)(0,0)[c]@xmath73 ( 21.9,2)(0,0)[c]@xmath73 ( 24.4,2)(0,0)[c]@xmath101 ( 26.8,2)(0,0)[c]@xmath101 ( 13.28,2)(1,0)1.45 ( 15.7,2)(1,0)1.25 ( 17.8,2)(1,0)0.9 ( 20.1,2)(1,0)1.4 ( 22.35,2)(1,0)1.6 ( 24.9,2)(1,0)1.5 ( 27.3,2)(1,0)1.5 ( 19.5,1.95)(0,0)[c]@xmath74 ( 29.7,1.95)(0,0)[c]@xmath74 ( 15.5,3)(0,0)[c]@xmath456 ( 17.8,1)(0,0)[c]@xmath103 ( 22,1)(0,0)[c]@xmath212 ( 24.4,1)(0,0)[c]@xmath446 ( 27,1)(0,0)[c]@xmath457 ( 8.0,2)(0,0)[c](1.6,1.2)@xmath447 ( 8.8,2)(1,0)1.7 ( 11.8,2)(0,0)[c](2.6,1.2)@xmath458 we remark that the subdiagram to the left of the odd simple root @xmath456 above becomes the dynkin diagram of @xmath459 .
recall the standard levi subalgebra @xmath323 of @xmath53 with ( opposite ) nilradical @xmath460 and @xmath461 ( see section [ sec : obcd ] ) .
[ lem : uinvar ] the sequences of odd reflections in [ oddreflection ] leave the sets of roots of @xmath460 and @xmath461 invariant .
this follows from the fact that the simple roots used in the sequences of odd reflections in section [ oddreflection ] are all roots of @xmath323 .
we denote by @xmath462 and @xmath463 the borel subalgebras of @xmath323 corresponding to the sets of simple roots @xmath464 and @xmath465 , respectively .
the sequences of odd reflections in [ oddreflection ] only affect the tail diagram ( 20,0)@xmath98(-20,8)@xmath466 and leaves the head diagram ( 20,0)@xmath71(-20,7)@xmath447 invariant . since the tail diagram is of type @xmath3 , the proofs of ( * ? ? ?
* lemma 3.2 ) and ( * ? ? ?
* corollary 3.3 ) can be adapted in a straightforward way to prove the following ( where lemma [ lem : uinvar ] is used ) .
[ prop : change ] let @xmath358 and @xmath96 . *
suppose that @xmath467 .
then the highest weight of @xmath468 with respect to the borel subalgebra @xmath462 is @xmath428 .
furthermore , @xmath469 and @xmath470 are highest weight @xmath53-modules of highest weight @xmath428 with respect to the new borel subalgebra @xmath434 .
* suppose that @xmath471 .
then the highest weight of @xmath468 with respect to the borel subalgebra @xmath463 is @xmath356 .
furthermore , @xmath469 and @xmath470 are highest weight @xmath53-modules of highest weight @xmath356 with respect to the new borel subalgebra @xmath435 .
[ sec : t ] by definition , @xmath2 and @xmath5 are naturally subalgebras of @xmath53 , @xmath473 and @xmath278 are subalgebras of @xmath323 , while @xmath474 and @xmath475 are subalgebras of @xmath476 .
also , we may regard @xmath477 and @xmath478 .
given a semisimple @xmath476-module @xmath479 , we define @xmath480 note that @xmath481 is an @xmath337-submodule of the @xmath482 , and @xmath483 is an @xmath475-submodule of @xmath482 . one checks that if @xmath479 is an @xmath323-module , then @xmath481 is an @xmath340-submodule of @xmath482 and @xmath483 is an @xmath278-submodule of @xmath482 .
furthermore , if @xmath479 is a @xmath53-module , then @xmath481 is a @xmath484-submodule of @xmath482 and @xmath483 is a @xmath485-submodule of @xmath482 .
the direct sum decomposition in @xmath482 gives rise to the natural projections @xmath486 that are @xmath337- and @xmath475-module homomorphisms , respectively .
if @xmath487 is an @xmath476-homomorphism , then the following induced maps @xmath488 : t(\wt{m } ) @>>>t(\wt{n } ) \qquad \hbox{and}\qquad \ov{t}[\wt{f } ] : \ov{t}(\wt{m } ) @>>>\ov{t}(\wt{n } ) \endcd\end{aligned}\ ] ] are also @xmath337- and @xmath475-module homomorphisms , respectively . also if @xmath487 is a @xmath489-homomorphism , then @xmath490 and @xmath491 $ ] ( respectively , @xmath492 and @xmath493 $ ] ) are @xmath484- ( respectively , @xmath485- ) homomorphisms .
[ lmod2lmod ] for @xmath358 , we have @xmath494 , and @xmath495 .
we shall prove the first formula using a character argument , and the second one can be proved similarly .
associated to partitions @xmath496 , we denote by @xmath497 and @xmath498 the schur and skew schur functions in the variables @xmath499 .
the hook schur functions associated to @xmath428 is defined to be ( cf .
@xcite ) @xmath500 for a dominant tuple @xmath301 , we have ( cf .
@xcite ) @xmath501 here @xmath502 for each @xmath162 , and @xmath503 denotes the irreducible @xmath504-module of highest weight @xmath505 .
note that @xmath506 .
as an @xmath277-module , @xmath468 is completely reducible . on the character level , applying @xmath34 to @xmath468 corresponds to setting @xmath507 in the character formula ( [ eqn : hookschur ] ) to zero . by ( [ hookschur : char ] )
, @xmath508 is an @xmath277-module with character @xmath509 , which is precisely the character of @xmath510 .
this proves the formula .
[ functor ] @xmath34 and @xmath472 define exact functors from @xmath31 to @xmath335 and from @xmath31 to @xmath17 , respectively .
the following theorem can be regarded as a weak version of the super duality which is to be established in theorem [ thm : equivalence ] .
[ matching : modules ] let @xmath358 .
if @xmath482 is a highest weight @xmath264-module of highest weight @xmath357 , then @xmath481 and @xmath483 are highest weight @xmath484- and @xmath485-modules of highest weights @xmath428 and @xmath356 , respectively .
furthermore , we have @xmath511 we will prove only the statements involving @xmath34 , and the statements involving @xmath472 can be proved in the same way . by proposition [ prop :
change ] , @xmath482 contains a @xmath512-highest weight vector @xmath513 of highest weight @xmath428 for @xmath514 .
the vector @xmath513 clearly lies in @xmath481 , and it is a @xmath515-singular vector since @xmath516 .
now @xmath481 is completely reducible over @xmath517 with all highest weights of its irreducible summands lying in @xmath518 .
thus to show that @xmath481 is a highest weight @xmath2-module it remains to show that any vector in @xmath481 of weight in @xmath518 is contained in @xmath519 .
this follows by the same argument in ( * ? ? ?
* lemma 3.5 ) , which only relies on the @xmath3-type tail diagram of @xmath53 .
let us write @xmath520 and @xmath521 .
we observe that all the weights in @xmath522 , @xmath510 , @xmath523 , and @xmath468 are of the form @xmath524 with @xmath525 .
since also @xmath526 , it follows by lemma [ lmod2lmod ] that @xmath527 .
since @xmath528 is a highest weight module of highest weight @xmath428 , we have @xmath529 . to show that @xmath34 sends irreducibles to irreducibles
we show that @xmath530 has no singular vector apart from the scalar multiples of a highest weight vector .
we argue by assuming otherwise and derive a contradiction .
if we have another singular vector of weight different from @xmath428 , then we can show , following the second part of the proof of ( * ? ? ?
* theorem 3.6 ) , that we also have a singular vector in @xmath470 of weight different from @xmath357 .
the argument there is applicable here , since it again only depends on the tail diagram , which is of type @xmath3 .
it can be shown that tilting modules exist in categories @xmath531 ( cf .
@xcite for type @xmath3 ) and that the functors @xmath34 and @xmath35 respect the tilting modules .
we choose not to develop the details in order to keep the paper to a reasonable length . by standard arguments
theorem [ matching : modules ] implies the following character formula .
[ character ] let @xmath358 , and write @xmath532 , @xmath533 .
then * @xmath534 , * @xmath535 .
[ rem : sol ] the transition matrix @xmath536 in theorem [ character ] is known according to the kazhdan - lusztig theory .
this is because the kazhdan - lusztig polynomials in the bgg category @xmath537 also determine the composition factors of generalized verma modules in the corresponding parabolic subcategory ( see e.g. ( * ? ? ?
* and proposition 7.5 ) ) .
hence theorem [ character ] and lemma [ lem : trunc ] provide a complete solution to the irreducible character problem in the category @xmath22 for the ortho - symplectic lie superalgebras . for a precise definition of homology groups of lie superalgebras with coefficients in a module and a precise formula for the boundary operator
we refer the reader to ( * ? ? ?
* section 4 ) or @xcite . for @xmath538
we denote by @xmath539 and @xmath540 .
furthermore let @xmath541 be the boundary operator of the complex of @xmath461-homology groups with coefficients in @xmath482 , regarded as a @xmath461-module .
the map @xmath542 is an @xmath323-module homomorphism and hence the homology groups @xmath543 are @xmath323-modules , for @xmath544 .
accordingly we let @xmath545 and @xmath546 stand for the boundary operator of the complex of @xmath289-homology with coefficients in @xmath388 and the boundary operator of the complex of @xmath547-homology with coefficients in @xmath548 , respectively .
similarly , @xmath549 and @xmath550 are @xmath340- and @xmath551-homomorphisms , respectively . [ boundary ] for @xmath552 and @xmath358 , we have * @xmath553 , and thus @xmath554 moreover , @xmath555=d$ ] .
* @xmath556 , and thus @xmath557 moreover , @xmath558=\ov{d}$ ] . we will prove ( i ) only .
it follows by definition of @xmath34 and @xmath461 that @xmath559 .
now , since all modules involved have weights of the form @xmath560 with @xmath525 , it follows that @xmath561 and @xmath562 have the same character .
complete reducibility of the @xmath277-modules @xmath561 and @xmath562 implies that @xmath563 as @xmath277-modules .
theorem [ matching : modules ] completes the proof of the first part of ( i ) .
the second part of ( i ) follows from the definitions of @xmath542 and @xmath564 ( see e.g. ( * ? ? ? * ( 4.1 ) ) ) . by lemma [ boundary ]
we have the following commutative diagram . @xmath565
thus @xmath34 induces an @xmath277-homomorphism from @xmath566 to @xmath567 .
similarly , @xmath472 induces an @xmath551-homomorphism from @xmath566 to @xmath568 . as an @xmath569-module
, @xmath570 is a direct sum of @xmath571 , @xmath313 , each appearing with finite multiplicity ( ( * ? ? ?
* section 3.2.3 ) ) . by (
* theorem 3.2 ) , @xmath572 as an @xmath569-module is completely reducible . write @xmath573 as @xmath569-modules .
it follows by lemmas [ lmod2lmod ] and [ lem : aux3 ] that @xmath574 , as @xmath340-modules .
similarly , @xmath575 , as @xmath551-modules .
the commutativity of ( [ compare - complexes ] ) and lemma [ boundary ] now allow us to adapt the proof of ( * ? ? ? * theorem 4.4 ) to prove the following .
[ matching : kl ] we have for @xmath576 * @xmath577 , as @xmath277-modules .
* @xmath578 , as @xmath551-modules . setting @xmath579 in theorem [ matching : kl ] and using theorem [ matching : modules ] we obtain the following .
[ matching : kl1 ] for @xmath358 and @xmath576 , we have * @xmath580 , as @xmath277-modules .
* @xmath581 , as @xmath551-modules .
we define parabolic kazhdan - lusztig polynomials in the categories @xmath40 , @xmath21 and @xmath36 for @xmath582 by letting @xmath583}\big { ) } ( -q)^{-n},\\ \ov{\ell}_{\mu^\natural\la^\natural}(q ) : = \sum_{n=0}^\infty\text{dim}_\c\big{(}\text{hom}_{\ov{\mf{l}}}\big { [ } l(\ov{\mf{l}},\mu^\natural ) , h_n\big{(}{\ov{\mf{u}}}_-;{l}(\la^\natural)\big { ) } \big{]}\big { ) } ( -q)^{-n},\\ \wt{\ell}_{\mu^\theta\la^\theta}(q ) : = \sum_{n=0}^\infty\text{dim}_\c\big{(}\text{hom}_{\wt{\mf{l}}}\big { [ } l(\wt{\mf{l}},\mu^\theta ) , h_n\big{(}{\wt{\mf{u}}}_-;{l}(\la^\theta)\big { ) } \big{]}\big { ) } ( -q)^{-n}.\end{aligned}\ ] ] by vogan s homological interpretation of the kazhdan - lusztig polynomials ( * ? ? ?
* conjecture 3.4 ) and the kazhdan - lusztig conjectures @xcite , proved in @xcite , @xmath584 coincides with the original definition and moreover @xmath585 ( cf .
theorem [ character ] ) .
the following reformulation of corollary [ matching : kl1 ] is a generalization of theorem [ character ] .
[ matching : klpol ] for @xmath586 we have @xmath587
[ [ section-5 ] ] the following is standard ( see , for example , ( * ? ? ?
* lemma 2.1.10 ) ) . [ ku ]
let @xmath588 .
then there exists a ( possibly infinite ) increasing filtration @xmath589 of @xmath484-modules such that * @xmath590 , * @xmath591 is a highest weight module of highest weight @xmath592 with @xmath593 , for @xmath594 , * the condition @xmath595 implies that @xmath596 , * for any weight @xmath294 of @xmath336 , there exists an @xmath597 such that @xmath598 .
similar statements hold for @xmath599 and @xmath600 .
let @xmath601 denote the full subcategory of @xmath31 consisting of finitely generated @xmath602-modules .
the categories @xmath603 and @xmath604 are defined in a similar fashion .
proposition [ ku ] implies the following .
[ filtration ] let @xmath605 .
then @xmath606 if and only if there exists a finite increasing filtration @xmath607 of @xmath484-modules such that @xmath591 is a highest weight module of highest weight @xmath592 with @xmath608 , for @xmath609 .
similar statements hold for @xmath610 and @xmath600 .
the following proposition is the converse to theorem [ matching : modules ] .
[ hw - onto ] * if @xmath611 is a highest weight @xmath484-module of highest weight @xmath612 , then there is a highest weight @xmath489-module @xmath613 of highest weight @xmath357 such that @xmath614 . *
if @xmath615 is a highest weight @xmath485-module of highest weight @xmath356 with @xmath358 , then there is a @xmath489-module @xmath613 of highest weight @xmath357 such that @xmath616
. we shall only prove ( i ) , as ( ii ) is similar .
we let @xmath182 be the kernel of the natural projection from the @xmath617 to @xmath611 .
now theorem [ matching : modules ] says that @xmath618 .
thus , by the exactness of functor @xmath34 , it suffices to prove that @xmath182 lifts to a submodule @xmath619 of @xmath469 such that @xmath620 .
there is an increasing filtration @xmath621 of @xmath2-modules for @xmath182 satisfying the properties of proposition [ ku ] . for each @xmath622 ,
let @xmath169 be a weight vector in @xmath623 such that @xmath624 is a non - zero highest weight vector of @xmath625 .
observe that @xmath626 and @xmath627 are completely reducible @xmath323- and @xmath277-modules , respectively .
then , for each @xmath622 , there is a highest weight vector @xmath628 of the @xmath323-module @xmath629 with respect to the borel subalgebra @xmath630 .
let @xmath631 be the submodule of @xmath469 generated by @xmath632 and set @xmath633 .
it is easy to see @xmath628 is a highest weight vector of the @xmath264-module @xmath634 .
let @xmath635 .
it is clear that @xmath636 for all @xmath170 .
this implies @xmath637 for all @xmath170 and hence @xmath620 .
define an equivalence relation @xmath640 on @xmath641 by letting @xmath642 if and only if @xmath643 lies in the root lattice @xmath644 of @xmath53 . for each such equivalence class @xmath645 $ ] , fix a representative @xmath645^o \in \wt{\h}^*$ ] and declare @xmath645^o$ ] to have @xmath146-grading @xmath646 . for @xmath647 , set ( cf . (
* -e ) and ( * ? ? ?
* section 2.5 ) for type @xmath3 ) @xmath648^o ) ( \widehat{e}_{r})\equiv { \epsilon } \,\,(\text{mod } 2 ) \big \ } , \text { for } \xx = \mf{b , c , d } , \\ { \wt{\h}^*}_\epsilon & = & \big \{\mu\in \wt{\h}^*\mid \sum_{i=1}^m ( \mu-[\mu]^o)(\widehat{e}_{-i } ) + \sum_{r\in \n}(\mu-[\mu]^o)(\widehat{e}_{r})\equiv { \epsilon } \,\,(\text{mod } 2 ) \big \ } , \text { for } \xx = \mf b^\bullet.\end{aligned}\ ] ] recall that @xmath649 is a semisimple @xmath476-module with @xmath650 . then @xmath651 acquires a natural @xmath146-grading @xmath652 given by @xmath653 which is compatible with the @xmath146-grading on @xmath53 . we define @xmath654 and @xmath638 to be the full subcategories of @xmath31 and @xmath601 , respectively , consisting of objects with @xmath146-gradation given by ( [ wt - z2-gradation ] ) .
note that the morphisms in @xmath654 and @xmath638 are of degree @xmath655 .
for @xmath538 , let @xmath656 denote the @xmath264-module @xmath482 equipped with the @xmath146-gradation given by ( [ wt - z2-gradation ] ) .
it is clear that @xmath657 is isomorphic to @xmath482 in @xmath31 .
thus @xmath31 and @xmath654 have isomorphic skeletons and hence they are equivalent categories . similarly , @xmath601 and @xmath638 are equivalent categories .
analogously define @xmath658 , @xmath659 , @xmath660 and @xmath639 to be the respective full subcategories of @xmath27 , @xmath661 , @xmath17 and @xmath603 consisting of objects with @xmath146-gradation given by ( [ wt - z2-gradation ] ) .
similarly , @xmath662 , @xmath663 , and also @xmath664 , @xmath665 .
( in case of @xmath40 and @xmath666 , these remarks are trivial except for the type @xmath667 which corresponds to a lie superalgebra ) .
recall the functors @xmath34 and @xmath472 from section [ tfunctors ] .
the following is the main result of this section .
[ thm : equivalence ] * @xmath668 is an equivalence of categories .
* @xmath669 is an equivalence of categories .
hence , the categories @xmath335 and @xmath17 are equivalent . since @xmath670 and @xmath671 it is enough to prove theorem [ thm : equivalence ] for @xmath660 and @xmath654 .
in order to keep notation simple we will from now on drop the superscript @xmath655 and use @xmath17 , @xmath31 , @xmath603 and @xmath601 to denote the respective categories @xmath660 , @xmath654 , @xmath639 and @xmath638 for the remainder of section [ sec : category ] .
henceforth , when we write @xmath672 , @xmath358 , we will mean the corresponding modules equipped with the @xmath146-gradation ( [ wt - z2-gradation ] ) .
similar convention applies to @xmath673 and @xmath674 . for @xmath675 and @xmath676 the @xmath170th extension group @xmath677
can be understood in the sense of baer - yoneda ( see e.g. ( * ? ? ?
* chapter vii ) ) and @xmath678 . in a similar way extensions in @xmath17 and @xmath31 are interpreted . from this viewpoint
the exact functors @xmath34 and @xmath472 induce natural maps on extensions by taking the projection of the corresponding exact sequences .
[ thm : equivalence - f.g .
] we have the following .
* @xmath679 is an equivalence of categories .
* @xmath680 is an equivalence of categories .
* the categories @xmath661 and @xmath603 are equivalent . theorem [ thm : equivalence - f.g .
] can be proved following a similar strategy as the one used to prove ( * ? ? ?
* theorem 5.1 ) . to avoid repeating similar arguments at great length
we will just point out the main differences between their proofs . in (
* section 5 ) the main point is to prove that the functor @xmath34 induces isomorphisms @xmath681 and @xmath682 , for @xmath683 irreducible , and @xmath684 having finite composition series . from this (
* theorem 5.1 ) is derived easily . as the isomorphism of the @xmath685 spaces imply the isomorphism of the @xmath686 spaces ( * ? ? ?
* lemma 5.12 ) , we are reduced to establish the isomorphism of the @xmath685 spaces . to prove the isomorphism of the @xmath685 spaces therein , the idea is to prove this isomorphism first for @xmath482 irreducible , and then to use induction on the length of the composition series of @xmath482 to establish the general case .
now , thanks to proposition [ hw - onto ] , we can proceed similarly as in ( * ? ? ?
* section 5 ) to prove theorem [ thm : equivalence - f.g . ] . for this purpose
we replace @xmath683 by a highest weight module , and @xmath482 and @xmath687 by finitely generated modules .
as finitely generated modules possess finite filtrations whose subquotients are highest weight modules ( cf .
proposition [ filtration ] ) , we can now borrow the same type of induction arguments from @xcite , now inducting on the length of such a filtration instead of the length of a composition series .
therefore , the proof of the isomorphisms is again reduced to a special case , namely when @xmath482 is a highest weight module .
this case can then be proved using similar arguments as the ones given in the proof of ( * ? ? ?
* lemmas 5.8 ) .
the case of @xmath472 is completely analogous . having theorem [ thm : equivalence - f.g . ] at our disposal we can now prove theorem [ thm : equivalence ] . since the proofs of ( i ) and ( ii ) are similar , we shall only prove ( i ) .
( iii ) follows from ( i ) and ( ii ) . for every @xmath688 ,
there is an increasing filtration @xmath689 of @xmath2-modules for @xmath388 with @xmath690 satisfying the properties of proposition [ ku ] .
the filtration @xmath691 of @xmath388 lifts to a filtration @xmath692 with @xmath693 such that @xmath694 by theorem [ thm : equivalence - f.g . ] .
it is clear that we have @xmath695 and @xmath696 .
it is well known that a full and faithful functor @xmath697 , satisfying the property that for every @xmath698 there exists @xmath699 with @xmath700 , is an equivalence of categories ( see e.g. ( * ? ? ?
* proposition 1.5.2 ) ) .
therefore it remains to show that @xmath34 is full and faithful . by proposition
[ ku ] , for @xmath701 , we may choose an increasing filtration of @xmath53-modules @xmath702 such that @xmath703 and @xmath704 is a highest weight module of highest weight @xmath705 with @xmath593 , for @xmath594
. then the direct limit of @xmath706 is @xmath707 and @xmath708 for every @xmath709 .
similarly we have @xmath710 and @xmath711 for @xmath712 .
furthermore , we have the following commutative diagram ( where @xmath713 ) : @xmath714 using a similar argument as the one given in ( * ? ? ?
* lemma 5.10 ) , where we replace the induction on the length of composition series therein by induction on the length of finite increasing filtration @xmath715 , we show that @xmath716 are isomorphisms for each @xmath170 .
therefore @xmath717 is an isomorphism and hence @xmath718 is an isomorphism .
this completes the proof .
the main purpose of this section is to determine the extremal weights of finite dimensional irreducible modules over the ortho - symplectic lie superalgebras with integral highest weights .
it follows that all such finite dimensional irreducible modules for the ortho - symplectic lie superalgebras are in the category @xmath22 .
we note that the finite dimensional irreducible modules of non - integral highest weights are typical and so their characters are known ( * ? ? ? * theorem 1 ) .
let us denote the weights of the natural @xmath719-module @xmath720 by @xmath721 for @xmath722 .
we call a weight _ integral _ , if it lies in @xmath54-span of the @xmath723 s and @xmath724 s . the _ standard _ borel subalgebra @xmath725 of @xmath719 is the one associated to the following set of simple roots ( 24,4 ) ( 6,2)(0,0)[c]@xmath73 ( 8.4,2)(0,0)[c]@xmath73 ( 10.5,1.95)(0,0)[c]@xmath74 ( 12.85,2)(0,0)[c]@xmath73 ( 15.25,2)(0,0)[c]@xmath101 ( 17.4,2)(0,0)[c]@xmath73 ( 19.6,1.95)(0,0)[c]@xmath74 ( 21.9,2)(0,0)[c]@xmath73 ( 24.2,2)(0,0)[c]@xmath73 ( 6.4,2)(1,0)1.55 ( 8.82,2)(1,0)0.8 ( 11.2,2)(1,0)1.2 ( 13.28,2)(1,0)1.45 ( 15.7,2)(1,0)1.25 ( 17.8,2)(1,0)1.0 ( 20.3,2)(1,0)1.2 ( 22.3,1.8)@xmath84 ( 5.6,1)(0,0)[c]@xmath726 ( 8.4,1)(0,0)[c]@xmath727 ( 12.8,1)(0,0)[c]@xmath728 ( 15.15,3)(0,0)[c]@xmath729 ( 17.4,1)(0,0)[c]@xmath730 ( 21.8,1)(0,0)[c]@xmath731 ( 24.5,1)(0,0)[c]@xmath732 an arbitrary dynkin diagram for @xmath719 always has a type @xmath3 end while the other end is a short ( even or odd ) root . starting from the type @xmath3 end , the simple roots for a borel subalgebra @xmath427 of @xmath719 give rise to a sequence of @xmath733 @xmath734 s , @xmath735 @xmath736 s , @xmath737 @xmath734 s , @xmath738 @xmath736 s , @xmath739 @xmath734 s , @xmath740 @xmath736 s and sequences of @xmath741 s : @xmath742 ( all the @xmath743 and @xmath744 are positive except possibly @xmath745 or @xmath746 ) .
note that a dynkin diagram contains a short _ odd _ root exactly when @xmath746 .
let @xmath747 for @xmath748 , and let @xmath749 .
note @xmath750 .
more precisely , there exist a permutation @xmath163 of @xmath751 and a permutation @xmath752 of @xmath753 , so that the simple roots for @xmath427 are given by @xmath754 @xcite [ kac hwt ] the irreducible @xmath719-module of integral highest weight @xmath761 with respect to the standard borel subalgebra is finite dimensional if and only if @xmath762 for some @xmath756-hook partition @xmath428 .
we denote by @xmath763 these irreducible @xmath719-modules with respect to the standard borel subalgebra , to distinguish from earlier notation used for irreducible modules with respect to different borel subalgebra .
actually the finite dimensionality criterion was given in @xcite in terms of dynkin labels , which is known to be equivalent to the more natural labeling above in terms of @xmath756-hook partitions ( cf .
same remark applies to the finite dimensionality criterion for @xmath764 in lemma [ hwt : standard : spo ] below .
( 24,4 ) ( 6,2)(0,0)[c]@xmath73 ( 8.4,2)(0,0)[c]@xmath101 ( 10.5,1.95)(0,0)[c]@xmath73 ( 12.85,2)(0,0)[c]@xmath101 ( 15.25,2)(0,0)[c]@xmath73 ( 17.4,2)(0,0)[c]@xmath101 ( 19.6,1.95)(0,0)[c]@xmath73 ( 21.9,2)(0,0)[c]@xmath101 ( 24.2,2 ) ( 6.4,2)(1,0)1.55 ( 8.82,2)(1,0)1.3 ( 10.8,2)(1,0)1.6 ( 13.28,2)(1,0)1.45 ( 15.7,2)(1,0)1.25 ( 17.8,2)(1,0)1.4 ( 19.9,2)(1,0)1.5 ( 22.3,1.8)@xmath84 ( 5.8,3)(0,0)[c]@xmath766 ( 8.4,1)(0,0)[c]@xmath767 ( 10.4,3)(0,0)[c]@xmath768 ( 12.8,1)(0,0)[c]@xmath769 ( 15.15,3)(0,0)[c]@xmath770 ( 17.4,1)(0,0)[c]@xmath771 ( 19.4,3)(0,0)[c]@xmath772 ( 21.8,1)(0,0)[c]@xmath773 ( 24,1)(0,0)[c]@xmath774 we read off from the above a signed sequence with indices @xmath775 . in particular , we obtain a sequence @xmath776 by ignoring the signs and indices . in this case , @xmath777 , and @xmath778 .
furthermore , the sequences @xmath779 and @xmath780 are @xmath781 and @xmath782 , respectively .
define the _ block frobenius coordinates _
@xmath783 of an @xmath756-hook partition @xmath428 associated to @xmath427 as follows . for @xmath784 ,
let @xmath785 it is elementary to read off the block frobenius coordinates of @xmath428 from the young diagram of @xmath428 in general , as illustrated by the next example .
[ frob : block ] consider the @xmath786-hook diagram @xmath787 the block frobenius coordinates associated with @xmath427 from example [ diag 9|10 ] for @xmath428 is : @xmath788 these are read off from the young diagram of @xmath428 by following the @xmath789 sequence @xmath776 as follows : ( 15,9 ) ( 0,6)(0,1)2 ( 0,8)(1,0)14 ( 14,8)(0,-1)1 ( 14,7)(-1,0)3 ( 11,7)(0,-1)1 ( 11,6)(-1,0)11 ( 0,6)(0,-1)6 ( 0,0)(1,0)2 ( 2,0)(0,1)6 ( 2,4)(1,0)6 ( 8,4)(0,1)2 ( 2,1)(1,0)1 ( 3,1)(0,1)1 ( 3,2)(1,0)1 ( 4,2)(0,1)2 ( 4,3)(1,0)3 ( 7,3)(0,1)1 ( 5.5,7.5)(0,0)[c]@xmath790 ( 5.5,6.5)(0,0)[c]@xmath791 ( 5.5,5.5)(0,0)[c]@xmath792 ( 5.5,4.5)(0,0)[c]@xmath793 ( 5.5,3.5)(0,0)[c]@xmath794 ( 0.5,3.4)(0,0)[c]@xmath795 ( 0.5,2)(0,0)[c]@xmath796 ( 0.5,2.7)(0,0)[c]@xmath797 ( 1.5,3.4)(0,0)[c]@xmath795 ( 1.5,2)(0,0)[c]@xmath796 ( 2.5,3.4)(0,0)[c]@xmath795 ( 2.5,2)(0,0)[c]@xmath796 ( 3.5,3.4)(0,0)[c]@xmath795 ( 3.5,2.3)(0,0)[c]@xmath796 ( 1.5,2.7)(0,0)[c]@xmath798 ( 2.5,2.7)(0,0)[c]@xmath799 ( 3.5,2.8)(0,0)[c]@xmath800 ( 4,3)(0,-1)3 ( 4,3)(1,0)10 ( -0.1,3)(1,0)0.2 ( 4,7.9)(0,1)0.2 ( -1.7,2.8)@xmath801 ( 3.2,8.3)@xmath802 [ hwt change ] let @xmath428 be an @xmath756-hook partition . let @xmath427 be a borel subalgebra of @xmath719 and retain the above notation .
then , the @xmath427-highest weight of the simple @xmath719-module @xmath763 is @xmath803 let us consider an odd reflection that changes a borel subalgebra @xmath804 to @xmath805 .
assume the theorem holds for @xmath804 .
we observe by lemma [ hwt odd ] that the statement of the theorem for @xmath805 follows from the validity of the theorem for @xmath804 .
the statement of the theorem is apparently consistent with a change of borel subalgebras induced from a real reflection , and all borel subalgebras are linked by a sequence of real and odd reflections .
hence , once we know the theorem holds for one particular borel subalgebra , it holds for all .
we finally note that the theorem holds for the standard borel subalgebra @xmath806 , which corresponds to the sequence of @xmath10 @xmath734 s followed by @xmath9 @xmath736 s with all signs @xmath807 and @xmath808 being positive , i.e. , @xmath809 .
recall that the following dynkin diagram of @xmath719 and of its ( trivial ) central extension @xmath815 has been in use from the point of view of super duality and it is opposite to the one associated to the standard borel @xmath725 .
( 24,4 ) ( 5.6,2)(0,0)[c]@xmath73 ( 8,2)(0,0)[c]@xmath73 ( 10.4,2)(0,0)[c]@xmath73 ( 14.85,2)(0,0)[c]@xmath101 ( 17.25,2)(0,0)[c]@xmath73 ( 19.4,2)(0,0)[c]@xmath73 ( 23.5,2)(0,0)[c]@xmath73 ( 8.35,2)(1,0)1.5 ( 10.82,2)(1,0)0.8 ( 13.2,2)(1,0)1.2 ( 15.28,2)(1,0)1.45 ( 17.7,2)(1,0)1.25 ( 19.81,2)(1,0)0.9 ( 22,2)(1,0)1 ( 6.8,2)(0,0)[c]@xmath80 ( 12.5,1.95)(0,0)[c]@xmath74 ( 21.5,1.95)(0,0)[c]@xmath74 ( 5.4,1)(0,0)[c]@xmath82 ( 7.8,1)(0,0)[c]@xmath76 ( 10.4,1)(0,0)[c]@xmath77 ( 14.7,1)(0,0)[c]@xmath816 ( 17.15,1)(0,0)[c]@xmath109 ( 19.5,1)(0,0)[c]@xmath108 ( 23.5,1)(0,0)[c]@xmath111 ( 0,1.2)(1.8,1.4)@xmath817 [ aux : finite1 ] an irreducible integral highest weight @xmath719-module with respect to the borel subalgebra corresponding to ( 22,0)@xmath820(-22,8)@xmath821 is finite dimensional if and only if the highest weight is of the form @xmath822 where @xmath755 is an @xmath756-hook partition . let us denote the weights of the natural @xmath764-module @xmath823 by @xmath824 for @xmath825 .
the _ standard _ borel subalgebra @xmath725 of @xmath764 is the one associated to the following set of simple roots ( 24,4.5 ) ( 6,2)(0,0)[c]@xmath73 ( 8.4,2)(0,0)[c]@xmath73 ( 12.85,2)(0,0)[c]@xmath73 ( 15.25,2)(0,0)[c]@xmath101 ( 17.4,2)(0,0)[c]@xmath73 ( 22,2)(0,0)[c]@xmath73 ( 24,3.8)(0,0)[c]@xmath73 ( 24,.3)(0,0)[c]@xmath73 ( 6.4,2)(1,0)1.55 ( 8.82,2)(1,0)0.8 ( 11.2,2)(1,0)1.2 ( 13.28,2)(1,0)1.45 ( 15.7,2)(1,0)1.25 ( 17.8,2)(1,0)0.9 ( 20.1,2)(1,0)1.4 ( 22.4,2)(1,1)1.4 ( 22.4,2)(1,-1)1.4 ( 10.5,1.95)(0,0)[c]@xmath74 ( 19.6,1.95)(0,0)[c]@xmath74 ( 6.1,1)(0,0)[c]@xmath726 ( 8.9,1)(0,0)[c]@xmath727 ( 12.8,1)(0,0)[c]@xmath826 ( 15.15,3)(0,0)[c]@xmath827 ( 17.4,1)(0,0)[c]@xmath828 ( 21.3,1)(0,0)[c]@xmath829 ( 26.5,3.8)(0,0)[c]@xmath830 ( 26.5,0.3)(0,0)[c]@xmath831 we will follow the notation for @xmath719 in subsection [ sec : type b ] for sets of simple roots in terms of signed @xmath789 sequences , so we have permutations @xmath835 , and signs @xmath836 .
we fix an ambiguity on the choice of the sign @xmath837 associated to a borel @xmath427 of @xmath834-shape , by demanding the total number of negative signs among @xmath838 to be always even .
let @xmath428 be an @xmath756-hook partition , and let the block frobenius coordinates @xmath783 be as defined in subsection [ sec : type b ] .
introduce the following weights : @xmath839 the weight @xmath840 will only be used for borel @xmath427 of @xmath834-shape .
note that @xmath841 and we shall denote @xmath842 .
[ hwt : standard : spo ] @xcite the irreducible @xmath764-module of integral highest weight of the form @xmath845 with respect to the standard borel subalgebra is finite dimensional if and only if @xmath846 is either @xmath847 or @xmath848 for some @xmath756-hook partition @xmath428 . 1 .
assume @xmath427 is of @xmath834-shape .
then , * @xmath851 is the @xmath427-extremal weight for the module @xmath852 .
* @xmath840 is of the @xmath427-extremal weight for the module @xmath853 .
assume @xmath427 is of @xmath832-shape .
then , * @xmath851 is the @xmath427-extremal weight for @xmath852 if @xmath854 .
* @xmath851 is the @xmath427-extremal weight for @xmath853 if @xmath855 .
( 24,5 ) ( 6,0.3)(0,0)[c]@xmath73 ( 6,3.8)(0,0)[c]@xmath73 ( 8.4,2)(0,0)[c]@xmath73 ( 12.85,2)(0,0)[c]@xmath73 ( 15.25,2)(0,0)[c]@xmath101 ( 17.4,2)(0,0)[c]@xmath73 ( 21.9,2)(0,0)[c]@xmath73 ( 6.4,0.4)(1,1)1.6 ( 6.4,3.7)(1,-1)1.6 ( 8.82,2)(1,0)0.8 ( 11.2,2)(1,0)1.2 ( 13.28,2)(1,0)1.45 ( 15.7,2)(1,0)1.25 ( 17.8,2)(1,0)0.9 ( 20.1,2)(1,0)1.4 ( 10.5,1.95)(0,0)[c]@xmath74 ( 19.6,1.95)(0,0)[c]@xmath74 ( 3,0.3)(0,0)[c]@xmath859 ( 4.5,3.8)(0,0)[c]@xmath76 ( 8.9,1)(0,0)[c]@xmath77 ( 12.8,1)(0,0)[c]@xmath79 ( 15.15,1)(0,0)[c]@xmath816 ( 17.8,1)(0,0)[c]@xmath109 ( 22,1)(0,0)[c]@xmath111 ( 0,1.2)(1.8,1.4)@xmath858 [ aux : finite2 ] an irreducible integral highest weight @xmath764-module with respect to the borel subalgebra corresponding to ( 24,0)@xmath861(-24,8)@xmath862 is finite dimensional if and only if the highest weight is of the form @xmath863 where @xmath755 is an @xmath756-hook partition .
[ finite : dim : character ] from corollaries [ aux : finite1 ] and [ aux : finite2 ] it follows that , after passing to the central extension @xmath11 on which the center @xmath280 acts as a scalar multiplication by @xmath864 , the weights in ( [ b - type : finite : hw2 ] ) and ( [ b - type : finite : hw1 ] ) lie in @xmath371 whenever @xmath865 .
hence , theorem [ character ] and lemma [ lem : trunc ] provide a complete solution to the _ finite dimensional _ irreducible character problem for the ortho - symplectic lie superalgebras .
recall @xcite that finite dimensional irreducible _ polynomial _
@xmath866-modules are exactly the highest weight modules @xmath867 with respect to the standard borel subalgebra parametrized by @xmath756-hook partitions @xmath428 .
one can assign to any borel subalgebra @xmath417 of @xmath866 an @xmath789 sequence as in section [ sec : type b ] , but now with @xmath868 , for all @xmath869 . by the same argument as for theorem [ hwt change ]
, we can show that the highest weights of the polynomial representations of @xmath866 with respect to @xmath417 is given by @xmath870 by flipping from the left to right the dynkin diagram of ( 23,0)@xmath71(-20,8)@xmath70 and changing all the simple roots therein to their opposites , we obtain a dynkin diagram(23,0)@xmath71(-20,8)@xmath871 corresponding to the opposite borel subalgebras , where @xmath872 .
similarly , by flipping the dynkin diagrams ( 23,0)@xmath71(-20,8)@xmath97 , ( 23,0)@xmath98(-20,8)@xmath99 and ( 23,0)@xmath873(-20,8)@xmath100 and changing all signs of the simple roots for @xmath124 , we obtain the dynkin diagrams ( 23,0)@xmath71(-20,8)@xmath874 , ( 23,0)@xmath98(-20,8)@xmath875 and ( 23,0)@xmath873(-20,8)@xmath876 , respectively , of the opposite borel subalgebras .
we form the diagrams corresponding to the borel subalgebras opposite to ( [ dynkin : combined ] ) as follows : @xmath877{{\ovalbox(1.6,1.2){$^\texttt{o}\mf{t}_n$ } } } } \put(5.8,0.5){\line(1,0){1.85 } } \put(8.5,0.5){\makebox(0,0)[c]{{\ovalbox(1.6,1.2){$^\texttt{o}\mf{k}^\xx$ } } } } \put(15,0.5){\makebox(0,0)[c]{{\ovalbox(1.6,1.2){$^\texttt{o}\ov{\mf{t}}_n$ } } } } \put(15.8,0.5){\line(1,0){1.85 } } \put(18.5,0.5){\makebox(0,0)[c]{{\ovalbox(1.6,1.2){$^\texttt{o}\mf{k}^\xx$ } } } } \put(25,0.5){\makebox(0,0)[c]{{\ovalbox(1.6,1.2){$^\texttt{o}\wt{\mf{t}}_n$ } } } } \put(25.8,0.5){\line(1,0){1.85 } } \put(28.5,0.5){\makebox(0,0)[c]{{\ovalbox(1.6,1.2){$^\texttt{o}\mf{k}^\xx$ } } } } \end{picture}\ ] ] the corresponding lie superalgebras are again @xmath2 , @xmath5 and @xmath53 , respectively .
the arguments in sections [ sec : o ] and [ sec : character ] can be adapted easily to allow us to compare correspondingly defined parabolic categories @xmath878 , @xmath879 and @xmath880 using these opposite borel subalgebras , whose precise definitions are evident .
we note that for the corresponding set of weights @xmath881 of the form @xmath882 to satisfy the corresponding dominant condition we require , besides the obvious dominant condition on the standard levi subalgebra of ( 23,0)@xmath71(-20,8)@xmath871 , also that @xmath883 is a partition .
this allows us to prove an analogous version of theorem [ character ] and thus to compute irreducible characters of lie superalgebras in terms of irreducible characters of lie algebras .
also the results in section [ sec : homology ] and section [ sec : category ] have fairly straightforward analogues in @xmath878 , @xmath879 and @xmath880 as well . in particular , we can prove equivalences of the corresponding finitely generated module subcategories following the strategy of section [ sec : category ] .
besides of its own interest , another virtue of this opposite version of super duality lies in the ease of calculation of finite dimensional irreducible characters of modules over the finite dimensional ortho - symplectic lie superalgebras . as the highest weight modules over @xmath5 in this setup already have highest weights over the standard borel subalgebras , the knowledge of extremal weights for finite dimensional irreducible modules is no longer needed to imply that solution of the irreducible character problem in the category @xmath879 and @xmath884 also solves the finite dimensional irreducible character problem .
bernstein and d.a .
leites , _ a formula for the characters of the irreducible finite dimensional representations of lie superalgebras of series @xmath886 and @xmath887 _ , c.r .
sci . * 33 * ( 1980 ) , 104951 [ in russian ] .
j. germoni , _ indecomposable representations of @xmath890 , @xmath891 and @xmath892_. colloquium on homology and representation theory ( spanish ) ( vaquerias , 1998 ) .
( cordoba ) * 65 * ( 2000 ) , 147163 .
d. leites , m. saveliev and v. serganova , _ embedding of osp(n/2 ) and the associated non - linear supersymmetric equations_. group theoretical methods in physics , vol .
i ( yurmala , 1985 ) , 255297 , vnu sci . press , utrecht , 1986 . | we formulate and establish a super duality which connects parabolic categories @xmath0 for the ortho - symplectic lie superalgebras and classical lie algebras of @xmath1 types .
this provides a complete and conceptual solution of the irreducible character problem for the ortho - symplectic lie superalgebras in a parabolic category @xmath0 , which includes all finite dimensional irreducible modules , in terms of classical kazhdan - lusztig polynomials . z c n f h n
g d l i v n e + u c d a x | by 2pt |
nature provides abundant examples of rhythmic systems and synchronization phenomena @xcite .
each rhythmic system is typically described by an ordinary differential equation that possesses a limit - cycle solution .
the phase reduction method for ordinary limit - cycle oscillators has been well established and successfully applied to analyze the synchronization properties of oscillators @xcite .
there also exist rhythmic spatiotemporal patterns described by limit - cycle solutions to partial differential equations @xcite .
we recently developed a phase description method for limit - cycle solutions to the following partial differential equations : the nonlinear fokker - planck equations that represent the collective dynamics of globally coupled noisy dynamical elements @xcite , the fluid equations that represent the dynamics of the temperature field in ordinary hele - shaw cells @xcite , and the reaction - diffusion equations that represent rhythmic spatiotemporal patterns in chemical and biological systems @xcite .
however , there are also examples of spatiotemporal rhythms in systems that further possess spatial translational symmetry ; these spatiotemporal rhythms can not be described by limit - cycle solutions .
for example , rotating annuli and spheres possess continuously rotational symmetry , i.e. , continuously translational symmetry in the rotating direction @xcite .
consequently , the emergence of spatiotemporal rhythms in such systems brings up two phase modes , i.e. , a spatial phase and a temporal phase .
such spatiotemporal rhythms are described by limit - torus solutions .
synchronization of spatiotemporal rhythms with two phase modes has been experimentally investigated using systems of rotating fluid annuli that exhibit traveling and oscillating convection , which is analogous to atmospheric circulation @xcite .
therefore , a phase description method for limit - torus solutions to partial differential equations is desirable . in this paper , as the first step , we consider oscillatory convection in a cylindrical hele - shaw cell that is laterally periodic .
an ordinary hele - shaw cell is a rectangular cavity where the gap between two vertical walls is much smaller than the extent of the other two spatial dimensions , and the fluid in the cavity exhibits oscillatory convection under the appropriate parameter conditions ( see refs .
@xcite and references therein ) .
the cylindrical hele - shaw cell is a cylindrical version of the ordinary hele - shaw cell that possesses spatial translational symmetry in the lateral direction owing to the cylindrical shape .
oscillatory convection in the cylindrical hele - shaw cell is therefore described by a limit - torus solution that possesses both spatial and temporal phases . here
, we formulate a theory for the phase description of oscillatory convection in the cylindrical hele - shaw cell .
the theory can be considered as a phase reduction method for limit - torus solutions to partial differential equations .
the theory can also be considered as a generalization of our phase description method for limit - cycle solutions to partial differential equations such as the nonlinear fokker - planck equations @xcite , fluid equations @xcite , and reaction - diffusion equations @xcite .
the phase reduction method for limit - torus solutions enables us to describe the dynamics of the oscillatory convection by two phases ( i.e. , spatial and temporal phases ) , and facilitates theoretical analysis of the spatiotemporal phase synchronization properties of the oscillatory convection . on the basis of phase reduction
, we characterize the spatiotemporal phase responses of oscillatory convection to weak impulses and analyze the spatiotemporal phase synchronization between weakly coupled systems exhibiting oscillatory convection .
this paper is organized as follows . in sec .
[ sec:2 ] , we formulate a theory for the phase description of oscillatory convection with a spatially translational mode ; supplemental information of the theory is given in app .
[ sec : a ] and app .
[ sec : b ] . in sec .
[ sec:3 ] , we illustrate the theory using a numerical analysis of the oscillatory convection . in sec .
[ sec:4 ] , we make a comparison between the theory and direct numerical simulations . concluding remarks are given in sec
. [ sec:5 ] .
in this section , we formulate a theory for the phase description of oscillatory convection in a cylindrical hele - shaw cell that is laterally periodic .
the theory can be considered as an extension of our phase description method for oscillatory convection in the ordinary hele - shaw cell @xcite to that in the cylindrical hele - shaw cell .
the dynamics of the temperature field @xmath0 in the cylindrical hele - shaw cell is described by the following dimensionless form ( see ref .
@xcite and references therein ) : @xmath1 the laplacian and jacobian are respectively given by @xmath2 where we assumed that the curvature effects due to the cylindrical shape are negligible ( see refs .
@xcite for curvature effects , although the subject of these references is not thermal convection but viscous fingering ) .
the first and second terms on the right - hand side of eq .
( [ eq : t ] ) represent diffusion and advection , respectively .
the stream function @xmath3 is determined from the temperature field @xmath0 as follows : @xmath4 where the rayleigh number is denoted by @xmath5 .
the stream function @xmath3 also gives the fluid velocity field @xmath6 , i.e. , @xmath7 figure [ fig:1 ] shows a schematic diagram of the cylindrical hele - shaw cell .
the system is defined in the following rectangular region : @xmath8 $ ] and @xmath9 $ ] .
because this hele - shaw cell has a cylindrical shape , the system possesses a @xmath10-periodicity with respect to @xmath11 .
the boundary conditions for the temperature field @xmath0 are given by @xmath12 where the temperature at the bottom ( @xmath13 ) is higher than that at the top ( @xmath14 ) .
the stream function @xmath3 satisfies the periodic boundary condition on @xmath11 and the dirichlet zero boundary condition on @xmath15 , i.e. , @xmath16 owing to the homogeneity of eqs .
( [ eq : t])([eq : p_t ] ) and the periodic boundary condition on @xmath11 , given in eqs .
( [ eq : bctx])([eq : bcpx ] ) , this system possesses continuous spatial translational symmetry with respect to @xmath11 .
we also note that no conserved quantity exists in this system . to simplify the boundary conditions in eq .
( [ eq : bcty ] ) , we consider the convective component @xmath17 of the temperature field @xmath0 as follows : @xmath18 substituting eq .
( [ eq : t - x ] ) into eq . ( [ eq : t ] ) and eq .
( [ eq : p_t ] ) , we derive the following equations : @xmath19 and @xmath20 applying eq .
( [ eq : t - x ] ) to eqs .
( [ eq : bctx])([eq : bcty ] ) , we obtain the following boundary conditions for the convective component @xmath17 : @xmath21 that is , the convective component @xmath17 satisfies the periodic boundary condition on @xmath11 and the dirichlet zero boundary condition on @xmath15 . in the derivation
below , it should be noted that eq .
( [ eq : p_x ] ) can also be written in the following form : @xmath22 where the green s function @xmath23 is the solution to @xmath24 under the periodic boundary condition on @xmath11 and the dirichlet zero boundary condition on @xmath15 . from the translational symmetry with respect to @xmath11 , the green s function @xmath23 possesses the following property : @xmath25 . in the following two subsections ,
we analyze the dynamical equation ( [ eq : x ] ) with eq .
( [ eq : p_x ] ) or eq .
( [ eq : p ] ) under the boundary conditions given by eqs .
( [ eq : bcxx])([eq : bcxy ] ) and eqs .
( [ eq : bcpx])([eq : bcpy ] ) . in general , a stable limit - torus solution to eq .
( [ eq : x ] ) , which represents oscillatory convection in the cylindrical hele - shaw cell , can be described by ( e.g. , see fig .
[ fig:4 ] in sec .
[ sec:3 ] ) @xmath26 the spatial phase and traveling velocity are denoted by @xmath27 and @xmath28 , respectively ; the temporal phase and oscillation frequency are denoted by @xmath29 and @xmath30 , respectively .
the spatial and temporal phases indicate the `` position '' and `` oscillation '' of the convection , respectively .
figure [ fig:2 ] shows a schematic diagram of the limit - torus solution @xmath31 .
the limit - torus solution @xmath31 satisfies the @xmath10-periodicity with respect to @xmath27 and the @xmath32-periodicity with respect to @xmath29 , i.e. , @xmath33 substituting eq .
( [ eq : x_x0 ] ) into eqs .
( [ eq : x])([eq : p_x ] ) , we find that the limit - torus solution @xmath31 satisfies the following equation : @xmath34 x_0(x - \phi , y , \theta ) = \nabla^2 x_0 + j(\psi_0 , x_0 ) - \frac{\partial \psi_0}{\partial x } , \label{eq : x0}\end{aligned}\ ] ] where the stream function @xmath35 is determined by ( e.g. , see fig .
[ fig:5 ] in sec .
[ sec:3 ] ) @xmath36 from eq .
( [ eq : t - x ] ) , the corresponding temperature field @xmath37 is given by ( e.g. , see fig .
[ fig:5 ] in sec .
[ sec:3 ] ) @xmath38 let @xmath39 represent a small disturbance to the limit - torus solution @xmath31 , and consider a slightly perturbed solution @xmath40 equation ( [ eq : x ] ) is then linearized with respect to @xmath39 as follows : @xmath41 here , the linear operator @xmath42 is explicitly given by @xmath43 u(x - \phi , y , \theta ) , \label{eq : call}\end{aligned}\ ] ] where @xmath44 similarly to the limit - torus solution @xmath31 , the function @xmath45 satisfies the periodic boundary condition on @xmath11 , the dirichlet zero boundary condition on @xmath15 , and the @xmath32-periodicity with respect to @xmath29 . in eq .
( [ eq : l ] ) , the function @xmath46 is the solution to @xmath47 under the periodic boundary condition on @xmath11 and the dirichlet zero boundary condition on @xmath15 .
note that the linear operator @xmath42 is periodic with respect to both @xmath27 and @xmath29 .
therefore , eq . ( [ eq : linear ] ) is a floquet - type system with two zero - eigenvalues ; one is associated with spatial translational symmetry breaking and the other is associated with temporal translational symmetry breaking .
defining the inner product of two functions as @xmath48\!\!\bigr ] } } = \frac{1}{2\pi } \int_0^{2\pi } d\theta \int_0 ^ 2 dx \int_0 ^ 1 dy \ , u^\ast(x - \phi , y , \theta ) u(x - \phi , y , \theta ) , \label{eq : inner}\end{aligned}\ ] ] we introduce the adjoint operator of the linear operator @xmath42 by @xmath49\!\!\bigr ] } } = { \ensuremath{\bigl[\!\!\bigl [ { \cal l}^\ast(x - \phi , y , \theta ) u^\ast(x - \phi , y , \theta ) , \ , u(x - \phi , y , \theta ) \bigr]\!\!\bigr]}}. \label{eq : operator}\end{aligned}\ ] ] by partial integration , the adjoint operator @xmath50 is explicitly given by @xmath51 u^\ast(x - \phi , y , \theta ) , \label{eq : callast}\end{aligned}\ ] ] where @xmath52 - \frac{\partial}{\partial y } \left [ u^\ast \frac{\partial \psi_0}{\partial x } \right ] + \frac{\partial}{\partial x } \bigl [ \psi_{u , x}^\ast - \psi_{u , y}^\ast \bigr ] .
\label{eq : last}\end{aligned}\ ] ] similarly to @xmath45 , the function @xmath53 also satisfies the periodic boundary condition on @xmath11 , the dirichlet zero boundary condition on @xmath15 , and the @xmath32-periodicity with respect to @xmath29 . in eq .
( [ eq : last ] ) , the two functions , i.e. , @xmath54 and @xmath55 , are the solutions to @xmath56 , \label{eq : past_uast_x } \\ \nabla^2 \psi_{u , y}^\ast(x - \phi , y , \theta ) & = -{\rm ra } \frac{\partial}{\partial y } \left [ u^\ast \frac{\partial x_0}{\partial x } \right ] , \label{eq : past_uast_y}\end{aligned}\ ] ] under the periodic boundary condition on @xmath11 and the dirichlet zero boundary condition on @xmath15 , respectively .
details of the derivation of the adjoint operator @xmath50 are given in app .
[ sec : a ] . in the calculation
below , we utilize the floquet eigenfunctions associated with the two zero - eigenvalues , i.e. , @xmath57 u_{\rm s}(x - \phi , y , \theta ) & = 0 , \\ { \cal l}(x - \phi , y , \theta ) u_{\rm t}(x - \phi , y , \theta ) = \left [ l(x - \phi , y , \theta ) + c \frac{\partial}{\partial x } - \omega \frac{\partial}{\partial \theta } \right ] u_{\rm t}(x - \phi , y , \theta ) & = 0 , \\ { \cal l}^\ast(x - \phi , y , \theta ) u_{\rm s}^\ast(x - \phi , y , \theta ) = \left [ l^\ast(x - \phi , y , \theta ) - c \frac{\partial}{\partial x } + \omega \frac{\partial}{\partial \theta } \right ] u_{\rm s}^\ast(x - \phi , y , \theta ) & = 0 , \label{eq : usast } \\ { \cal l}^\ast(x - \phi , y , \theta ) u_{\rm t}^\ast(x - \phi , y , \theta ) = \left [ l^\ast(x - \phi , y , \theta ) - c \frac{\partial}{\partial x } + \omega \frac{\partial}{\partial \theta } \right ] u_{\rm t}^\ast(x - \phi , y , \theta ) & = 0 .
\label{eq : utast}\end{aligned}\ ] ] note that the two right zero eigenfunctions , i.e. , @xmath58 for the spatial phase @xmath27 and @xmath59 for the temporal phase @xmath29 , can be chosen as ( e.g. , see fig .
[ fig:7 ] in sec .
[ sec:3 ] ) @xmath60 which are confirmed by differentiating eq .
( [ eq : x0 ] ) with respect to @xmath27 and @xmath29 , respectively . for the inner product ( [ eq : inner ] ) with the two right zero eigenfunctions ( [ eq : us])([eq : ut ] ) , the corresponding two left zero eigenfunctions , i.e. , @xmath61 and @xmath62 , are orthonormalized as @xmath63\!\!\bigr ] } } = \frac{1}{2\pi } \int_0^{2\pi } d\theta \int_0 ^ 2 dx \int_0 ^ 1 dy \ , u_p^\ast(x - \phi , y , \theta ) u_q(x - \phi , y , \theta ) = \delta_{pq},\end{aligned}\ ] ] for @xmath64 .
here , we note that the following equation holds ( see also refs .
@xcite ) : @xmath65 \nonumber \\ & = \int_0 ^ 2 dx \int_0 ^ 1 dy \ , \left [ u_p^\ast(x - \phi , y , \theta ) \frac{\partial}{\partial \theta } u_q(x - \phi , y , \theta ) + u_q(x - \phi , y , \theta ) \frac{\partial}{\partial \theta } u_p^\ast(x - \phi , y , \theta ) \right ] \nonumber \\ & = \frac{1}{\omega } \int_0 ^ 2 dx \int_0 ^ 1 dy \ , \biggl [ u_p^\ast(x - \phi , y , \theta ) \left\ { l(x - \phi , y , \theta ) + c \frac{\partial}{\partial x } \right\ } u_q(x - \phi , y , \theta ) \nonumber \\ & \qquad\qquad\qquad\qquad\quad - u_q(x - \phi , y , \theta ) \left\ { l^\ast(x - \phi , y , \theta ) - c \frac{\partial}{\partial x } \right\ } u_p^\ast(x - \phi , y , \theta ) \biggr ] \nonumber \\ & = 0.\end{aligned}\ ] ] therefore , the following orthonormalization condition is independently satisfied for each @xmath29 : @xmath66 here , we describe a numerical method for obtaining the left zero eigenfunctions . from eqs .
( [ eq : usast])([eq : utast ] ) , the left zero eigenfunctions , @xmath61 and @xmath62 , satisfy @xmath67 u_p^\ast(x - \phi , y , \theta),\end{aligned}\ ] ] for @xmath68 , which can be transformed into @xmath69 u_p^\ast(x - \phi , y , -\omega s ) , \label{eq : adjoint}\end{aligned}\ ] ] by substituting @xmath70 . a relaxation method using eq .
( [ eq : adjoint ] ) , which can also be called the adjoint method ( see refs .
@xcite for limit - cycle solutions to ordinary differential equations and refs .
@xcite for limit - cycle solutions to partial differential equations ) , is convenient to obtain the left zero eigenfunctions for the limit - torus solution . in the following two subsections ,
we derive a set of phase equations for oscillatory convection in the cylindrical hele - shaw cell using the limit - torus solution and its floquet zero eigenfunctions . in this subsection , we consider oscillatory cylindrical - hele - shaw convection with a weak perturbation to the temperature field @xmath0 described by the following equation : @xmath71 the weak perturbation is denoted by @xmath72 . substituting eq .
( [ eq : t - x ] ) into eq .
( [ eq : tp ] ) , we obtain the following equation for the convective component @xmath17 : @xmath73 using the idea of phase reduction @xcite , we derive a set of phase equations from the perturbed equation ( [ eq : xp ] ) .
that is , using the left zero eigenfunctions , i.e. , @xmath61 and @xmath62 , we project the dynamics of the perturbed equation ( [ eq : xp ] ) onto the unperturbed limit - torus solution with respect to the spatial and temporal phases as follows : @xmath74 \nonumber \\ & = \int_0 ^ 2 dx \int_0 ^ 1 dy \ , u_{\rm s}^\ast(x - \phi , y , \theta ) \left [ \nabla^2 x + j(\psi , x ) - \frac{\partial \psi}{\partial x } + \epsilon p(x , y , t ) \right ] \nonumber \\ & \simeq \int_0 ^ 2 dx \int_0 ^ 1 dy \ , u_{\rm s}^\ast(x - \phi , y , \theta ) \left [ \nabla^2 x_0 + j(\psi_0 , x_0 ) - \frac{\partial \psi_0}{\partial x } + \epsilon p(x , y , t ) \right ] \nonumber \\ & = \int_0 ^ 2 dx \int_0 ^ 1 dy \ , u_{\rm s}^\ast(x - \phi , y , \theta ) \left [ - c \frac{\partial x_0}{\partial x } + \omega \frac{\partial x_0}{\partial \theta } + \epsilon p(x , y , t ) \right ] \nonumber \\ & = \int_0 ^ 2 dx \int_0 ^ 1 dy \ , u_{\rm s}^\ast(x - \phi , y , \theta ) \biggl [ - c u_{\rm s}(x - \phi , y , \theta ) + \omega u_{\rm t}(x - \phi , y , \theta ) + \epsilon p(x , y , t ) \biggr ] \nonumber \\ & = -c + \epsilon \int_0 ^ 2 dx \int_0 ^ 1 dy \ , u_{\rm s}^\ast(x - \phi , y , \theta ) p(x , y , t ) , \label{eq : project_phi}\end{aligned}\ ] ] and @xmath75 \nonumber \\ & = \int_0 ^ 2 dx \int_0 ^ 1 dy \ , u_{\rm t}^\ast(x - \phi , y , \theta ) \left [ \nabla^2 x + j(\psi , x ) - \frac{\partial \psi}{\partial x } + \epsilon p(x , y , t ) \right ] \nonumber \\ & \simeq \int_0 ^ 2 dx \int_0 ^ 1 dy \ , u_{\rm t}^\ast(x - \phi , y , \theta ) \left [ \nabla^2 x_0 + j(\psi_0 , x_0 ) - \frac{\partial \psi_0}{\partial x } + \epsilon p(x , y , t ) \right ] \nonumber \\ & = \int_0 ^ 2 dx \int_0 ^ 1 dy \ , u_{\rm t}^\ast(x - \phi , y , \theta ) \left [ - c \frac{\partial x_0}{\partial x } + \omega \frac{\partial x_0}{\partial \theta } + \epsilon p(x , y , t ) \right ] \nonumber \\ & = \int_0 ^ 2 dx \int_0 ^ 1 dy \ , u_{\rm t}^\ast(x - \phi , y , \theta ) \biggl [ - c u_{\rm s}(x - \phi , y , \theta ) + \omega u_{\rm t}(x - \phi , y , \theta ) + \epsilon p(x , y , t ) \biggr ] \nonumber \\ & = \omega + \epsilon \int_0 ^ 2 dx \int_0 ^ 1 dy \ , u_{\rm t}^\ast(x - \phi , y , \theta ) p(x , y , t ) , \label{eq : project_theta}\end{aligned}\ ] ] where we approximated @xmath17 by the unperturbed limit - torus solution @xmath31 in eqs . ( [ eq : project_phi])([eq : project_theta ] ) . therefore , the two phase equations describing the oscillatory cylindrical - hele - shaw convection with weak perturbation are approximately obtained in the following forms : @xmath76 where the _ phase sensitivity functions _ for the spatial and temporal phases are defined as ( e.g. , see fig .
[ fig:8 ] in sec . [ sec:3 ] )
@xmath77 the phase equations ( [ eq : phi_p])([eq : theta_p ] ) are the main results of this paper .
these equations can also be considered as an extension of that describing oscillatory convection in the ordinary hele - shaw cell @xcite . as found from eqs .
( [ eq : phi_p])([eq : theta_p ] ) , the spatial and temporal phases are coupled ; therefore , nontrivial spatiotemporal phase dynamics are revealed .
furthermore , we consider the case of the perturbation described by a product of two functions as follows : @xmath78 that is , the space - dependence and time - dependence of the perturbation are separated . in this case , the phase equations ( [ eq : phi_p])([eq : theta_p ] ) are written in the following forms : @xmath79 where the _ effective phase sensitivity functions _ for the spatial and temporal phases are given by ( e.g. , see fig .
[ fig:10 ] in sec .
[ sec:3 ] ) @xmath80 we note that the forms of eqs .
( [ eq : phi_q])([eq : theta_q ] ) are essentially the same as those of the phase equations which are derived for a perturbed limit - torus oscillator described by a finite - dimensional dynamical system ( see refs .
@xcite ) .
finally , it should be noted that we can also consider oscillatory cylindrical - hele - shaw convection with weak boundary forcing as mentioned in app .
[ sec : b ] . in this subsection
, we consider weakly coupled systems of oscillatory cylindrical - hele - shaw convection described by the following equation : @xmath81 for @xmath82 or @xmath83 .
two identical systems of oscillatory cylindrical - hele - shaw convection are mutually coupled through corresponding temperatures at each spatial point , where the coupling parameter is denoted by @xmath84 .
substituting eq .
( [ eq : t - x ] ) into eq .
( [ eq : tsigma ] ) for each @xmath85 , we obtain the following equation for the convective component @xmath86 : @xmath87 as in eq .
( [ eq : p_x ] ) , the stream function @xmath88 of each system is determined by @xmath89 here , we assume that unperturbed oscillatory cylindrical - hele - shaw convection is a stable limit - torus solution and that the coupling between the systems of oscillatory convection is sufficiently weak . under this assumption , as in the preceding subsection , we obtain a set of phase equations from eq .
( [ eq : xsigma ] ) as follows : @xmath90 where @xmath91 , \\ \tilde{\gamma}_{\rm t}\left ( \phi_\sigma - \phi_\tau , \theta_\sigma , \theta_\tau \right ) & = \int_0 ^ 2 dx \int_0 ^ 1 dy \ , z_{\rm t}(x - \phi_\sigma , y , \theta_\sigma ) \bigl [ x_0(x - \phi_\tau , y , \theta_\tau ) - x_0(x - \phi_\sigma , y , \theta_\sigma ) \bigr].\end{aligned}\ ] ] these two functions , i.e. , @xmath92 and @xmath93 , depend on the spatial phase difference between the systems and the temporal phases of both systems . introducing the slow phase variables as @xmath94
we rewrite eqs .
( [ eq : phi_tildegamma])([eq : theta_tildegamma ] ) as @xmath95 by applying the averaging method with respect to the temporal phases , eqs .
( [ eq : phi_tildegamma])([eq : theta_tildegamma ] ) are written in the following forms : @xmath96 where the _ phase coupling functions _ for the spatial and temporal phases are given by ( e.g. , see fig .
[ fig:12 ] in sec .
[ sec:3 ] ) @xmath97 therefore , we obtain the following phase equations : @xmath98 where the phase coupling functions are explicitly described as @xmath99 , \label{eq : gamma_s}\\[3 mm ] & \gamma_{\rm t}\left ( \phi_\sigma - \phi_\tau , \theta_\sigma - \theta_\tau \right ) \nonumber \\ & = \frac{1}{2\pi } \int_0^{2\pi } d\lambda \int_0 ^ 2 dx \int_0 ^ 1 dy \ , z_{\rm t}(x - \phi_\sigma , y , \lambda + \theta_\sigma ) \bigl [ x_0(x - \phi_\tau , y , \lambda + \theta_\tau ) - x_0(x - \phi_\sigma , y , \lambda + \theta_\sigma ) \bigr ] .
\label{eq : gamma_t}\end{aligned}\ ] ] the phase coupling functions for the spatial and temporal phases , i.e. , @xmath100 and @xmath101 , depend only on the spatial and temporal phase differences .
let the spatial and temporal phase differences respectively be defined as @xmath102 from eqs .
( [ eq : phi_gamma])([eq : theta_gamma ] ) , we obtain the following equations by subtraction : @xmath103 where @xmath104 these two functions , i.e. , @xmath105 and @xmath106 , satisfy the following properties : @xmath107 which represent the anti - symmetry with respect to the origin , i.e. , @xmath108 .
finally , we note that the forms of eqs .
( [ eq : phi_gamma])([eq : theta_gamma ] ) are the same as that of the phase equations which are derived from weakly coupled limit - torus oscillators described by finite - dimensional dynamical systems ( see refs .
that is , a system of oscillatory convection with a spatially translational mode can be reduced to a set of phase equations , similarly to an ordinary limit - torus oscillator .
in this section , we illustrate the theory developed in sec .
[ sec:2 ] using a numerical analysis of the oscillatory convection in the cylindrical hele - shaw cell . for numerical simulations using the pseudospectral method performed in sec .
[ subsec:3b ] , we first describe a spectral transformation . considering the boundary conditions for @xmath17 , given in eqs .
( [ eq : bcxx])([eq : bcxy ] ) , we introduce the following spectral decomposition : @xmath109 where the spectral transformation of @xmath17 is given by @xmath110 similarly , the limit - torus solution @xmath31 introduced in eq .
( [ eq : x_x0 ] ) is decomposed as @xmath111 the corresponding spectral transformation of the limit - torus solution is described by @xmath112 where we assign the origin of the spatial phase , i.e. , @xmath113 , to the spatial pattern @xmath114 that satisfies the following property : @xmath115 which is unique when a pair of vortices exist in the system , as considered below ( e.g. , see fig .
[ fig:4 ] and fig .
[ fig:5 ] in sec .
[ subsec:3b ] ) . when visualizing the temporal phase of the limit - torus in the infinite - dimensional state space , we project the limit - torus solution @xmath31 onto the @xmath116-@xmath117 plane as @xmath118 which are real numbers and depend only on the temporal phase @xmath29 . when determining the spatial phase of the limit - torus solution , we introduce the following complex order parameter : @xmath119 which corresponds to eq .
( [ eq : hjk ] ) with @xmath120 and @xmath121 . from eq .
( [ eq : phi_zero ] ) , the spatial phase @xmath27 is determined by @xmath122 as in @xmath17 , considering the boundary conditions for @xmath123 and @xmath124 , i.e. , the periodic boundary condition on @xmath11 and the dirichlet zero boundary condition on @xmath15 as described in eqs .
( [ eq : bcuastx])([eq : bcuasty ] ) , we also introduce the following spectral decompositions : @xmath125 where the spectral transformations are given by @xmath126 the spatial power spectra of the phase sensitivity functions averaged over the temporal phase are defined as @xmath127 we first summarize our numerical simulations in this paper . as mentioned in the preceding subsection , we applied the pseudospectral method , which is composed of a fourier expansion with @xmath128 modes for the periodic boundary condition on @xmath11 and a sine expansion with @xmath129 modes for the dirichlet zero boundary condition on @xmath15 .
the initial values were prepared such that the system exhibits oscillatory convection with a pair of vortices . because the rayleigh number was fixed at @xmath130 , the traveling velocity and oscillation frequency were @xmath131 and @xmath132 , respectively .
as mentioned below , this limit - torus solution possesses reflection symmetry because the traveling velocity is exactly zero , i.e. , @xmath131 . here
, we note that the theory developed in sec .
[ sec:2 ] is applicable for the case of non - zero traveling velocity .
figure [ fig:3 ] shows the limit - torus orbit projected onto the @xmath116-@xmath117 plane , which was obtained from our numerical simulations of the dynamical equation ( [ eq : x ] ) .
snapshots of the limit - torus solution @xmath31 are shown in fig .
[ fig:4 ] .
in addition , several quantities associated with the limit - torus solution are shown as follows : snapshots of the temperature field @xmath133 and the stream function @xmath35 are shown in fig .
[ fig:5 ] ; snapshots of the fluid velocity @xmath134 are shown in fig .
[ fig:6 ] ; snapshots of the right zero eigenfunctions , @xmath58 and @xmath59 , are shown in fig .
[ fig:7 ] ; snapshots of the phase sensitivity functions , @xmath123 and @xmath124 , are shown in fig .
[ fig:8 ] .
we note that the phase sensitivity functions , @xmath123 and @xmath124 , were obtained using the adjoint method , i.e. , the relaxation method for eq .
( [ eq : adjoint ] ) with the orthonormalization condition given by eq .
( [ eq : orthonormal ] ) . as seen in fig .
[ fig:5 ] , the spatial phase @xmath27 can be considered as the `` position '' of the hot plume in the convection , whereas the temporal phase @xmath29 represents the `` oscillation '' of the convection .
we also note that the phase sensitivity functions , @xmath123 and @xmath124 , are spatially localized as seen in fig .
[ fig:8 ] .
namely , when the spatial phase is @xmath135 , the amplitudes of @xmath123 and @xmath124 with respect to the temporal phase @xmath29 in the bottom - left , bottom - right , and top - center regions of the system are much larger than those in the other regions ; this fact reflects the dynamics of the spatial pattern of the convective component @xmath31 shown in fig .
[ fig:4 ] . furthermore ,
in this case , the limit - torus solution @xmath31 and the phase sensitivity functions , @xmath123 and @xmath124 , possess line symmetry .
as mentioned above , in this case , the traveling velocity is zero , i.e. , @xmath136 the limit - torus solution representing the oscillatory convection with eq .
( [ eq : c_zero ] ) possesses the following reflection symmetry ( see fig .
[ fig:4 ] ) : @xmath137 from eq .
( [ eq : us ] ) and eq .
( [ eq : ut ] ) , the right zero eigenfunctions , @xmath58 and @xmath59 , respectively possess the following reflection anti - symmetry and reflection symmetry ( see fig .
[ fig:7 ] ) : @xmath138 therefore , the phase sensitivity functions , or the left zero eigenfunctions , also possess the following properties ( see fig .
[ fig:8 ] ) : @xmath139 moreover , in this case , the limit - torus solution @xmath31 and the phase sensitivity functions , @xmath123 and @xmath124 , also possess point symmetry . for each @xmath29 ,
the limit - torus solution and phase sensitivity functions possess the following properties with respect to a certain point of the system ( see fig . [ fig:4 ] and fig . [ fig:8 ] ) : @xmath140 where @xmath141 and @xmath142 . in this subsection
, we calculate the effective phase sensitivity functions obtained in sec .
[ subsec:2e ] . before illustrating the effective phase sensitivity functions , the spatial power spectra of the phase sensitivity functions averaged over the temporal phase as defined in eq .
( [ eq : ps ] ) and eq .
( [ eq : pt ] ) , i.e. , @xmath143 and @xmath144 , are shown in fig .
[ fig:9](a ) and fig .
[ fig:9](b ) , respectively . owing to the point - symmetry of the phase sensitivity functions , given in eqs .
( [ eq : zs_ps])([eq : zt_ps ] ) , both @xmath143 and @xmath144 exhibit checkerboard patterns .
the power of the mode @xmath145 is the largest for @xmath143 and @xmath144 . to illustrate the effective phase sensitivity functions
, we consider a corresponding spatial pattern @xmath146 where @xmath147 and @xmath148 .
figure [ fig:9](c ) shows the spatial pattern @xmath149 , for which the effective phase sensitivity functions are shown in fig .
[ fig:10 ] with respect to @xmath27 and @xmath29 .
in addition , the effective phase sensitivity functions with @xmath150 are shown in fig .
[ fig:11 ] as a function of @xmath29 . as seen in fig .
[ fig:10 ] and fig .
[ fig:11 ] , the effective phase sensitivity functions , @xmath151 and @xmath152 , exhibit both positive and negative values .
namely , when the effective phase sensitivity function for the spatial phase , @xmath151 , is positive ( negative ) , the spatial phase @xmath27 is advanced ( delayed ) by applying a positive perturbation . similarly , when the effective phase sensitivity function for the temporal phase , @xmath152 , is positive ( negative ) , the temporal phase @xmath29 is advanced ( delayed ) by applying a positive perturbation . in this subsection , we calculate the phase coupling functions obtained in sec .
[ subsec:2f ] .
the anti - symmetric components of the phase coupling functions , @xmath105 and @xmath106 , are shown in fig .
[ fig:12](a ) and fig .
[ fig:12](b ) , respectively . as shown in eqs .
( [ eq : gsa_as])([eq : gta_as ] ) , both @xmath105 and @xmath106 possess anti - symmetry with respect to the origin , i.e. , @xmath108 . here
, we describe another symmetry of the anti - symmetric components of the phase coupling functions in this reflection symmetric case .
the phase coupling functions given by eqs .
( [ eq : gamma_s])([eq : gamma_t ] ) are written in the following forms : @xmath153 , \\
\gamma_{\rm t}(\vardelta\phi , \vardelta\theta ) & = \frac{1}{2\pi } \int_0^{2\pi } d\lambda \int_0 ^ 2 dx \int_0 ^ 1 dy \ , z_{\rm t}(x - \vardelta\phi , y , \lambda + \vardelta\theta ) \bigl
[ x_0(x , y , \lambda ) - x_0(x - \vardelta\phi , y , \lambda + \vardelta\theta ) \bigr].\end{aligned}\ ] ] from eqs . (
[ eq : x0_rs])([eq : zs_rs])([eq : zt_rs ] ) , the phase coupling functions , @xmath154 and @xmath155 , respectively possess the following reflection anti - symmetry and reflection symmetry : @xmath156 therefore , the anti - symmetric components of the phase coupling functions , @xmath105 and @xmath106 , also possess the following properties ( see fig .
[ fig:12 ] ) : @xmath157 in addition , considering the symmetries given by eqs .
( [ eq : gsa_as])([eq : gta_as ] ) and eqs .
( [ eq : gsa_rs])([eq : gta_rs ] ) , we also obtain the following properties ( see fig . [ fig:12 ] ) : @xmath158 from the above symmetries , in this case , it is sufficient to investigate only the following region : @xmath159 \times [ 0 , 1]$ ] .
figure [ fig:12](c ) shows the nullclines , fixed points , and typical orbits of the phase differences in this region .
the typical orbits were obtained from the numerical simulations of eqs .
( [ eq : dphi])([eq : dtheta ] ) . as seen in fig .
[ fig:12](c ) , the spatial and temporal in - phase state , i.e. , @xmath160 , is globally stable under the phase reduction approximation .
the spatial phase difference @xmath161 monotonously decreases to zero , whereas the temporal phase difference @xmath162 first oscillates and then becomes zero .
in this section , we compare the theoretical values obtained in sec .
[ sec:3 ] with direct numerical simulations of oscillatory convection . in this subsection
, we make a comparison of the effective phase sensitivity functions between the theoretical values obtained in sec .
[ subsec:3c ] , i.e. , fig .
[ fig:11 ] , and direct numerical simulations of oscillatory convection with weak impulses described by eq .
( [ eq : xp ] ) using eqs .
( [ eq : p_aq])([eq : a_cs ] ) .
the comparison of the effective phase sensitivity functions , @xmath163 and @xmath164 , between the direct numerical simulations with impulse intensity @xmath84 and the theoretical curves are shown in fig .
[ fig:13 ] .
the simulation results agree quantitatively with the theory . in this subsection
, we make a comparison on time evolution of phase differences between the theoretical values obtained in sec .
[ subsec:3d ] , i.e. , fig .
[ fig:12](c ) , and direct numerical simulations of two weakly coupled systems of oscillatory convection described by eq .
( [ eq : xsigma ] ) .
the comparison of the time evolution of the phase differences between the direct numerical simulations with coupling intensity @xmath84 and the theoretical curves are shown in fig .
[ fig:14 ] .
the simulation results agree quantitatively with the theory .
our investigations in this paper are summarized as follows . in sec .
[ sec:2 ] , we formulated the theory for the phase description of oscillatory convection with a spatially translational mode .
in particular , we derived the phase sensitivity functions for the spatial and temporal phases .
details of the derivation of the adjoint operator , which provides the phase sensitivity functions , are given in app .
[ sec : a ] .
treatments of the boundary forcing are described in app .
[ sec : b ] .
we also derived the phase coupling functions for the spatial and temporal phases . in sec .
[ sec:3 ] , we illustrated the theory using the numerical analysis of the oscillatory convection .
in particular , we obtained the phase sensitivity functions and phase coupling functions . in sec .
[ sec:4 ] , we made comparisons between the theory and direct numerical simulations : the spatiotemporal phase responses of oscillatory convection to weak impulses and the spatiotemporal phase synchronization between two weakly coupled systems of oscillatory convection .
the theoretical predictions were successfully confirmed by the direct numerical simulations . here , we give some concluding remarks .
first , we summarize three types of solutions to partial differential equations for the field @xmath165 and the corresponding phase description methods considered so far . ( a ) a traveling solution : @xmath166 with @xmath167 .
( b ) an oscillating solution : @xmath168 with @xmath169 .
( c ) a traveling and oscillating solution : @xmath170 with @xmath167 and @xmath169 .
the phase description method for type ( a ) has been developed in refs .
@xcite ( see also refs .
@xcite ) ; this method is closely related to the phase reduction approach to spatially periodic patterns in that the phase is associated with spatial translational symmetry breaking @xcite ( see also refs .
@xcite ) . as mentioned in refs .
@xcite , type ( a ) can be considered as a special case of type ( b ) .
type ( b ) can be considered as a limit - cycle solution and possesses only one phase variable , whereas type ( c ) can be considered as a limit - torus solution and possesses two phase variables .
the phase description method developed in this paper belongs to type ( c ) , and it can be considered as a generalization of our phase description method for type ( b ) developed in refs .
@xcite .
second , we note the spatial reflection symmetry of spatial patterns .
consider the pattern formation in a system of spatial reflection symmetry : when a spatial pattern does not break the reflection symmetry , the traveling velocity is zero , @xmath131 ; meanwhile , when a spatial pattern does break the reflection symmetry , the traveling velocity becomes non - zero , @xmath171 .
the traveling velocity of the oscillatory cylindrical - hele - shaw convection is zero ; however , the phase description method itself is applicable for the case of non - zero traveling velocity . if the reflection symmetry of limit - torus solution , i.e. , eq .
( [ eq : x0_rs ] ) , is lost , then those of the phase sensitivity functions , i.e. , eqs .
( [ eq : zs_rs])([eq : zt_rs ] ) , are lost , and those of the phase coupling functions , i.e. , eqs .
( [ eq : gs_rs])([eq : gt_rs ] ) and eqs .
( [ eq : gsa_rs])([eq : gta_rs ] ) , are also lost .
third , we note ubiquitousness of the limit - torus solutions to partial differential equations .
a limit - torus solution to an ordinary differential equation represents a quasi - periodic oscillator , and a phase description method for the quasi - periodic oscillator has also been developed @xcite
. however , there have been few studies on the synchronization of quasi - periodic oscillators ; this may be due to the fact that limit - cycle or chaotic oscillations are ubiquitous , but quasi - periodic oscillations are rather rare in ordinary differential equations . in contrast , a limit - torus solution is ubiquitous in a partial differential equation that possesses some spatial translational symmetry . from this point of view , a systematic analysis of a set of phase equations , such as eqs .
( [ eq : phi_q])([eq : theta_q ] ) , eqs .
( [ eq : phi_gamma])([eq : theta_gamma ] ) , and eqs .
( [ eq : dphi])([eq : dtheta ] ) , is meaningful and important .
it should also be noted that these phase equations are universal and invariant for limit - torus solutions .
finally , we remark the broad applicability of our phase description approach , which is not restricted to oscillatory cylindrical - hele - shaw convection .
the phase description method can be generalized to traveling and oscillating localized convection ( i.e. , a traveling breather ) in a binary fluid system ( see , e.g. , ref .
@xcite ) or traveling and oscillating convection in a rotating fluid annulus system ( see , e.g. , ref .
similar phase description methods can also be developed for oscillating spots of zero traveling velocity ( see , e.g. , ref .
@xcite ) or traveling breathers of non - zero traveling velocity ( see , e.g. , ref .
@xcite ) in reaction - diffusion systems . y.k . is grateful to members of both the earth evolution modeling research team and the nonlinear dynamics and its application research team at ifree / jamstec for fruitful comments .
is also grateful for financial support by jsps kakenhi grant number 25800222 .
h.n . is grateful for financial support by jsps kakenhi grant numbers 25540108 and 22684020 , crest kokubu project of jst , and first aihara project of jsps .
in this appendix , we describe the details of the derivation of the adjoint operator @xmath50 given in eqs .
( [ eq : callast])([eq : last ] ) ( see also , e.g. , refs .
@xcite for mathematical terms ) .
the derivation procedure is similar to that performed in ref .
@xcite . from eqs .
( [ eq : call])([eq : l ] ) , the linear operator @xmath42 is given by the following form : @xmath172 by partial integration , each term of the inner product @xmath173\!]}}$ ] is transformed into @xmath174\!\!\biggr ] } } & = \frac{1}{2\pi } \int_0^{2\pi } d\theta \int_0 ^ 1 dy \ , \left\ { \left [ u^\ast \ , \frac{\partial u}{\partial x } \right]_{x=0}^{x=2 } - \left [ \frac{\partial u^\ast}{\partial x } \ , u \right]_{x=0}^{x=2 } \right\ } + { \ensuremath{\biggl[\!\!\biggl [ \frac{\partial^2 u^\ast}{\partial x^2 } , u \biggr]\!\!\biggr ] } } , \\ { \ensuremath{\biggl[\!\!\biggl [ u^\ast , \frac{\partial^2 u}{\partial y^2 } \biggr]\!\!\biggr ] } } & = \frac{1}{2\pi } \int_0^{2\pi } d\theta \int_0 ^ 2 dx \ ,
\left\ { \left [ u^\ast \ , \frac{\partial u}{\partial y } \right]_{y=0}^{y=1 } - \left [ \frac{\partial u^\ast}{\partial y } \ , u \right]_{y=0}^{y=1 } \right\ } + { \ensuremath{\biggl[\!\!\biggl [ \frac{\partial^2 u^\ast}{\partial y^2 } , u \biggr]\!\!\biggr ] } } , \\ { \ensuremath{\biggl[\!\!\biggl [ u^\ast , - \frac{\partial \psi_0}{\partial y } \frac{\partial u}{\partial x } \biggr]\!\!\biggr ] } } & = -\frac{1}{2\pi } \int_0^{2\pi } d\theta \int_0 ^ 1 dy \ , \left [ u^\ast \ , \frac{\partial \psi_0}{\partial y } \ , u \right]_{x=0}^{x=2 } + { \ensuremath{\biggl[\!\!\biggl [ \frac{\partial}{\partial x } \left [ u^\ast \frac{\partial \psi_0}{\partial y } \right ] , u \biggr]\!\!\biggr ] } } , \\ { \ensuremath{\biggl[\!\!\biggl [ u^\ast , \frac{\partial \psi_0}{\partial x } \frac{\partial u}{\partial y } \biggr]\!\!\biggr ] } } & = \frac{1}{2\pi } \int_0^{2\pi } d\theta \int_0 ^ 2 dx \ , \left [ u^\ast \ , \frac{\partial \psi_0}{\partial x } \ , u \right]_{y=0}^{y=1 } + { \ensuremath{\biggl[\!\!\biggl [ -\frac{\partial}{\partial y } \left [ u^\ast \frac{\partial \psi_0}{\partial y } \right ] , u \biggr]\!\!\biggr ] } } , \\ { \ensuremath{\biggl[\!\!\biggl [ u^\ast , \frac{\partial \psi_u}{\partial x } \left ( \frac{\partial x_0}{\partial y } - 1 \right ) \biggr]\!\!\biggr ] } } & = \frac{1}{2\pi } \int_0^{2\pi } d\theta \int_0 ^ 1 dy \ , \left [ u^\ast \ , \left ( \frac{\partial x_0}{\partial y } - 1 \right ) \
, \psi_u \right]_{x=0}^{x=2 } + { \ensuremath{\biggl[\!\!\biggl [ -\frac{\partial}{\partial x } \left [ u^\ast \left ( \frac{\partial x_0}{\partial y } - 1 \right ) \right ] , \psi_u \biggr]\!\!\biggr ] } } , \label{eq : pd_p_u_1 } \\ { \ensuremath{\biggl[\!\!\biggl [ u^\ast , -\frac{\partial \psi_u}{\partial y } \frac{\partial x_0}{\partial x } \biggr]\!\!\biggr ] } } & = -\frac{1}{2\pi } \int_0^{2\pi } d\theta \int_0 ^ 2 dx \ , \left [ u^\ast \ , \frac{\partial x_0}{\partial x } \
, \psi_u \right]_{y=0}^{y=1 } + { \ensuremath{\biggl[\!\!\biggl [ \frac{\partial}{\partial y } \left [ u^\ast \frac{\partial x_0}{\partial x } \right ] , \psi_u \biggr]\!\!\biggr ] } } , \label{eq : pd_p_u_2 } \\ { \ensuremath{\biggl[\!\!\biggl [ u^\ast , c \frac{\partial u}{\partial x } \biggr]\!\!\biggr ] } } & = \frac{c}{2\pi } \int_0^{2\pi } d\theta \int_0 ^ 1 dy \ , \biggl [ u^\ast \ , u \biggr]_{x=0}^{x=2 } + { \ensuremath{\biggl[\!\!\biggl [ -c \frac{\partial u^\ast}{\partial x } , u \biggr]\!\!\biggr ] } } , \\ { \ensuremath{\biggl[\!\!\biggl [ u^\ast , - \omega \frac{\partial u}{\partial \theta } \biggr]\!\!\biggr ] } } &
= -\frac{\omega}{2\pi } \int_0 ^ 2 dx \int_0 ^ 1 dy \ , \biggl [ u^\ast \ , u \biggr]_{\theta=0}^{\theta=2\pi } + { \ensuremath{\biggl[\!\!\biggl [ \omega \frac{\partial u^\ast}{\partial \theta } , u \biggr]\!\!\biggr]}}.\end{aligned}\ ] ] using the green s function @xmath23 given in eq .
( [ eq : g ] ) , the function @xmath46 given in eq .
( [ eq : p_u ] ) can also be written in the following form : @xmath175 in eqs .
( [ eq : pd_p_u_1])([eq : pd_p_u_2 ] ) , we perform the following manipulations : @xmath176 , \psi_u \biggr]\!\!\biggr ] } } \nonumber \\ & = -\frac{1}{2\pi } \int_0^{2\pi } d\theta \int_0 ^ 2 dx \int_0 ^ 1 dy \ , \frac{\partial}{\partial x } \left [ u^\ast \left ( \frac{\partial x_0}{\partial y } - 1 \right ) \right ] \ , \psi_u \nonumber \\ & = -\frac{1}{2\pi } \int_0^{2\pi } d\theta \int_0 ^ 2 dx \int_0 ^ 1 dy \int_0 ^ 2 dx ' \int_0 ^ 1 dy ' \ , g(x , y , x ' , y ' ) \frac{\partial u'}{\partial x ' } \frac{\partial}{\partial x } \left [ u^\ast \left ( \frac{\partial x_0}{\partial y } - 1 \right ) \right ] \nonumber \\ & = -\frac{1}{2\pi } \int_0^{2\pi } d\theta \int_0 ^ 2 dx \int_0 ^ 1 dy \int_0 ^ 2 dx ' \int_0 ^ 1 dy ' \ , g(x ' , y ' , x , y ) \frac{\partial u}{\partial x } \frac{\partial}{\partial x ' } \left [ { u^\ast } ' \left ( \frac{\partial x_0'}{\partial y ' } - 1 \right ) \right ] \nonumber \\ & = -\frac{1}{2\pi } \int_0^{2\pi } d\theta \int_0 ^ 2 dx \int_0 ^ 1 dy \ , \psi_{u , x}^\ast \frac{\partial u}{\partial x } \nonumber \\ & = -\frac{1}{2\pi } \int_0^{2\pi } d\theta \int_0 ^ 1 dy \ , \biggl [ \psi_{u , x}^\ast \ , u \biggr]_{x=0}^{x=2 } + { \ensuremath{\biggl[\!\!\biggl [ \frac{\partial \psi_{u , x}^\ast}{\partial x } , u \biggr]\!\!\biggr]}},\end{aligned}\ ] ] and @xmath177 , \psi_u \biggr]\!\!\biggr ] } } \nonumber \\ & = \frac{1}{2\pi } \int_0^{2\pi } d\theta \int_0 ^ 2 dx \int_0 ^ 1 dy \ , \frac{\partial}{\partial y } \left [ u^\ast \frac{\partial x_0}{\partial x } \right ] \ , \psi_u \nonumber \\ & = \frac{1}{2\pi } \int_0^{2\pi } d\theta \int_0 ^ 2 dx \int_0 ^ 1 dy \int_0 ^ 2 dx ' \int_0 ^ 1 dy ' \ , g(x , y , x ' , y ' ) \frac{\partial u'}{\partial x ' } \frac{\partial}{\partial y } \left [ u^\ast \frac{\partial x_0}{\partial x } \right ] \nonumber \\ & = \frac{1}{2\pi } \int_0^{2\pi } d\theta \int_0 ^ 2 dx \int_0 ^ 1 dy \int_0 ^ 2 dx ' \int_0 ^ 1 dy ' \ , g(x ' , y ' , x , y ) \frac{\partial u}{\partial x } \frac{\partial}{\partial y ' } \left [ { u^\ast } ' \frac{\partial x_0'}{\partial x ' } \right ] \nonumber \\ & = \frac{1}{2\pi } \int_0^{2\pi } d\theta \int_0 ^ 2 dx \int_0 ^ 1 dy \ , \psi_{u , y}^\ast \frac{\partial u}{\partial x } \nonumber \\ & = \frac{1}{2\pi } \int_0^{2\pi } d\theta \int_0 ^ 1 dy \ , \biggl [ \psi_{u , y}^\ast \ , u \biggr]_{x=0}^{x=2 } + { \ensuremath{\biggl[\!\!\biggl [ -\frac{\partial \psi_{u , y}^\ast}{\partial x } , u \biggr]\!\!\biggr]}},\end{aligned}\ ] ] where we used the following abbreviations : @xmath178 and also defined the following functions : @xmath179 , \label{eq : past_x } \\
\psi_{u , y}^\ast(x - \phi , y , \theta ) & = \int_0 ^ 2 dx ' \int_0 ^ 1 dy ' \ , g(x ' , y ' , x , y ) \frac{\partial}{\partial y ' } \left [ { u^\ast}(x ' - \phi , y ' , \theta ) \frac{\partial x_0(x ' - \phi , y ' , \theta)}{\partial x ' } \right ] .
\label{eq : past_y}\end{aligned}\ ] ] here , we note that eqs .
( [ eq : past_uast_x])([eq : past_uast_y ] ) can be derived by applying the laplacian to eqs .
( [ eq : past_x])([eq : past_y ] ) , respectively . in this way , the adjoint operator @xmath50 , defined in eq .
( [ eq : operator ] ) , is obtained as @xmath180 - \frac{\partial}{\partial y } \left [ u^\ast \frac{\partial \psi_0}{\partial x } \right ] + \frac{\partial \psi_{u , x}^\ast}{\partial x } - \frac{\partial \psi_{u , y}^\ast}{\partial x } - c \frac{\partial u^\ast}{\partial x } + \omega \frac{\partial u^\ast}{\partial \theta}.\end{aligned}\ ] ] similarly to the boundary conditions for @xmath45 as @xmath181 the adjoint boundary conditions are given by @xmath182 which represent the periodic boundary condition on @xmath11 , the dirichlet zero boundary condition on @xmath15 , and the @xmath32-periodicity with respect to @xmath29 . in fact , under these adjoint boundary conditions , the bilinear concomitant @xmath183 = { \ensuremath{[\ ! [ u^\ast(x - \phi , y , \theta ) , { \cal l}(x - \phi , y , \theta ) u(x - \phi , y , \theta ) ] \ ! ] } } - { \ensuremath{[\ ! [ { \cal l}^\ast(x - \phi , y , \theta ) u^\ast(x - \phi , y , \theta ) , u(x - \phi , y , \theta ) ] \!]}}$ ] becomes zero , i.e. , @xmath184 = & + \frac{1}{2\pi } \int_0^{2\pi } d\theta \int_0 ^ 1 dy \
, \left [ u^\ast \ , \frac{\partial u}{\partial x } \right]_{x=0}^{x=2 } \nonumber \\ & - \frac{1}{2\pi } \int_0^{2\pi } d\theta \int_0 ^ 1 dy \ , \left [ \frac{\partial u^\ast}{\partial x } \ , u \right]_{x=0}^{x=2 } \nonumber \\ & + \frac{1}{2\pi } \int_0^{2\pi } d\theta \int_0 ^ 2 dx \ , \left [ u^\ast \ , \frac{\partial u}{\partial y } \right]_{y=0}^{y=1 } \nonumber \\ & - \frac{1}{2\pi } \int_0^{2\pi } d\theta \int_0 ^ 2 dx \ , \left [ \frac{\partial u^\ast}{\partial y } \ , u \right]_{y=0}^{y=1 } \nonumber \\ & -\frac{1}{2\pi } \int_0^{2\pi } d\theta \int_0 ^ 1 dy \ , \left [ u^\ast \ , \frac{\partial \psi_0}{\partial y } \ , u \right]_{x=0}^{x=2 } \nonumber \\ & + \frac{1}{2\pi } \int_0^{2\pi } d\theta \int_0 ^ 2 dx \ , \left [ u^\ast \ , \frac{\partial \psi_0}{\partial x } \ , u \right]_{y=0}^{y=1 } \nonumber \\ & + \frac{1}{2\pi } \int_0^{2\pi } d\theta \int_0 ^ 1 dy \ , \left [ u^\ast \ , \left ( \frac{\partial x_0}{\partial y } - 1 \right ) \ , \psi_u \right]_{x=0}^{x=2 } \nonumber \\ & - \frac{1}{2\pi } \int_0^{2\pi } d\theta \int_0 ^ 2 dx \ , \left [ u^\ast \ , \frac{\partial x_0}{\partial x } \ , \psi_u \right]_{y=0}^{y=1 } \nonumber \\ & - \frac{1}{2\pi } \int_0^{2\pi } d\theta \int_0 ^ 1 dy \ ,
\biggl [ \psi_{u , x}^\ast \ , u \biggr]_{x=0}^{x=2 } \nonumber \\ & + \frac{1}{2\pi } \int_0^{2\pi } d\theta \int_0 ^ 1 dy \ , \biggl [ \psi_{u , y}^\ast \ , u \biggr]_{x=0}^{x=2 } \nonumber \\ & + \frac{c}{2\pi } \int_0^{2\pi } d\theta \int_0 ^ 1 dy \ , \biggl [ u^\ast \ , u \biggr]_{x=0}^{x=2 } \nonumber \\ & - \frac{\omega}{2\pi } \int_0 ^ 2 dx \int_0 ^ 1 dy \ , \biggl [ u^\ast \ , u \biggr]_{\theta=0}^{\theta=2\pi } = 0 .
\label{eq : cals}\end{aligned}\ ] ] each term of the bilinear concomitant @xmath183 $ ] vanishes for the following reasons : the 1st , 2nd , 5th , 7th , 9th , 10th , and 11th terms in eq .
( [ eq : cals ] ) become zero owing to the @xmath10-periodicity with respect to @xmath11 for all the functions ; the 3rd , 4th , 6th , and 8th terms , the dirichlet zero boundary condition on @xmath15 for @xmath185 and @xmath186 ; the last term ( i.e. , the 12th term ) , the @xmath32-periodicity with respect to @xmath29 for all the functions .
in this appendix , we consider oscillatory cylindrical - hele - shaw convection with weak boundary forcing described by the following equation : @xmath187 where the stream function @xmath3 is determined from the temperature field @xmath0 as @xmath188 the above two equations are a reproduction of eqs .
( [ eq : t])([eq : p_t ] ) for readability .
the boundary conditions for the temperature field @xmath0 are now given by @xmath189 where the weak boundary forcing applied at the bottom ( @xmath13 ) and the top ( @xmath14 ) is described by @xmath190 and @xmath191 , respectively .
the stream function @xmath3 satisfies the dirichlet zero boundary condition on @xmath15 , i.e. , @xmath192 extending the transformation given by eq .
( [ eq : t - x ] ) , we consider the following transformation of the temperature field ( see also , e.g. , ref .
@xcite for this type of transformation ) : @xmath193 where the function @xmath194 is defined as @xmath195 .
\label{eq : ptpb}\end{aligned}\ ] ] we note that @xmath196 and @xmath197 . by considering the boundary forcing , the stream function @xmath3 is also decomposed as @xmath198 substituting eqs .
( [ eq : t - xwp])([eq : p - pxpp ] ) into eq .
( [ eq : tb ] ) , we derive the following equation : @xmath199 where the first - order terms associated with the boundary forcing are given by @xmath200 p + j(\psi_x , p ) + j(\psi_p , x ) - \frac{\partial \psi_p}{\partial x}.\end{aligned}\ ] ] applying eqs .
( [ eq : t - xwp])([eq : p - pxpp ] ) to eq .
( [ eq : p_tb ] ) and also considering the linearity of eq .
( [ eq : p_tb ] ) , we obtain the following equations : @xmath201 we note that eq .
( [ eq : px_x ] ) corresponds to eq .
( [ eq : p_x ] ) . from eqs .
( [ eq : bctb])([eq : bctt ] ) and eqs .
( [ eq : t - xwp])([eq : ptpb ] ) , as in eq .
( [ eq : bcxy ] ) , the convective component @xmath17 satisfies the dirichlet zero boundary condition on @xmath15 , i.e. , @xmath202 the two stream functions , @xmath203 and @xmath204 , also satisfy the dirichlet zero boundary condition on @xmath15 , i.e. , @xmath205 in this way , oscillatory convection with boundary forcing is found to be exactly described by eq .
( [ eq : xb ] ) , which possesses the form of eq .
( [ eq : xp ] ) from the viewpoint of the phase reduction .
when the boundary forcing is absent , i.e. , @xmath206 , the system is assumed to exhibit oscillatory cylindrical - hele - shaw convection described by the following limit - torus solution : @xmath207 which is a reproduction of eq .
( [ eq : x_x0 ] ) for readability . under the assumption that the boundary forcing is sufficiently weak , as in sec .
[ subsec:2e ] , using the phase sensitivity functions given by eqs .
( [ eq : zs])([eq : zt ] ) , we derive a set of phase equations from eq .
( [ eq : xb ] ) as follows : @xmath208 where the _ effective boundary forcing function _ is given by @xmath209 p + j(\psi_0 , p ) + j(\psi_p , x_0 ) - \frac{\partial \psi_p}{\partial x}.\end{aligned}\ ] ] we note that the two functions , @xmath210 and @xmath211 , are given by eq .
( [ eq : x0 ] ) and eq .
( [ eq : p0 ] ) , respectively .
we also note that the second - order term of eq .
( [ eq : xb ] ) , i.e. , @xmath212 , is negligible owing to the smallness of @xmath84 .
furthermore , we consider the case in which the weak forcing is spatially homogeneous as follows : @xmath213 for this case , the effective boundary forcing function is simplified as @xmath214 \frac{\partial \psi_0}{\partial x } - \bigl [ y \ , \dot{q}_{\rm t}(t ) + ( 1 - y ) \ , \dot{q}_{\rm b}(t ) \bigr].\end{aligned}\ ] ] in summary , the phase description method is applicable to the oscillatory cylindrical - hele - shaw convection with weak forcing applied to the boundary given in eqs .
( [ eq : bctb])([eq : bctt ] ) as well as to the bulk given in eq .
( [ eq : tp ] ) . because the spatial extent of the cylindrical - hele - shaw cell is quasi - two - dimensional , external forcing can be easily applied not only to the boundary but also to the bulk in experiments
; however , in general , although external forcing can be easily applied to boundary , it may be difficult to apply external forcing to bulk of fluid systems in experiments .
in fact , in the rotating fluid annulus experiments @xcite , perturbations are applied to the boundary .
therefore , the treatments of boundary forcing as developed in this appendix are required to apply the phase description method to such fluid systems .
izhikevich , dynamical systems in neuroscience : the geometry of excitability and bursting , mit press , cambridge , ma , 2007 .
ermentrout , d.h .
terman , mathematical foundations of neuroscience , springer , new york , 2010 .
y. kawamura , h. nakao , collective phase description of oscillatory convection , chaos 23 ( 2013 ) 043129 .
[ arxiv:1110.1128 ] + y. kawamura , h. nakao , noise - induced synchronization of oscillatory convection and its optimization , phys .
e 89 ( 2014 ) 012912 .
[ arxiv:1401.4223 ] h. nakao , t. yanagita , y. kawamura , phase description of stable limit - cycle solutions in reaction - diffusion systems , procedia iutam 5 ( 2012 ) 227 - 233 ; h. nakao , t. yanagita , y. kawamura , phase - reduction approach to synchronization of spatiotemporal rhythms in reaction - diffusion systems , phys .
x ( 2014 ) in press .
m. ghil , s. childress , topics in geophysical fluid dynamics : atmospheric dynamics , dynamo theory , and climate dynamics , springer , new york , 1987 .
castrejn - pita , p.l .
read , synchronization in a pair of thermally coupled rotating baroclinic annuli : understanding atmospheric teleconnections in the laboratory , phys .
104 ( 2010 ) 204501 .
a. bernardini , j. bragard , h. mancini , synchronization between two hele - shaw cells , math .
. eng . 1 ( 2004 ) 339 - 346 ; a. bernardini , synchronization between two hele - shaw cells , ph.d . thesis , university of navarra , 2005 .
y. kawamura , h. nakao , k. arai , h. kori , y. kuramoto , collective phase sensitivity , phys .
( 2008 ) 024101 .
[ arxiv:0807.1285 ] + y. kawamura , collective phase dynamics of globally coupled oscillators : noise - induced anti - phase synchronization , physica d 270 ( 2014 ) 20 - 29 .
[ arxiv:1312.7054 ] y. kawamura , h. nakao , k. arai , h. kori , y. kuramoto , phase synchronization between collective rhythms of globally coupled oscillator groups : noisy identical case , chaos 20 ( 2010 ) 043109 .
[ arxiv:1007.4382 ] j. lber , m. br , h. engel , front propagation in one - dimensional spatially periodic bistable media , phys . rev .
e 86 ( 2012 ) 066210 .
y. pomeau , p. manneville , stability and fluctuations of a spatially periodic convective flow , j. de phys .
40 ( 1979 ) 609 - 612 . | we formulate a theory for the phase description of oscillatory convection in a cylindrical hele - shaw cell that is laterally periodic .
this system possesses spatial translational symmetry in the lateral direction owing to the cylindrical shape as well as temporal translational symmetry .
oscillatory convection in this system is described by a limit - torus solution that possesses two phase modes ; one is a spatial phase and the other is a temporal phase .
the spatial and temporal phases indicate the `` position '' and `` oscillation '' of the convection , respectively . the theory developed in this paper
can be considered as a phase reduction method for limit - torus solutions in infinite - dimensional dynamical systems , namely , limit - torus solutions to partial differential equations representing oscillatory convection with a spatially translational mode .
we derive the phase sensitivity functions for spatial and temporal phases ; these functions quantify the phase responses of the oscillatory convection to weak perturbations applied at each spatial point . using the phase sensitivity functions , we characterize the spatiotemporal phase responses of oscillatory convection to weak spatial stimuli and analyze the spatiotemporal phase synchronization between weakly coupled systems of oscillatory convection . _
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ * highlights : * + we develop a phase reduction theory for oscillatory convection with a spatial mode .
+ the theory can be considered as a phase description method for limit - torus solutions .
+ we derive phase sensitivity functions for spatial and temporal phases of convection .
+ we can quantify spatiotemporal phase responses of convection to weak perturbations .
+ we can analyze spatiotemporal phase synchronization between weakly coupled systems .
+ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ |
the problem of fermionic pairing and superfluidity with mismatched fermi surfaces has been widely investigated in recent years from theoretical standpoints @xcite as well as experimentally @xcite . a mismatch in fermi surfaces can be easily realized in a two - species fermionic system where the pairing species have unequal populations or different masses/ chemical potentials .
typically , such a system would consist of two different fermionic atoms ( e.g. , @xmath0li and @xmath1k ) or alternatively , two hyperfine states of the same atom ( e.g. , states @xmath2 and @xmath3 of @xmath0li atoms ) .
several phases have been proposed to describe the possible ground state of such a system , including the fflo phase @xcite(pairing with non - zero centre - of - mass momentum , where the order parameter shows a spatial variation ) , the gapless bp ( breached pair ) phase @xcite ( pairing with zero centre - of mass momentum : where the order parameter is non zero but the excitation energy becomes zero ) , and the inhomogeneous phase - separated state @xcite , where any two pure states coexist . in this note ,
our focus is on the homogeneous and gapless breached pair state , also known as the sarma state .
sarma @xcite , in the early studies of superconductivity , predicted a spatially isotropic , homogenous and uniform state with gapless excitation modes in the presence of a magnetic field .
however , for weak coupling bcs theory , this gapless breached pair state marks the maximum of the thermodynamic potential , and thus , can not be the stable ground state of the system .
this is the well - known sarma instability . in the last few years
, several mechanisms were put forward to avoid this instability . according to forbes
@xcite , a stable sarma state is possible in a model with finite range interaction where the momentum dependence of the pairing gap cures the instability .
it has also been proposed by a number of workers @xcite that the breached pair state becomes stable in the deep bec regime , if the bcs - bec crossover picture is taken into account .
he et al.,in a very recent work @xcite has argued that the breached pair state can be a possible ground state in the weak coupling region for a two - band fermi system .
+ to study the breached pair state , we start with a two - species fermionic system . in addition to the weak bcs attraction ( denoted by @xmath4 ) , we consider a strong interaction ( @xmath5 ) of the feshbach variety which couples a fermion of type @xmath6 with a @xmath7 fermion to form a bosonic molecule @xmath8 .
our model resembles the one used in @xcite , but we extend it to cover the two - species case .
the system is described by the hamiltonian : @xmath9\ ] ] here @xmath10 , @xmath11 are the creation and annihilation operators for atom @xmath6 , while @xmath12 , @xmath13 are the corresponding operators for atom @xmath7 . also , @xmath14 = @xmath15 and @xmath16 = @xmath17 , while @xmath18 , @xmath19 are the respective kinetic energies and @xmath20 , @xmath21 the respective chemical potentials for species @xmath6 and @xmath7 .
the annihilation and creation operators for the composite boson @xmath8 are @xmath22 and @xmath23 ( we restrict ourselves to the case where only zero - momentum bosons are formed).the chemical potential for the bosons is @xmath24 , and @xmath25 is the threshold energy of the composite bose particle energy band . + using this hamiltonian , we study the effect of the feshbach parameters first on the gapped bcs superfluid state , and then on the gapless sarma state , taking a variational approach .
the stability of such a gapless state is analysed ensuring the positivity of the superfluid density @xcite . considering the bcs - bec crossover picture ,
we show that the breached pair state is stable only in a narrow region in the bec side , bounded by two magnetic field values .
we want to study the ground state of the two - fermion system by a variational method .
we take @xmath26 as the trial form of the ground state of the system . here
= @xmath28 = @xmath29 , i.e , the probability of the pair ( @xmath30 , @xmath31)being occupied is @xmath32 , and the probability that it is unoccupied is @xmath33 , while @xmath34 is the ground state for the condensate part of the boson subsystem .
our variational prescription is that @xmath34 has to be chosen in such a manner that it is an eigenstate of the annihilation operator b , which would make it a coherent state . in terms of fock states
@xmath34 = @xmath35 . using the normalization condition
, we arrive at @xmath36 .
the ground state of the system is consequently given by @xmath37 , @xmath38 being the expectation value of the total number of bosons in the condensed state . from the hamiltonian ( [ hamil ] ) and the ground state wave function ( [ grnd ] ) , the ground state energy of the system would be @xmath39 minimizing @xmath40 with respect to @xmath41 and @xmath42 we get @xmath43 where @xmath44 .
+ @xmath45 is defined by the relation @xmath46 we note that this expression matches with the one obtained by ohashi et al @xcite using diagrammatics . in order to avoid a possible ultraviolet divergence , the summation in the right hand side of equation([gap ] )
has to be carried over upto a cutoff . for a metallic superconductor , this cutoff
is played by @xmath47 , @xmath48 being the debye frequency . in ultracold atomic systems ,
the cutoff is determined by the range of the interatomic potential .
now , if we choose @xmath49 , the usual form of the gap equation follows , and we can identify @xmath50 as the gap in the excitation spectrum .
a similar relation is obtained in finite temperature systems as well , if we incorporate the appropriate fermi distribution functions in the expression of the free energy . in terms of this gap parameter , @xmath51 and @xmath41
can be expressed as @xmath52 and @xmath53 .
+ the critical temperature @xmath54 ( temperature at which the gap vanishes ) is proportional to @xmath55 , the bcs gap in the weak coupling limit , which is given by @xcite @xmath56 here @xmath57 is the density of states in the fermi level , @xmath58 is the base of natural logarithms , and @xmath59 is the fermi energy .
thus the value of @xmath55 ( and @xmath60 ) can be raised or lowered by adjusting @xmath25 , the tunable parameter depending on the feshbach resonance process .
+ next we study gapless excitations.the quasiparticle dispersions as obtained from standard bcs - like treatment is of the form @xcite : @xmath61 -@xmath62 ) curve for @xmath63 , @xmath64,title="fig : " ] -@xmath62 ) curve for @xmath63 , @xmath64,title="fig : " ] -@xmath62 ) curve for @xmath63 , @xmath64,title="fig : " ] -@xmath62 ) curve for @xmath63 , @xmath64,title="fig : " ] -@xmath62 ) curve for @xmath63 , @xmath64,title="fig : " ] -@xmath62 ) curve for @xmath63 , @xmath64,title="fig : " ] depending on the values of particle masses and corresponding chemical potentials , these quasiparticle excitations can be negative , thus leading to gapless excitations .
this is possible only if the magnitude of @xmath50 is less than a critical value @xmath65 @xcite .
@xmath66 when @xmath67 , both @xmath68 and @xmath69 remain positive for all values of @xmath62 .
this corresponds to usual bcs pairing .
when @xmath70 , either @xmath68 or @xmath69 crosses zero at the points : @xmath71^{1/2}\ ] ] the difference between @xmath72 and @xmath73 gives the span over which we get a gapless region in the parameter space .
the state with gapless excitations marks the coexistence of the superfluid and normal components at zero temperature , and is called the sarma phase , or the breached pair state .
when we include the feshbach term in the hamiltonian , we have a control over this sarma phase as well . in figures(1,2,3 )
, we plot the @xmath40-@xmath62 curve for a two - species system . here
we scale all energies by @xmath74 ( fermi energy of species @xmath7 ) and all momenta by @xmath75 ( fermi momentum of species @xmath7 ) .
it is evident that in this convention , @xmath76 and @xmath77(in the bcs limit ) .
we choose @xmath78 , @xmath79 and multiply quantities @xmath80 , @xmath5 , @xmath81 and @xmath45 by @xmath57 , the density of states at the fermi level to get dimensionless quantities
@xmath82 , @xmath83 , @xmath84 and @xmath85 .
. had there been no feshbach coupling @xmath5 , we would get a gapless region from @xmath62=1.01 to @xmath62=1.85 ( in units of @xmath75 ) as seen from figure-1 .
+ if @xmath64 is introduced in the system , and we choose @xmath87 in such a way that @xmath88 , we get the sarma phase for a wider region , from @xmath62=1.00 to @xmath62=1.95 ( in units of @xmath75 ) as seen from figure-2 . + if , on the other hand , @xmath89 , the gapless phase vanishes entirely ( figure-3 ) . +
the stability of a superfluid phase has been studied in different ways like demanding positivity of the superfluid density @xcite , ensuring non - negative eigenvalues of the number susceptibility matrix of the system@xcite , minimization of thermodynamic potential @xcite etc . here
we adopt the first criterion and demand that the superfluid density @xmath90 must be positive .
if the fermions are charged , this is equivalent to saying that the meissner mass - squared must be positive @xcite , since @xmath90 , the superfluid density , and @xmath91 , the meissner mass squared obey the relation : @xmath92 @xcite , @xmath93 being the electronic charge and @xmath94 , the mass of the particle . in our treatment , we work with the superfluid density directly , since we are talking of a charge neutral fermi system .
however , the term meissner mass " can still be used just to continue the analogy with superconductors , as it only serves to express the superfluid density to within a multiplicative constant
. for convenience , we assume that the two species have equal masses . therefore , equation ( [ p2 - 11 ] ) takes a simpler form : @xmath95}^{\frac{1}{2}}\ ] ] where @xmath96 and @xmath97 .
thus , @xmath98 here .
he et al @xcite have shown that when the two species have equal masses , the superfluid density can be expressed as @xmath99 where @xmath100 .
we note that here @xmath101 marks the bare fermion density or the density of the atoms that have not been part of the condensate yet ( i.e , unpaired fermions , fermions forming the cooper pair and fermions that constitute the non - condensate bosons all are counted in n ) .
since the total number of bare fermi atoms is conserved , so @xmath102= constant , @xmath38 being the expectation value of the total number of bosons in the condensate . if @xmath103 , for @xmath90 is to be positive , @xmath104 has to be less than 1 , which is satisfied if @xmath105 , and @xmath106 .
+ now when both @xmath72 and @xmath73 are real and there is a breached pair phase between them ( this state has sometimes been termed as bp2 state in the literature @xcite ) , the bare fermion density can be written as : @xmath107\ ] ] here the first term denotes the contribution from the gapless region between @xmath72 and @xmath73 , i.e , the normal component .
the next two integrals take care of the contributions from the gapped superfluid regimes , one from momenta @xmath108 to @xmath72 and the other from @xmath73 to infinity .
the terms , when rearranged , gives @xmath109 in the weak coupling limit , the last two integrals are very small , and hence , can be neglected .
therefore , @xmath110=1 , and @xmath90 is negative .
thus , the sarma state is unstable here .
+ + we shift our focus to a special case , where only one of @xmath72 and @xmath73 is real ( in the literature , this is called the bp1 state @xcite ) . in this case the @xmath40-@xmath62 curve resembles figure 4 , and the bare fermion density is @xmath111 this yields @xmath112 , and thus the stability criterion is fulfilled .
-@xmath62 ) curve when @xmath72 is imaginary , title="fig : " ] -@xmath62 ) curve when @xmath72 is imaginary , title="fig : " ] now , from equation([p2 - 12 ] ) , @xmath72 is imaginary if @xmath113 , i.e , @xmath114 .
so , sarma phase is stable only in a region where the chemical potential of one species is positive , and the other , negative , provided the magnitude of their product is greater than @xmath115 .
+ leggett has shown @xcite that in the bcs - bec crossover picture , the chemical potential can be determined by solving the gap and the number equations . extending these equations to the two species case , we get @xmath116 where @xmath117 is the @xmath118 scattering length and @xmath119 and @xmath120 correspond to the fermi wave number of the more and the less populated species respectively . let @xmath121 denote the wave number corresponding to the breaching point for the bp1 state , i.e , @xmath122 . converting the sums into integrals
, we obtain that in the weak coupling limit , @xmath123 , as expected .
in contrast , in the strong coupling limit , @xmath124 provided @xmath125 ( a condition which is satisfied if the population imbalance is small compared to the total population ) . in this case @xmath126
asymptotically approaches @xmath127 , i.e , half the binding energy of the molecule .
+ if we think of @xmath20 and @xmath21 separately , the first one differs from the other by the fermi energy of the excess fermions .
moreover , in the presence of a magnetic field @xmath128 , there is an asymmetry between the chemical potentials given by @xmath129 , @xmath130 being the fermion magneton .
so we have @xmath131 and @xmath21 in the crossover picture ] therefore , @xmath132 our domain of interest is when @xmath20 is positive , and @xmath21 is negative , i.e , while approaching the bec side , @xmath21 has already crossed zero but @xmath20 has not .
now , @xmath20 and @xmath21 becomes zero at magnetic field values @xmath133 and @xmath134 respectively , as sketched in figure ( 5 ) , where @xmath135\\ \label{h2 } h_2=(2/m_b)\left [ -\frac{\hbar^2}{2ma_{s2}^2}+\frac{2\epsilon_{b_f}(k_{b_f}a_{s2})}{3\pi}\left(1+\frac{2k'^3a_{s2}^3}{\pi}\right ) -\frac { \hbar^2 k'^2}{4m}\right]\end{aligned}\ ] ] here @xmath136 , @xmath137 are the respective values of the scattering length at @xmath133 and @xmath134 .
between these two magnetic field values , the sarma state will be stable .
experiments with population - imbalanced fermionic systems have been done by zwierlein et al .
@xcite and partridge et al @xcite.they obtained the signature of superfluidity in an unequal mixture of two spin states of @xmath0li atoms , and a quantum phase transition between the superfluid state and the normal state was observed at a critical polarization .
another method for experimental detection of the breached pair phase has been suggested by yi et al .
however , no clear signature of this gapless phase has been obtained till date .
we now use the data obtained from these experiments on population - imbalanced gas of @xmath0li atoms @xcite to make an estimate of the magnetic field values corresponding to the breached pair sarma state . to be able to use the expressions ( [ h1],[h2 ] ) ,
we need to know the scattering length as a function of the magnetic field .
this is provided by @xmath138 , where @xmath139 is the background scattering length , @xmath140 is the width of the resonance , and @xmath141 is the position of the resonance peak .
we use this expression for scattering length , and put @xmath142 ( @xmath143=bohr radius ) , which is the singlet scattering length for @xmath0li .
we also take @xmath144 , @xmath145 as population of the two species , @xmath141= 834 g , @xmath140= 300 g ( these values correspond to experimental data for feshbach resonances in @xmath0li as reported in @xcite . ) we now solve equations ( [ h1],[h2 ] ) explicitly to find @xmath133 to be 832.40 g and @xmath134 to be 833.95 g. it would be interesting to speculate that the superfluid observed by zwierlein et al . in this range
is of the bp1 variety .
this estimate of @xmath133 and @xmath134 is not a highly accurate one .
in fact , the range of the breached pair state should get shifted a bit towards lower values of the magnetic field .
actually , for deriving equations ( [ h1],[h2 ] ) , we had assumed the magnitudes of the chemical potentials to be very large compared to @xmath50 and @xmath146 , an assumption which does not hold at points where @xmath20 and @xmath21 become zero .
however , the values are not unreasonable in view of the fact that in the bcs - bec crossover picture , the chemical potential falls quite sharply after crossing zero and quickly becomes a large negative quantity .
here we have studied a two - species fermionic system in the presence of feshbach resonance , taking a variational route with an explicit construct of the ground state .
the gapless breached pair state was discussed , and its stability was analysed ensuring the positivity of the superfluid density .
we showed that a breached pair state with two fermi surfaces is always unstable , while its single fermi - surface counterpart is stable when the chemical potential of the two pairing species bear opposite signs .
the stability of the breached pair state is indeed a widely debated issue .
although it has often been suggested that the bp1 state might be stable in deep bec region , nothing , to the best of our knowledge , was said anything about how ` deep ' that really is . in this note , however , we observe that , the requirement that @xmath147 and @xmath148 should be of opposite signs , automatically puts two bounds in terms of the feshbach magnetic fields , between which the gapless state is stable . moreover ,
this stable breached pair state is obtained not in deep bec , but near the vicinity of the point when the average chemical potential crosses zero , i.e , right after the onset of condensation .
gubankova et al @xcite , while discussing the stability of breached pair states by analysing the number susceptibility , commented that stable gapless states with a single fermi surface exist for negative average chemical potential . in a recent paper ,
a. mishra et al.@xcite reached the same conclusion by comparing the thermodynamic potentials of the condensed phase and the normal phase .
although the criterion they arrive at ( the negativity of the average chemical potential ) does not fully match with ours ( the chemical potential of the two species to have opposite signs ) , there is definitely a region of overlap . in their experiment ,
zwierlein et al . observed that superfluidity breaks down when the pairing gap @xmath50 becomes small compared to the chemical potential difference @xmath149 .
we note that this matches with the stability criterion for the bp2 state , since , if @xmath150 and @xmath151 , the state becomes unstable , as seen from equation ( [ m2 ] ) . as for the bp1 state
, we have shown that this state is stable in a region where @xmath20 and @xmath21 have opposite signs , which can be achieved by keeping the system between two specific magnetic field values .
for simplicity we took the two species to have same masses in our calculation , but the treatment should be extendable to a situation where the fermion species are of different masses , as for example in a @xmath0li-@xmath152k system .
the author would like to express her gratitude to her supervisor prof .
j. k. bhattacharjee for his keen interest and helpful comments during the course of this work .
constructive suggestions from the referee are also thankfully acknowledged .
w. v. liu and f. wilczek , phys .
lett . * 90 * , 047002 ( 2003 ) .
p. f. bedaque , h. caldas , and g. rupak , phys .
lett . * 91 * , 247002 ( 2003 ) .
wu and s. yip , phys .
a * 67 * , 053603 ( 2003 ) h. caldas , c.w .
morais and a.l .
mota , phys .
d * 72 * , 045008 ( 2005 ) j. carlson and s. reddy , phys .
lett . * 95 * , 060401 ( 2005 ) .
d. t. son and m. a. stephanov , phys
a * 74 * 013614 ( 2006 ) .
d. e. sheehy and l. radzihovsky , phys .
. lett . * 96 * , 060401 ( 2006 ) .
d. e. sheehy and l. radzihovsky , annals of physics * 322 * , 1790 ( 2007 ) c. c. chien , q. chen , y. he and k. levin , phys rev lett .
* 98 * , 110404 , ( 2007 ) q. chen , y. he , c. c. chien and k. levin , phys rev b * 75 * , 014521 , ( 2007 ) t. koponen , j. kinnunen , j.p .
martikainen , l.m .
jensen , p. torma , new journal of physics b * 8 * , 179 ( 2006 ) m. iskin and c. a. r. sa de melo , phys .
lett . * 97 * , 100404 ( 2006 ) k. b. gubbels , m. w. j. romans , and h. t. c. stoof , phys . rev
. lett . * 97 * , 210402 ( 2006 ) pairing in fermionic systems , edited by a. sedrakian , j. w. clark , and m. alford , world scientific publishing co , singapore , 2006 g.sarma , j.phys.chem.solid 24,1029(1963 ) a. i. larkin , yu .
n. ovchinnikov , sov .
jetp * 20 * , ( 1965 ) p. fulde , r. a. ferrel , phys .
a , * 135 * , 550 , ( 1964 ) m. m. forbes , e. gubankova , w. v. liu and f. wilczek , phys .
rev . lett.*94 * , 017001(2005 ) h.hu and x.j .
liu , phys .
rev . a , 73 , 051603(r)(2006 ) m. mannarelli , g. nardulli and m. ruggieri , phys . rev .
a.74 , 033606(2006 ) ; c .- h .
wu , and s .- k .
yip , phys .
b * 73 * , 132506 ( 2006 ) . a. mishra and h. mishra the eur . phys .
j. d. , * 53*,75,(2009 ) l. he and p. zhuang , phys .
b * 79 * , 024511 ( 2009 ) y. ohashi and a. griffin , phys . rev
. lett . * 89 * , 130402 ( 2002 ) m. holland , s.j.j.m.f .
kokkelmans , m.l .
chiofalo , r. walser , phys .
rev . lett.*87 * , 120406 ( 2001 ) r.p .
feynman , statistical mechanics , a set of lectures , addison - wesley , new york , 1988 c. j. pethick and h. smith , bose - einstein condensation in dilute gases , cambridge univ .
press 2002 l. he , m. jin , p. zhuang , phys .
b * 73 * , 214527 ( 2006 ) l. he , m. jin , p. zhuang , phys . rev .
b * 73*,024511 ( 2006 ) m. kitazawa , d.h .
rischke , i. shovkovy , phys . lett .
b * 637 * , 367 ( 2006 ) e. gubankova , a. schmitt , f. wilczek , phys . rev .
b * 74 * , 064505 ( 2006 ) a.j .
leggett , quantum liquids , bose condensation and cooper pairing in condensed - matter systems , oxford university press , oxford , 2006 m. w. zwierlein , a. schirotzek , c. h. schunck and w. ketterle , science * 311 * , 492 ( 2006 ) m. w. zwierlein , c.h .
schunck , a. schirotzek , and w. ketterle , nature ( london ) * 442 * , 54 ( 2006 ) g.b .
partridge , w. li , r. i. kamar , y. a. liao , and r. g. hulet , science * 311 * , 503 ( 2006 ) c. h. schunck , m. w. zwierlein , c. a. stan , s. m. f. raupach , and w. ketterle , a. simoni , e. tiesinga , c. j. williams , p. s. julienne , pra * 71 * , 045601(2005 ) w. yi and l .- m .
duan , phys .
lett . * 97 * , 120401 ( 2006 ) h. caldas and a.l .
mota , j stat .
mech , p08013 ( 2008 ) | we investigate the phenomenon of fermionic pairing with mismatched fermi surfaces in a two - species system in the presence of feshbach resonance , where the resonantly - paired fermions combine to form bosonic molecules .
we observe that the feshbach parameters control the critical temperature of the gapped bcs superfluid state , and also determine the range over which a gapless breached pair state may exist . demanding the positivity of the superfluid density , it is shown that although a breached pair state with two fermi surfaces is always unstable , its single fermi - surface counterpart can be stable if the chemical potentials of the two pairing species have opposite signs .
this condition is satisfied only over a narrow region in the bec side , characterized by an upper and a lower limit for the magnetic field .
we estimate these limits for a mixture of two hyperfine states of @xmath0li using recent experimental data . |
recent works have emphasized that globular clusters ( gcs ) are stellar systems which exhibit strong dynamical evolution in the potential well of their host galaxy ( gnedin & ostriker 1997 , hereafter go97 , murali & weinberg 1997 ) . since their birth , gcs
suffer internal evolution on their own : at the very beginning gcs evolve rapidly because of the fast stellar evolution of their massive stars ( vesperini & heggie 1997 , portegies zwart et al .
the finite size of the gcs produces two - body relaxation between the member stars ; equipartition of energy heats the lighter stars which diffuse outwards into the halo while the heavier stars sink slowly towards the contracting core ( spitzer & hart 1971 ) . typically , the relaxation time @xmath1 ( hereafter we will refer to the relaxation time at the half mass radius ; see , e.g. , binney & tremaine 1987 ) in a gc is of the order of @xmath2 yr ( go97 ) , which is significantly less than the age of all galactic globular clusters . owing to the negative specific heat in self - gravitating stellar systems ( antonov 1962 , lynden - bell & wood 1968 ) ,
related to the virial equilibrium , the core is contracting monotonically during a process called the `` gravothermal catastrophe '' , leading to core collapse , characterized by very high stellar densities ( up to @xmath3 ) .
it has been eventually recognized that this process is not so catastrophic after all , since the cluster core does not collapse for ever but bounces back towards lower stellar density phases ( hnon 1975 ) .
the key role of binaries ( either primordial or formed via encounters during the high stellar density phase ) has been emphasized in the internal dynamics of the gcs : these binaries act as a heating source and slow down the core collapse ( goodman & hut 1989 ) and even reverse it .
many authors have studied numerically the post core - collapse phase , using conducting - gas - sphere , fokker - planck , and n - body codes , in simulations computed well into core collapse and beyond , leading to the discovery of possible post - collapse gravothermal oscillations ( see meylan & heggie 1997 for a review ) .
globular clusters evolve dynamically , even when considering only relaxation , which causes stars to escape , consequently cluster cores to contract and envelopes to expand . any galaxy through its gravitational potential well influences the dynamical evolution of its globular clusters , accelerating their destruction .
the stars in the globular cluster halo are stripped by the tidal field of the galaxy : the outward diffusion of the stars towards the sub - thermalised halo is speeded up and the core contracts even more ( spitzer & chevalier 1973 ) .
moreover the gravitational shocks heat up the outer parts of the globular cluster , increasing its loss of stars ( aguilar et al .
1988 , weinberg 1994 , kundic & ostriker 1995 , leon et al .
shocks are caused by the tidal field of the galaxy : interactions with the disk , the bulge and , at a lower level , with the giant molecular clouds ( gmcs , see spitzer 1958 ) , heat up the outer regions of each star clusters .
the disk - shocking occurs when the gc crosses the thin disk where it is compressed by the varying z - component of the galactic plane potential ; this has been found to dominate the heating of gcs ( chernoff et al .
a gc globally gains energy during the crossing and exhibits peculiar transient deformation ( leon et al .
2000 ) . the shocks with the bulge and the gmcs are similar in their physical processes :
the gcs suffer an elongation aligned parallel to the density gradient in the bulge or the gmc .
these processes combined with the internal dynamical evolution have probably destroyed an important fraction of the primordial gcs and are still currently at play .
go97 estimate that _
`` half of the present clusters are to be destroyed within the next hubble time''_.
all gcs are expected to have already lost an important fraction of their mass , deposited in the form of individual stars in the halo of the galaxy .
the mass - loss rate is a function of the total mass of the cluster , its structural parameters like the concentration @xmath4 ( where @xmath4 = log ( @xmath5 ) with and are the tidal and core radii ) , and its orbital motion around the galactic center .
till recently , the only way to investigate the orbital history of a globular cluster , apart from proper motions , was to derive its tidal radius , which gives an indication of its perigalactic distance .
unfortunately , there are difficulties in defining the tidal radius , both theoretically and observationally ; and , as expected , there are discrepancies between the theoretical and observational values of the tidal radius ( see , e.g. , odenkirchen et al . 1997 ,
scholz et al . 1998 ) .
n - body simulations of globular clusters embedded in a realistic galactic potential ( oh & lin 1992 ; johnston et al . 1997 ; combes , leon & meylan 1999 , hereafter clm99 ) have been performed in order to study the amount of mass loss for different kinds of orbits and different kinds of clusters , along with the dynamics and the mass function in tidal tails .
the 2-d structure of such tidal tails appears to be a good tracer of gravitational shocks and should be a tracer of the potential well
. moreover the detection of unbound stars released by the clusters is the only way to measure directly the mass loss rate of the cluster .
grillmair et al . (
1995 ) in an analysis of star - count in the outer parts of a few galactic globular clusters found extra - cluster overdensities that they associated partly with stars stripped into the galaxy field .
similar tidal interaction remnants around globular clusters have also been found in external galaxies : grillmair et al.(1996 ) observed three clusters in m31 which exhibit departures from a king profiles .
leon et al . ( 1999 ) find tidal extensions in the outskirts of interacting binary clusters and isolated clusters in the large magellanic cloud ( lmc ) .
not surprisingly , galactic tidal forces play also an essential role in the evolution of smaller - scale galactic clusters , the open clusters : they exhibit tidal tails in their neighborhood ( odenkirchen 1998 ; bergond et al .
2000 ) . in this work
we study the 2-d structure of the tidal tails associated with 20 galactic globular clusters by using the wavelet transform to detect weak structures at large scale and filter the strong background noise for the low galactic latitude clusters .
we analyze with great care the observational bias which can be very important . in section 2
we present the observations , in section 3 we describe the data reduction , in section 4 we present the results with , for each globular cluster , detailed comments about related observational bias .
section 5 presents the general discussion of all results .
.list of observed globular clusters . about plates / films , eso and src stand for the plates of the corresponding two surveys .
all the other plates / films are from our observations , except the two plates e131 and e1069 which are from the poss i. [ cols="<,<,^,^ " , ] let us compare the amount of tidally stripped stars obtained in our simulations and observations . because of the magnitude limitation of the plates and films , it is likely that we underestimate the observed tidal tails .
the mass of the tidal tails in the case of ngc 5139 has been computed for a salpeter law : if we assume a steep slope @xmath6= 2.8 for the mass function in the tidal tails , we get a tail mass equal to about 1 % of the total mass of the cluster , equal to 5.1 10@xmath7 . in spite of the great uncertainty on the star counts of the tidal tails ,
such a large mass is an upper limit for the tidal tail mass from the simulations ( clm99 ) .
it confirms both that ngc 5139 has a genuinely large total mass ( meylan et al .
1995 ) and that the spectrum mass is likely less steep than @xmath8 in the outer part .
all these considerations point toward a mass for the tidal tails between 0.6 and 1.0 % of the total mass of ngc 5139 . from n - body simulation performed by moore ( 1996 )
we can note that the presence of tidal tails is the indication of low dark matter content in globular clusters .
ngc 7492 and pal 12 provide an interesting comparison because of their similar characteristics : low masses ( 6 and 2 @xmath9 @xmath10 , respectively ) , low concentrations ( 1.0 and 0.9 , respectively ) , and large distances from the galactic center ( 23.5 and 14.7 kpc , respectively ) .
they are also both strongly influenced by the galaxy : go97 compute @xmath11 = 77.8 and 17.9 , respectively .
nevertheless , their tidal tails appear strongly different , with a very extended structure for pal 12 and very tiny one for ngc 7492 .
the last gravitational shock suffered by ngc 7492 , if any , has to have occured much before the last one for pal 12 .
the long time since the last gravitational shock suffered ngc 7492 has allowed the surface density of the unbound stars to fade along the cluster orbit .
we estimate the last tidal shock suffered by pal 12 to be about 350 myr .
we have observed 20 galactic globular clusters with multi - color schmidt plates and films on wide fields .
field and cluster stars are sorted in the color - magnitude plane . a star - count analysis is performed on the color selected stars to study the overdensities that can be attributed to the stars stripped from the globular clusters by tidal shocks ( disk / bulge ) as well as from internal dynamical evolution .
we use the wavelet transform in order to enhance the weak tidal structures at large scales and in order to filter the high background noise at low galactic latitude . after highlighting the observational biases resulting from dust extinction and background galaxy clustering at faint magnitudes , we reach the following conclusions : * all the clusters observed , which do not suffer from strong observational biases , present tidal tails , tracing their dynamical evolution in the galaxy ( evaporation , tidal shocking , tidal torquing , and bulge shocking ) . * the clusters in the following sub - sample ( viz .
ngc 104 , ngc 288 , ngc 2298 , ngc 5139 , ngc 5904 , ngc 6535 , and ngc 6809 ) exhibit tidal extensions resulting from a shock , i.e. tails aligned with the tidal field gradient . * the clusters in another sub - sample ( viz .
ngc 1261 , ngc 1851 , ngc 1904 , ngc 5694 , ngc 5824 , ngc 6205 , ngc 7492 , pal 5 , and pal 12 ) present extensions which are likely tracing the orbital path of the cluster with various degrees of mass loss .
* ngc 7492 is a striking case because of its very small extension and its high destruction rate driven by the galaxy as computed by go97 .
its dynamical `` twin '' for such an evolution , namely pal 12 , exhibits , on the contrary , a large extension tracing its orbital path , with a possible shock which happened more than 350 myr . * the velocity diffusion of the stripped stars is tentatively estimated , in one case ( viz .
ngc 6254 ) , to be similar to the cluster velocity dispersion . *
thanks to the relatively small distance of ngc 5139 and its high release of unbound stars during its current disk shocking , we estimate the mass loss to be between 0.6 and 1 % of the cluster total mass , taking into account a possible mass segregation in the cluster halo . this mass loss rate is consistent with our estimates from n - body simulations ( clm99 ) . *
the second break in the surface density slope , apart from the break at the tidal radius ( cf .
the case of ngc 6254 ) could be an indicator of some recent gravitational shocks , with the indicator displaying a range of values between -0.9 and -2 .
the latter is likely overestimated because of the uncorrected crowding towards the clusters .
the use of better quality data , e.g. wide - field ccd observations , combined with the present star - count method will allow in a near future to get rid easily of the observational biases and to obtain a better color selection thanks to a more accurate photometry
. these observations will provide more precise observational estimates of the mass loss rates for different regimes of galaxy - driven cluster evolution . with the help of numerical simulations and accurate proper motions
, it will be possible to constrain efficiently the parameters describing the galactic potential ( disk scale - height , surface density , bulge size ) . in case of a flat dark matter halo ( pfenniger et al.1994 ) , the tidal shock on the globular clusters
would be enhanced , depending on the surface density and the scale - height of this dark matter halo .
pal 2 is a good candidate to probe such dark matter halo flattening , because of its small distance to the galactic plane ( z = -2.2 kpc ) and its relatively large distance to the galactic center ( 21.6 kpc ) , placing this cluster in a region where the tidal shocking by the disk only is expected to be low .
we acknowledge warmly the eso schmidt telescope operators o. and g. pizarro , and the efficiency of the schmidt astronomers b. reipurth and j. brewer for the conduction of the observations .
we would like to think the whole mama team for its efficient support .
we are greatful to r. le poole ( sterrenwacht leiden ) for facilitating the use of poss i plates .
we thank e. bertin , l. cambresy , j. guibert , c. loir and m. odenkirchen for very helpful discussions .
we acknowledge the w.e .
harris s catalog of parameters for the galactic globular clusters .
o. gnedin gave us kindly an electronic version of his galactic globular cluster destruction rates .
and we are very grateful to the referee , c. grillmair , for his enlightening suggestions and comments which helped to improve this paper .
abell , g. o. , corwin , g. j. , olowin , p. 1989 , apjs , 70 , 1 alcaino , g. , liller , w. 1987 , aj , 94 , 1585 aguilar , l. a. , hut , p. , ostriker , j.p .
1988 , apj , 335,720 anderson , j. , 1998 , phd thesis , berkeley university anderson , j. , king , i. r. 1996 , asp conf .
92 : formation of the galactic halo ... inside and out , 257 antonov v.a .
1962 , vest .
7 , 135 armandroff , t. e. , da costa , g. s. 1991 , aj , 101 , 1329 bahcall , j. n. , soneira , r. m. , schmidt , m. 1983 , apj , 265 , 730 berger , j. , cordoni , j. p. , fringant , a. m. , guibert , j. , moreau , o. , reboul , h. , vanderriest , c. 1991 , a&as , 87 , 389 bergond , g. , leon , s. , guibert , j. 2000 , in prep .
bertin , e. , arnouts , s. 1996 , a&as , 117 , 393 binney , j. , tremaine , s. , 1987 , in `` galactic dynamics '' , princeton series bijaoui , a. 1991 , `` the wavelet transform '' , data analysis workshop , 3 caldwell , j. a. r. , ostriker , j. p. 1981 ,
apj , 251 , 61 cambrzy , l. 1999 , a&a , 345 , 965 cannon , r.d . ,
stobbie , r.s . , 1973 , mnras , 162 , 207 chernoff , d.f . , kochanek , c.s . , shapiro , s.l . 1986 , apj , 309 , 183 chereul , e. , crz , m. , bienaym , o. 1999 , a&a , 135 , 5 combes , f. , leon , s. , meylan . , s. , 1999 , a&a , 352 , 149 ( clm99 ) dinescu , d. i. , girard , t. m. , van altena , w. f. , mendez , r. a. , lopez , c. e. 1997 , aj , 114 , 1014 djorgovski s.g . , meylan g. 1994 , 108 , 1292 geisler , d. , piatti , a. e. , claria , j. j. , minniti , d. 1995 , aj , 109 , 605 gill , a.g . , henriksen , r.n .
1990 , apj , 365 , 27 gillett , f. c. , neugebauer , g. , emerson , j. p. , rice , w. l. 1986 , apj , 300 , 722 gnedin , o.y . ,
ostriker , j.p .
1997 , apj , 474 , 223 ( go97 ) goodman , j. , hut , p. 1989
, nature , 339 , 40 grillmair , c.j .
1992 , phd thesis , australian national university grillmair , c. j. , ajhar , e. a. , faber , s. m. , baum , w. a. , holtzman , j. a. , lauer , t. r. , lynds , c. r. , oneil , e. j. , j .. 1996 , aj , 111 , 2293 grillmair , c. j. , freeman , k. c. , irwin , m. , quinn , p. j. 1995 , aj , 109 , 2553 grossmann , a. , morlet , j 1985 , in streit l. ( ed ) mathematics+physics , lecture on recent results .
world scientific , singapore gunn , j.e .
, griffin , r.f .
1979 , aj , 84 , 752 harris , w.e .
1996 , aj , 112 , 1487 hesser , j. e. , shawl , s. j. , meyer , j. e. 1986 , pasp , 98 , 403 hurley , d. j. c. , richer , h. b. , fahlman , g. g. 1989 , aj , 98 , 2124 johnston , k. v. , sigurdsson , s. , hernquist , l. 1998 , astro - ph/9805291 kundic , t. ostriker , j. p. 1995 , apj , 438 , 702 lehmann , i. , scholz , r .- d .
1997 , a&a , 320 , 776 leon , s. , combes , f. , leeuwin f. 2000 , in prep .
leon , s. , bergond , g. , vallenari , a. , 1999 , a&a , 344 , 450 longaretti , p .- y . ,
lagoute , c. 1996 , 308 , 453 lynden - bell d. , wood , r. , 1968 , mnras , 138 , 495 merritt , d. , meylan , g. , mayor , m. , 1997 , aj , 114 , 1074 - 1086 meylan , g. , 1987 , a&a , 184 , 144 - 154 meylan , g. , 1989 , a&a , 214 , 106 - 112 meylan , g. , mayor , m. 1986 , a&a , 166 , 122 meylan , g. , mayor , m. , duqueynnoy , a. , dubath , p. 1995 , a&a , 303 , 761 meylan , g. , mayor , m. 1991 , a&a , 250 , 113 meylan , g. , heggie , d.c . , 1997 , a&ar , 8 , 1 moore , b. , 1996 , apj , 461 , l13 mould , j. r. , watson , a. m. , gallagher , j. s. , i. , ballester , g. e. , burrows , c. j. , casertano , s. , clarke , j. t. , crisp , d. , griffiths , r. e. , hester , j. j. , hoessel , j. g. , holtzman , j. a. , scowen , p. a. , stapelfeldt , k. r. , trauger , j. t. , westphal , j. a. 1996 , pasp , 108 , 682 murali , c. weinberg , m. d. 1997 , mnras , 291 , 717 murtagh , f. , starck , j. -l . ,
bijaoui , a. 1995 , a&as , 112 , 179 odenkirchen , m. , brosche , p. , geffert , m. , tucholke , h - j . , 1997 , new astronomy , 2 , 477 odenkirchen , m. 1998 , private communication oh , k.s . , lin , d.n.c
. 1992 , apj , 386 , 519 ostriker , j.p . &
caldwell , j. a. 1983 , in kinematics , dynamics and structure of the milky way , ed . w. l. h. shuter ( dordrecht : reidel ) , 24 pfenniger , d. , combes , f. , martinet , l. 1994 , aa , 285 , 79 portegies zwart , s. f. , hut , p. , verbunt , f. 1997 , a&a , 328 , 130 vesperini , e. , heggie , d. c. 1997 , mnras , 289 , 898 pryor , c. , meylan , g. 1993 , in `` structure and dynamics of globular clusters '' djorgovski & meylan ( eds ) , a.s.p .
vol 50 saviane , i. , piotto , g. , fagotto , f. , zaggia , s. , cappacioli , m. , aparicio , a. 1998 , a&a , 333 , 479 scholz , r. -d . , irwin , m. , odenkirchen , m. , meusinger , h. 1998 , a&a , 333 , 531 slezak , e. , durret , f. , gerbal , d. 1994 , aj , 108 , 1996 schweitzer , a. e. , cudworth , k. m. , majewski , s. r. 1993 , `` the globular cluster - galaxy connection .
globular clusters within the context of their parent galaxies '' , asp , p. 113
spitzer , l.j .
1958 , apj , 127 , 17 spitzer , l.j . ,
hart , m.h .
1971 , apj , 166 , 483 spitzer , l.j .
, chevalier r.a .
1973 , 183 , 565 starck , j .-
l , siebenmorgen , r. , gredel , r. 1997 , apj , 482 , 1011 trager s.c .
, king i.r . , djorgovski s. 1995 , aj , 109 , 218 vesperini , e. heggie , d. c. 1997 , mnras , 289 , 898 webbink , r. f. 1981 , apjs , 45 , 259 weinberg , m.d .
1994 , aj , 108 , 1414 zaggia , s.r .
, piotto , g. , capaccioli , m. 1997 , a&a , 324 , 1004 zoccali , m. , piotto , g. , zaggia , s. r. , capaccioli , m. 1998 , a&a , 331 , 541 | large - field multi - color images of 20 galactic globular clusters are used to investigate the presence of tidal tails around these stellar systems . field and
cluster stars are sorted with the help of color - magnitude diagrams , and star - count analysis is performed on the selected cluster stars in order to increase the signal - to - noise ratio of their surface density .
we study the overdensities of these stars using the wavelet transform of the star counts in order to filter the background density noise and to detect the weak structures , at large scale , formed by the numerous stars previously members of the clusters .
we associate these stellar overdensities with the stars evaporated from the clusters because of dynamical relaxation and/or tidal stripping from the clusters by the galactic gravitational field .
we take into account the strong observational biases induced by the clustering of galactic field stars and of background galaxies , along with the fluctuations of the background due to dust extinction .
most of the globular clusters in our sample display strong evidence of tidal interactions with the galactic plane in the form of large and extended deformations .
these tidal tails exhibit projected directions preferentially towards the galactic center .
all the clusters observed , which do not suffer from strong observational biases , present such tidal tails , tracing their dynamical evolution ( evaporation , tidal shocking , tidal torquing , and bulge shocking ) in the galaxy .
the clusters exhibit different regimes of mass loss rate , detected using the radial density slope in the outer parts of the clusters .
for ngc 5139 ( @xmath0 centauri ) , we estimate , taking into account the possible presence of mass segregation in its outer parts , that about 0.6 to 1 % of its mass has been lost during the current disk shocking event . in the case of ngc 6254
, we tentatively estimate , in the cluster reference frame , for the radial diffusion velocity of the stars stripped , a value of the order of the velocity dispersion in the cluster itself .
the sizes and orientations of these observed tidal tails are perfectly reproduced by n - body simulations of globular clusters in the galactic potential well .
we present these results in a companion paper ( combes , leon & meylan 1999 ) . as a by - product of this study
, we detect several new galaxy clusters towards the different fields studied at high galactic latitude .
the estimation of the tidal radius of some of the globular clusters could have been overestimated because of these galaxy clusters . |
one of the strongest motivations for physics beyond the standard model ( sm ) is the observational evidence for gravitationally interacting non - baryonic dark matter ( dm ) @xcite .
if dark matter is permitted interactions with the sm , as predicted by many sm extensions , then possible signatures of these interactions could arise in direct , indirect and collider dark matter searches .
furthermore , if the scale of the new physics mediating the dark matter - standard model interactions is large relative to the energies relevant to the various dark matter searches , then this ultraviolet ( uv ) scale can be integrated out of the full lagrangian theory and we can relate potential dark matter signatures using the contact operator approximation in a generalized effective field theory ( eft ) @xcite .
a well motivated class of dark matter candidates comes from models with light dark matter ( ldm ) .
as opposed to the more ubiquitous weakly interacting massive particles ( wimps ) , ldm particles have masses typically below the weak scale and , thus , could produce nuclear recoils at direct detection experiments with energies near or below threshold for detection .
assuming the contact operator approximation is valid , we can use eft to relate complementary bounds on dark matter - standard model interactions from monojet searches at high energy colliders @xcite to limits on ldm scattering off of nuclei , possibly constraining dark matter - nucleon interactions for dark matter masses beyond the reach of direct detection . in a similar fashion , searches for invisible
bound state decays at high luminosity colliders can constrain ldm interactions with the sm independent of any particular uv physics model .
previous studies have considered a variety of invisible bound state decays @xcite and , in a related work , the authors of @xcite have thoroughly explored the complementary aspects of @xmath1 and @xmath2 decays to purely invisible final states . in this work ,
we explore constraints from the decay @xmath3 and explore the possible complementarity of searches between different bound state decays , as well as the relationship between disparate dark matter detection strategies .
in particular , bound state quarkonium decays to @xmath4 are , at quark level in the matrix element , identical to monophoton searches .
however , in the non - relativistic limit , many of the possible dm - sm interaction structures can be constrained by a particular combination of bound state meson decays to @xmath5 final states and , thus , could possibly offer an important complement to monophoton searches . for simplicity and due to the relative lack of data available for other choices of initial state
, we exclusively calculate constraints from the decays of @xmath1 . in principle
, we will see that constraints from different combinations of initial bound states and decay channels ( i.e. , @xmath6 or @xmath7 ) can constrain the same interaction structures . results for @xmath8 decays into final states with scalar dark matter were also presented in @xcite , although the constraints on dark matter - quark contact operators from radiative decays were approximated with limits from @xmath9 decays assuming an on - shell mediator @xcite . in this paper
, we use the limit on @xmath10 decays for a particular dm - sm interaction structure and dark matter spin to obtain limits on all contact operators , of dimension six or lower , coupling scalar , fermion or vector ldm to bottomonium . in section ii , we discuss the current experimental constraints on bound state decays to @xmath5 and on dark matter annihilation from observations of dwarf spheroidal galaxies , then calculate the relevant decay widths and annihilation cross sections , respectively . in section iii , we present the associated limits on the strength of the dm - sm couplings and compare to results from monojet dark matter searches and direct dark matter detection .
we assume the dark matter - quark interaction is mediated by a uv physics model with heavy degrees of freedom that can be integrated out of the lagrangian and , thus , can be parametrized as an effective four - point interaction , which can be written as lorentz - invariant contractions of quark and dark matter bilinear structures .
while the matrix element for quark / antiquark annihilation , ( @xmath11 ) , and the corresponding matrix element for bound state meson decay can , in principle , receive contributions from amplitudes with a photon emitted by our heavy mediator , these contributions will necessarily be suppressed by extra powers of the uv scale relative to diagrams with photons emitted by the initial state quarks @xcite .
we show the diagrams which yield the leading order contributions to the matrix elements for our bound state meson decays in figure [ fig : diagrams ] .
in contrast to the contact operators considered in purely invisible decays , the quark bilinears in the operators mediating decays to dark matter with a photon in the final state need not share the angular momentum and @xmath12 properties of the heavy quarkonium state .
assuming weak interactions are negligible , the charge conjugation transformation required of the quark bilinears in the dm - sm interaction structures is determined by the presence or absence of a photon in the final state . for the purely invisible decays of the @xmath1 with @xmath13 , the quark bilinear must be @xmath14 or @xmath15 , where @xmath16 is a spatial index @xcite . for @xmath17 decays ,
the final state photon is intrinsically @xmath18-odd and , thus , the associated quark bilinears in the operators mediating single - photon decays should be even under charge conjugation .
therefore , the quark bilinear for heavy quarkonium decays to @xmath7 must be @xmath19 , @xmath20 , @xmath21 or @xmath22 and the effective operators which can yield a nonvanishing matrix element for @xmath10 decays are orthogonal to the operators which allow @xmath23 decays .
when single - photon decays are considered in tandem with operators permitting purely invisible decays , all possible dm - sm interaction structures of dimension six or lower can be constrained by @xmath8 decays .
the interaction structures and [ tab : structsinv ] , we refer to spin-0 , spin-1/2 and spin-1 dark matter fields with @xmath24 , @xmath25 and @xmath26 , respectively . in all other contexts , unless specifically noted , a dark matter field of arbitrary spin will be denoted @xmath25 . ] permitting @xmath13 bound state decays to @xmath5 are listed in table [ tab : structs ] , along with the angular momentum and @xmath12 properties of other possible bound states with non - vanishing matrix elements for decays to @xmath6 or @xmath7 final states and other dark matter searches , here we note the relationship between the constraints arising from a variety of bound state decays . in the appendix
, we update the constraints from purely invisible decays and calculate the relevant branching fractions and cross sections for the v7@xmath27 and v9@xmath27 operators not considered in previous work .
+ + + ] .
if the interaction structures can permit @xmath28-wave dark matter annihilation , then a bound can also be set by fermi observations of photons originating from dwarf spheroidal galaxies .
lastly , we indicate whether or not constraints on spin - independent ( si ) or spin - dependent ( sd ) nucleon scattering can be inferred from babar or fermi limits
. .interaction structures that can mediate @xmath1 decays to @xmath29 . invisible ( radiative ) identifies the @xmath30 bound states that can be annihilated for @xmath31 ( @xmath29 ) final states .
note that we only consider @xmath28- and @xmath32-wave bound states .
if the interaction structures can permit @xmath28-wave dark matter annihilation , then a bound can also be set by fermi observations of photons originating from dwarf spheroidal galaxies .
lastly , we indicate whether or not constraints on spin - independent ( si ) or spin - dependent ( sd ) nucleon scattering can be inferred from babar or fermi limits . [ cols="^,^,^,^,^,^",options="header " , ]
as noted in the text , the various dm - sm interaction structures will have complementarity constraints due to decays of different bound states to invisible final states .
for example , assuming fermionic dark matter interacts through our f5 operator , which arises from integrating out a vector mediator , constraints from @xmath33 decays will be directly related @xmath34 , @xmath35 and @xmath36 decays . as a result , between decays to @xmath31 and @xmath29 final states , every low - lying bottomonium state can be used to constrain a variety of dark matter interaction structures . as pointed out by the authors of @xcite , a complete set of dimension 6 or lower
contact operators mediating spin-1 dark matter interactions with quarks should include @xmath18-odd v(7 - 10)@xmath27 interaction structures in addition to the @xmath18-even v(7 - 10)@xmath37 operators considered in @xcite .
we have included both groups of operators in our analysis of bound state decays to @xmath38 in this work .
we calculate the branching fractions for bound state decays to @xmath31 through the v7@xmath27 and v9@xmath27 operators not considered in @xcite , _
v7_- ( |x x ) & = & b ( e^+ e^- ) m^4 512 ^2 ^2 q_b^2 ^4 ( 1 - 4 m_x^2 m^2 ) ^3/2 m^2 m_x^2 ( 1 + m^2 4 m_x^2 ) , + b_v9_- ( |x x ) & = & b ( e^+ e^- ) m^4 128 ^2 ^2 q_b^2 ^4 ( 1 - 4 m_x^2 m^2 ) ^1/2 ( 1 + m^2 2 m_x^2 ) . note that , while the v(7 - 10)@xmath37 operators allow for the dark matter fields to be real or complex , the v(7 - 10)@xmath27 operators are only non - vanishing if we assume complex dark matter .
the v7@xmath37 and v9@xmath37 branching fractions to @xmath31 calculated in @xcite also assumed operators with complex dark matter , albeit with an extra factor of 2 in the normalization of the operators . for completeness
, we also calculate the dark matter annihilation cross section for the v9@xmath27 interaction structure , _
a^v9_- v & = & m_x^2 3 ^4 ( 1 - m_q^2 m_x^2 ) ^1/2 ( 1 + m_q^2 2 m_x^2 ) .
we calculate the differential branching fractions for @xmath10 , with respect to photon energy and scattering angle , assuming polarized @xmath1 produced in @xmath39 transitions . for relativistic @xmath40 annihilating through a photon , the resulting @xmath41
is produced polarized with its spin axis lying along the beam line .
we assume the dipion transition is dominated by @xmath42 gluon radiation , thus the daughter @xmath1 should also be polarized along the beam line .
we also assume a negligible differential boost between the rest frame of the @xmath41 and that of the @xmath1 . the polarization of the @xmath1 and the lack of relative angular momentum between the @xmath1 and the dipion system have been confirmed in the angular distributions of charged leptons produced in subsequent @xmath1 decays and in the dipion system , respectively @xcite .
we can define the kinematics of the @xmath43 decay , where @xmath25 now denotes the invisible system consisting of our dark matter particles .
since we are considering the rest frame of the @xmath41 , and , to a reasonable approximation , that of the @xmath1 , we define @xmath44
. the photon will be emitted at an angle , @xmath45 , relative to the beam line with an energy , @xmath46 , yielding @xmath47 .
finally , we only consider @xmath1 polarizations with spin projections on the beam axis , @xmath48 .
now we can calculate the matrix elements squared for our polarized @xmath1 decays , given the relevant dm - sm interaction structures , and then integrate over the full phase space of the invisible system , leaving phase space distributions only in our observable kinematic variables .
we write the differential branching fractions using abbreviated notation for some of the phase space factors , ( a , b , c ) & = & a^2 + b^2 + c^2 - 2 a b - 2 ac - 2 b c , with @xmath49 and @xmath50 .
we also denote the possible angular dependencies for @xmath10 decays in our framework , @xmath51 and @xmath52 . the following are the polarized differential branching fractions as functions of dark matter mass , @xmath1 mass , mediation scale , the branching fraction to @xmath53 , and phase space factors : & = & 3 b ( e^+ e^- ) m^2 512 ^3 ^6 _ ^1/2 _
x^3/2 x^2 f_^+ , + d b_f2,f4^pol ( |x x ) d x^2 d & = & 3 b ( e^+ e^- ) m^2 512 ^3 ^6 _ ^1/2 _ x^1/2 x^2 f_^+ , + d b_f7^pol ( |x x ) d x^2 d & = & 3 b ( e^+ e^- ) 128 ^3 ^4 _ ^1/2 , + d b_f8^pol ( |x x ) d x^2 d & = & b ( e^+ e^- ) 64 ^3 ^4 _ ^1/2 , + d b_s1,s2^pol ( |x x ) d x^2 d & = & 3 b ( e^+ e^- ) m^2 1024 ^3 ^4 _ ^1/2 _
x^1/2 f_^+ , + d b_s4^pol ( |x x ) d x^2 d & = & b ( e^+ e^- ) m^2 1024 ^3 ^4 _ ^1/2 _ x^3/2 , + d b_v1,v2^pol ( |x x ) d x^2 d & = & 3 b ( e^+ e^- ) m^2 1024 ^3 ^4 _ ^1/2 _ x^1/2 f_^+ , + d b_v4^pol ( |x x ) d x^2 d & = & b ( e^+ e^- ) m^2 1024 ^3 ^4 _ ^1/2 _ x^3/2 , + d b_v8_+^pol ( |x x ) d x^2 d & = & b ( e^+ e^- ) 4096 ^3 ^4 _ ^1/2 _ x^3/2 , + d b_v8_-^pol ( |x x ) d x^2 d & = & b ( e^+ e^- ) m^2 1024 ^3 ^4 _ ^1/2 _ x^3/2 x^2 m_x^2 , + d b_v10_+^pol ( |x x ) d x^2 d & = & b ( e^+ e^- ) 1024 ^3 ^4 _ ^1/2 , + d b_v10_-^pol ( |x x ) d x^2 d & = & b ( e^+ e^- ) m^2 512 ^3 ^4 _ ^1/2 _ x^1/2 . note if we integrate these differential branching fractions over the entire kinematic range , which is equivalent to assuming an isotropically polarized bound state , we recover the fully integrated branching fractions reported in eq .
[ eqn : fullwidths ] .
we show the analytic results of the relevant integrals over the phase space of the invariant mass squared of the invisible system , @xmath54 , noting the change of integration variables to @xmath55 .
the integrals are all of the form i_n^m ( x ) & = & _ 1^x ( 1 - 1 x ) ^n x^ m dx , where @xmath56 and @xmath57 is an integer such that @xmath58 . for calculation of the partially integrated branching fractions , the allowed range of @xmath54
is determined by the photon detection threshold , @xmath59 , as well as the dark matter mass , such that @xmath60 .
the full branching fractions are integrated over the entire kinematic range with @xmath61 .
i_1/2 ^ -1 ( x ) & = & - 2 + i_1/2 ^ 0 ( x ) & = & + i_1/2 ^ 1 ( x ) & = & 1 4 + i_1/2 ^ 2 ( x ) & = & 1 24 + i_1/2 ^ 3 ( x ) & = & 1 192 + i_1/2 ^ 4 ( x ) & = & 1 1920 + i_3/2 ^ -1 ( x ) & = & 2 + i_3/2 ^ 0 ( x ) & = & + i_3/2 ^ 1 ( x ) & = & 1 4 + i_3/2 ^ 2 ( x ) & = & 1 24 + i_3/2 ^ 3 ( x ) & = & 1 64 + i_3/2 ^ 4 ( x ) & = & 1 640 note that @xmath62 for all @xmath63 and @xmath57 , thus the integrals need only be evaluated at @xmath64
. 99 e. kolb and m. turner , front . phys . * 69 * , 1 - 547 ( 1990 ) .
m. beltran , d. hooper , e. w. kolb and z. c. krusberg , phys .
d * 80 * , 043509 ( 2009 ) [ arxiv:0808.3384 [ hep - ph ] ] .
q. -h .
cao , c. -r .
chen , c. s. li and h. zhang , jhep * 1108 * , 018 ( 2011 ) [ arxiv:0912.4511 [ hep - ph ] ] .
j. goodman , m. ibe , a. rajaraman , w. shepherd , t. tait and h. yu , phys . lett .
b * 695 * , 185 ( 2011 ) [ arxiv:1005.1286 [ hep - ph ] ] .
a. rajaraman , t. m. p. tait and a. m. wijangco , phys .
dark univ .
* 2 * , 17 ( 2013 ) [ arxiv:1211.7061 [ hep - ph ] ] .
h. dreiner , d. schmeier and j. tattersall , europhys .
lett . * 102 * , 51001 ( 2013 ) [ arxiv:1303.3348 [ hep - ph ] ] .
j. kumar and d. marfatia , phys .
d * 88 * , 014035 ( 2013 ) [ arxiv:1305.1611 [ hep - ph ] ] .
g. busoni , a. de simone , e. morgante , and a. riotto , phys .
b * 728 * , 412 ( 2014 ) [ arxiv:1307.2253 [ hep - ph ] ] .
a. difranzo , k. i. nagao , a. rajaraman and t. m. p. tait , jhep * 1311 * , 014 ( 2013 ) [ arxiv:1308.2679 [ hep - ph ] ] . m. r. buckley , phys . rev .
d * 88 * , 055028 ( 2013 ) [ arxiv:1308.4146 [ hep - ph ] ] .
o. buchmueller , m. j. dolan , and c. mccabe , jhep * 1401 * , 025 ( 2014 ) [ arxiv:1308.6799 [ hep - ph ] ]
. g. busoni , a. de simone , j. gramling , e. morgante , and a. riotto , [ arxiv:1402.1275 [ hep - ph ] ] .
a. alves , s. profumo , f. s. queiroz , and w. shepherd , [ arxiv:1403.5027 [ hep - ph ] ] . m. a. fedderke , j .- y .
chen , e. w. kolb , and l .- t .
wang , [ arxiv:1404.2283 [ hep - ph ] ] . f. deramo and m. procura , jhep * 1504 * , 054 ( 2015 ) doi:10.1007/jhep04(2015)054 [ arxiv:1411.3342 [ hep - ph ] ] .
a. birkedal , k. matchev and m. perelstein , phys .
d * 70 * , 077701 ( 2004 ) [ arxiv : hep - ph/0403004 ] .
j. l. feng , s. su and f. takayama , phys .
* 96 * , 151802 ( 2006 ) [ hep - ph/0503117 ] .
m. beltran , d. hooper , e. w. kolb , z. c. krusberg and t. m. p. tait , jhep * 1009 * , 037 ( 2010 ) [ arxiv:1002.4137 [ hep - ph ] ] .
y. bai , p. j. fox and r. harnik , jhep * 1012 * , 048 ( 2010 ) [ arxiv:1005.3797 [ hep - ph ] ] .
j. goodman , m. ibe , a. rajaraman , w. shepherd , t. tait and h. yu , phys .
d * 82 * , 116010 ( 2010 ) [ arxiv:1008.1783 [ hep - ph ] ] .
a. rajaraman , w. shepherd , t. m. p. tait and a. m. wijangco , phys .
d * 84 * , 095013 ( 2011 ) [ arxiv:1108.1196 [ hep - ph ] ] .
p. j. fox , r. harnik , j. kopp and y. tsai , phys .
d * 85 * , 056011 ( 2012 ) [ arxiv:1109.4398 [ hep - ph ] ] .
y. bai , a. rajaraman , [ arxiv : 1109.6009 [ hep - ph ] ] .
j. goodman and w. shepherd , [ arxiv:1111.2359 [ hep - ph ] ] . y. bai and t. m. tait , phys .
b * 723 * , 384 ( 2013 ) [ arxiv:1208.4361 [ hep - ph ] ] .
p. agrawal and v. rentala , [ arxiv:1312.5325 [ hep - ph ] ] . m. papucci , a. vichi , and k. m. zurek , [ arxiv : 1402.2285 [ hep - ph ] ] .
j. kumar , d. marfatia and d. yaylali , arxiv:1508.04466 [ hep - ph ] .
r. essig , j. mardon , m. papucci , t. volansky and y. zhong , [ arxiv:1309.5084 [ hep - ph ] ] .
p. fayet , phys .
d * 75 * , 115017 ( 2007 ) [ hep - ph/0702176 ] .
p. fayet , phys .
d * 81 * , 054025 ( 2010 ) [ arxiv:0910.2587 [ hep - ph ] ] .
b. mcelrath , [ arxiv:0712.0016 [ hep - ph ] ] .
r. cotta , a. rajaraman , t. tait and a. wijangco , [ arxiv:1305.6609 [ hep - ph ] ] . k. schmidt - hoberg , f. staub and m. w. winkler , phys .
b * 727 * , 506 ( 2013 ) [ arxiv:1310.6752 [ hep - ph ] ] .
a. badin and a. petrov , [ arxiv:1005.1277 [ hep - ph ] ] .
d. mckeen , phys .
d * 79 * , 114001 ( 2009 ) [ arxiv:0903.4982 [ hep - ph ] ] .
n. fernandez , j. kumar , i. seong and p. stengel , phys .
d * 90 * , no . 1 , 015029 ( 2014 ) [ arxiv:1404.6599 [ hep - ph ] ] . g. yeghiyan , phys .
d * 80 * , 115019 ( 2009 ) [ arxiv:0909.4919 [ hep - ph ] ] .
b. aubert _ et al . _ [ babar collaboration ] , [ arxiv:0808.0017 [ hep - ex ] ] .
j. beringer _
( particle data group ) phys .
d * 86 * , 010001 ( 2012 ) .
r. balest _ et al . _
[ cleo collaboration ] , phys .
d * 51 * , 2053 ( 1995 ) .
del amo sanchez _ et al . _
[ babar collaboration ] , phys .
lett . * 107 * , 021804 ( 2011 ) [ arxiv:1007.4646 [ hep - ex ] ] .
r. essig , n. sehgal and l. strigari , phys .
d * 80 * , 023506 ( 2009 ) [ arxiv:0902.4750 [ hep - ph ] ] .
a. geringer - sameth and s. koushiappas , phys .
lett . * 107 * , 241303 ( 2011 ) [ arxiv:1108.2914 [ astro-ph.co ] ] .
a. geringer - sameth and s. koushiappas , phys .
d * 86 * , 021302(r ) ( 2012 ) [ arxiv:1206.0796 [ astro-ph.he ] ] .
the fermi - lat collaboration , phys .
lett . * 107 * , 241302 ( 2011 ) [ arxiv:1108.3546[astro - ph.he ] ] .
the fermi - lat collaboration , phys .
d * 89 * , 042001 ( 2014 ) [ arxiv:1310.0828 [ astro-ph.he ] ] .
m. ackermann _ et al .
_ [ fermi - lat collaboration ] , phys . rev .
lett . * 115 * , no .
23 , 231301 ( 2015 ) doi:10.1103/physrevlett.115.231301 [ arxiv:1503.02641 [ astro-ph.he ] ] . m. g. baring , t. ghosh , f. s. queiroz and k. sinha , arxiv:1510.00389 [ hep - ph ] .
t. sjstrand , s. mrenna and p. skands , jhep * 0605 * , 026 ( 2006 ) , [ arxiv : hep - ph/0603175 ] . j. f. gunion , d. hooper and b. mcelrath , phys .
d * 73 * , 015011 ( 2006 ) [ hep - ph/0509024 ] .
u. haisch and f. kahlhoefer , jcap * 1304 * , 050 ( 2013 ) [ arxiv:1302.4454 [ hep - ph ] ] .
a. crivellin , f. deramo and m. procura , phys .
lett . * 112 * , 191304 ( 2014 ) [ arxiv:1402.1173 [ hep - ph ] ] .
g. aad _ et al .
_ [ atlas collaboration ] , eur .
j. c * 75 * , no . 7 , 299 ( 2015 ) [ eur . phys .
j. c * 75 * , no .
9 , 408 ( 2015 ) ] [ arxiv:1502.01518 [ hep - ex ] ] .
g. aad _ et al . _ [ atlas collaboration ] , phys . rev .
d * 91 * , no . 1
, 012008 ( 2015 ) [ arxiv:1411.1559 [ hep - ex ] ] . v. khachatryan _ et al . _
[ cms collaboration ] , eur .
j. c * 75 * , no . 5 , 235 ( 2015 ) [ arxiv:1408.3583 [ hep - ex ] ] .
v. khachatryan _ et al . _
[ cms collaboration ] , arxiv:1410.8812 [ hep - ex ] .
u. haisch and e. re , jhep * 1506 * , 078 ( 2015 ) doi:10.1007/jhep06(2015)078 [ arxiv:1503.00691 [ hep - ph ] ] .
o. buchmueller , m. j. dolan , s. a. malik and c. mccabe , jhep * 1501 * , 037 ( 2015 ) doi:10.1007/jhep01(2015)037 [ arxiv:1407.8257 [ hep - ph ] ] .
c. kelso , j. kumar , p. sandick and p. stengel , phys .
d * 91 * , 055028 ( 2015 ) [ arxiv:1411.2634 [ hep - ph ] ] .
j. m. alarcon , j. martin camalich and j. a. oller , phys .
d * 85 * , 051503 ( 2012 ) doi:10.1103/physrevd.85.051503 [ arxiv:1110.3797 [ hep - ph ] ] .
j. m. alarcon , l. s. geng , j. martin camalich and j. a. oller , phys .
b * 730 * , 342 ( 2014 ) doi:10.1016/j.physletb.2014.01.065 [ arxiv:1209.2870 [ hep - ph ] ] .
m. hoferichter , j. ruiz de elvira , b. kubis and u. g. meiner , phys .
lett . * 115 * , no . 9 , 092301 ( 2015 ) doi:10.1103/physrevlett.115.092301 [ arxiv:1506.04142 [ hep - ph ] ] .
j. ellis , k. a. olive and p. sandick , new j. phys . * 11 * , 105015 ( 2009 ) [ arxiv:0905.0107 [ hep - ph ] ] .
g. angloher _ et al . _
[ cresst collaboration ] , arxiv:1509.01515 [ astro-ph.co ] .
r. agnese et al . , [ arxiv:1402.7137 [ hep - ex ] ] .
d. s. akerib _ et al . _
[ lux collaboration ] , arxiv:1310.8214 [ astro-ph.co ] .
c. amole _ et al . _ [ pico collaboration ] , phys .
lett . * 114 * , no .
23 , 231302 ( 2015 ) [ arxiv:1503.00008 [ astro-ph.co ] ] .
s. archambault _ et al .
_ [ picasso collaboration ] , phys .
b * 711 * , 153 ( 2012 ) [ arxiv:1202.1240 [ hep - ex ] ] . c. savage , g. gelmini , p. gondolo and k. freese , jcap * 0904 * , 010 ( 2009 ) [ arxiv:0808.3607 [ astro - ph ] ] .
g. angloher , m. bauer , i. bavykina , a. bento , c. bucci , c. ciemniak , g. deuter and f. von feilitzsch _
et al . _ ,
j. c * 72 * , 1971 ( 2012 ) [ arxiv:1109.0702 [ astro-ph.co ] ] . c. e. aalseth _
et al . _ ,
arxiv:1401.6234 [ astro-ph.co ] .
r. agnese _ et al . _
[ cdms collaboration ] , phys .
* 111 * , 251301 ( 2013 ) [ arxiv:1304.4279 [ hep - ex ] ] . | we investigate constraints on the interactions of light dark matter with standard model quarks in a framework with effective contact operators mediating the decay of heavy flavor bound state quarkonium to dark matter and a photon .
when considered in combination with decays to purely invisible final states , constraints from heavy quarkonium decays at high intensity electron - positron colliders can complement missing energy searches at high energy colliders and provide sensitivity to dark matter masses difficult to probe at direct and indirect detection experiments .
we calculate the approximate limits on the branching fraction for @xmath0 decays to dark matter and a photon . given the approximate limits on the branching fractions for all dimension 6 or lower contact operators , we present the corresponding limits on the interaction strength for each operator and the inferred limits on dark matter - nucleon scattering .
complementary constraints on dark matter annihilation from gamma - ray searches from dwarf spheroidal galaxies are also considered . |
let @xmath0 denote the set of all fillings of a @xmath1 rectangular array with the integers @xmath9 such that that the elements increase from bottom to top in each column .
we let @xmath10 denote the cell in @xmath6-th row from the bottom and @xmath11-th column from the left of the @xmath1 rectangle and we let @xmath12 denote the element in cell @xmath10 of @xmath13 . for example
, the elements of @xmath14 are pictured below .
it is easy to see that latexmath:[\[\label{number } @xmath5 , allowing all permutations of the elements in each column gives rise to @xmath16 fillings of the @xmath1 rectangle with the numbers @xmath2 . since there are @xmath17 fillings of the @xmath1 rectangle with the numbers @xmath2 , ( [ number ] ) easily follows . given a partition @xmath18 where @xmath19 , we let @xmath20 denote the ferrers diagram of @xmath21 , i.e. @xmath20 is the set of left - justified rows of squares where the size of the @xmath6-th row , reading from top to bottom , is @xmath22 .
thus a @xmath1 rectangular array corresponds to the ferrers diagram corresponding to @xmath23 . if @xmath5 and the integers are increasing in each row , reading from left to right , then @xmath7 is a standard tableau of shape @xmath23 .
we let @xmath24 denote the set of all standard tableaux of shape @xmath23 and let @xmath25 .
one can use the frame - robinson - thrall hook formula @xcite to show that @xmath26 where @xmath27 and @xmath28 for @xmath29 .
the goal of this paper is to study pattern matching conditions in @xmath0 .
clearly , when @xmath30 , we have @xmath31 , where @xmath32 is the symmetric group , so our results can be viewed as generalizations of results on pattern matching in @xmath32 .
our long term goal is to extend the work of this paper to the study pattern matching conditions in standard tableaux of rectangular shape as well as pattern matching conditions in column strict tableaux of rectangular shapes over a fixed alphabet @xmath33 .
the study of pattern matching conditions in @xmath0 is considerably easier than the study of pattern matching conditions in standard tableaux or column strict tableax .
nevertheless , we shall see that the study of pattern matching conditions in @xmath0 requires us to prove some interesting results on pattern matching conditions in standard tableaux .
in particular , we shall generalize two classical results on permutations . given a permutation @xmath34 ,
we let @xmath35 we let @xmath36 and @xmath37 .
then the generating function for the distribution of rises or descents in @xmath32 is given by @xmath38 see stanley @xcite .
a permutation @xmath39 is alternating if @xmath40 or , equivalently , if @xmath41 equals the set of odd numbers which are less than @xmath42 .
let @xmath43 denote the number of alternating permutations of @xmath32 .
then andr @xcite , @xcite proved that @xmath44 and @xmath45 if @xmath7 is any filling of a @xmath1-rectangle with distinct positive integers such that elements in each column increase , reading from bottom to top , then we let @xmath46 denote the element of @xmath0 which results from @xmath7 by replacing the @xmath6-th smallest element of @xmath7 by @xmath6 . for example , figure [ figure : red ] demonstrates a filling , @xmath7 , with its corresponding reduced filling , @xmath46 . if @xmath5 and @xmath47 , then we let @xmath48 $ ] be the filling of the @xmath49 rectangle where the elements in column @xmath50 of @xmath48 $ ] equal the elements in column @xmath51 in @xmath7 for @xmath52
. we can then extend the usual pattern matching definitions from permutations to elements of @xmath0 as follows .
[ def1 ] let @xmath3 be an element of @xmath53 and @xmath5 where @xmath54 .
then we say 1 .
@xmath3 * occurs * in @xmath7 if there are @xmath55 such that @xmath56 ) } = p$ ] , 2 .
@xmath7 * avoids * @xmath3 if there is no occurrence of @xmath3 in @xmath7 , and 3 .
there is a * @xmath3-match in @xmath7 starting at position @xmath6 * if @xmath57 ) } = p$ ] .
again , when @xmath30 , then @xmath58 , where @xmath32 is the symmetric group , and our definitions reduce to the standard definitions that have appeared in the pattern matching literature .
we note that kitaev , mansour , and vella @xcite have studied pattern matching in matrices which is a more general setting than the one we are considering in this paper .
we let @xmath59 denote the number of @xmath3-matches in @xmath7 and let @xmath60 be the maximum number of nonoverlapping @xmath3-matches in @xmath7 , where two @xmath3-matches are said to overlap if they share a common column .
for example , if we consider the fillings @xmath61 and @xmath62 shown in figure [ figure : pmatch ] , then it is easy to see that there are no @xmath3-matches in @xmath7 but there is an occurrence of @xmath3 in @xmath7 , since @xmath63 ) } = p$ ] . also , there are 2 @xmath3-matches in @xmath64 starting at positions 1 and 2 , respectively , so @xmath65 and @xmath66
. one can easily extend these notions to sets of elements of @xmath53 .
that is , suppose that @xmath67 .
then @xmath5 has an @xmath68-match at place @xmath6 provided @xmath69 ) } \in \upsilon$ ] .
let @xmath70 and @xmath71 be the number of @xmath68-matches and nonoverlapping @xmath68-matches in @xmath7 , respectively . to generalize ( [ risgf ] ) , ( [ alte ] ) , and ( [ alto ] ) , one must first generalize the notion of a rise in a permutation @xmath72 . as one may view a rise in @xmath73 as a pattern match of the permutation @xmath74 , a natural analogue of a rise in @xmath0
is a pattern match of one or more patterns , @xmath75 , which have strictly increasing rows , i.e. a match of a standard tableaux of shape @xmath76 .
consequently , there are many analogues of rises in our setting . for example , when @xmath77 , the two standard tableaux of shape @xmath78 , which we denote @xmath79 and @xmath80 , are shown in figure [ figure:2standard ] .
thus , when considering the analogues of a rise in @xmath81 , we should study matches of the patterns @xmath79 and @xmath80 as well as matches of the set of patterns @xmath82 . if @xmath83 , then we define @xmath84 to be the set of @xmath5 with @xmath85 , i.e. the set of @xmath5 with the property that there are @xmath3-matches in @xmath7 starting at positions @xmath86 .
we let @xmath87 , and by convention , define @xmath88 . for example , if @xmath3 is the element of @xmath89 that has the elements @xmath90 in the first column and the elements @xmath91 in the second column , then it is easy to see that @xmath92 for all @xmath93 since the only element of @xmath5 with @xmath94 has the entries @xmath95 in the @xmath6-th column for @xmath96 . more generally ,
if @xmath68 is a subset of @xmath89 , then we define @xmath97 to be the set of @xmath5 such that @xmath98 and let @xmath99 .
again , we use the convention that @xmath100 .
for example , if @xmath101 is the set of all standard tableaux of shape @xmath4 , then it easy to see that a @xmath5 has @xmath98 if and only if @xmath7 is standard tableaux of shape @xmath23 , so @xmath102 if @xmath68 is a subset of @xmath89 , then we shall be interested in the following three generating functions : @xmath103 where @xmath104 is equal to the number of @xmath5 that have no @xmath68-matches .
if @xmath68 consists of a single element @xmath3 , then we write @xmath105 , @xmath106 , and @xmath107 for @xmath108 , @xmath109 , and @xmath110 , respectively . clearly , the generating functions @xmath108 where @xmath111 can be viewed as analogues of the generating function for rises as described in ( [ risgf ] )
. we shall prove the following general theorems concerning the generating functions @xmath108 , @xmath109 , and @xmath110 . [ thm : d ] for all @xmath112 , @xmath113 [ thm : n ] for all @xmath112 , @xmath114 theorem [ thm : d ] is proved by applying a ring homomorphism defined on the ring , @xmath115 , of symmetric functions over infinitely many variables , @xmath116 , to a simple symmetric function identity .
there has been a long line of research , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , that shows that a large number of generating functions for permutation statistics can be obtained by applying homomorphisms defined on @xmath115 to simple symmetric function identities .
theorem [ thm :
n ] is an analogue of a result of kitaev @xcite . by our remarks
above , we have the following corollaries .
let @xmath117 denote the set of standard tableaux of shape @xmath4 , then @xmath118 and @xmath119 let @xmath3 be the element of @xmath117 that has @xmath90 in the first column , then @xmath120 and @xmath121 we have developed similar results for other 2-column patterns ; the key is to be able to compute @xmath122 .
for example , we can prove the following . 1 . if @xmath123 , then @xmath124 where @xmath125 is the @xmath42-th catalan number .
2 . if @xmath126 , then @xmath127 .
let @xmath128 denote the set of positive integers @xmath129 , @xmath130 denote the set of odd numbers in @xmath128 , and @xmath131 denote the set of even numbers in @xmath128 .
for @xmath132 , let @xmath133 = \{1 , \ldots , n\}$ ] , @xmath134 \cap \mathbb{o}$ ] , and @xmath135 \cap \mathbb{e}$ ] .
now suppose that @xmath68 is a subset of @xmath89 and @xmath5 .
then we let 1 .
@xmath136 denote the set of @xmath6 such that there is an @xmath68-match of @xmath7 starting at position @xmath6 , 2 .
@xmath137 denote the set of @xmath5 such that @xmath138 , and 3 .
@xmath139 denote the number of @xmath6 such that @xmath140 .
we say that @xmath5 is * @xmath68-alternating * if @xmath141 .
that is , @xmath5 is @xmath68-alternating if @xmath142 ) } \in \upsilon$ ] for all @xmath143 and @xmath144 ) } \not \in \upsilon$ ] for all @xmath145 .
we let @xmath146 denote the number of @xmath68-alternating elements of @xmath0 .
we shall prove the following general theorems about @xmath68-alternating fillings .
[ thm : eax ] let @xmath68 be a subset of @xmath89 .
then @xmath147 [ thm : oax ] let @xmath68 be a subset of @xmath89
. then @xmath148 theorems [ thm : eax ] and [ thm : oax ] are also proved by applying a ring homomorphism defined on @xmath115 to simple symmetric function identities . putting @xmath149 in theorems [ thm : eax ] and [ thm : oax ] , we obtain the following generating functions for @xmath68-alternating elements of @xmath0 which can be viewed as generalizations of ( [ alte ] ) and ( [ alto ] ) . @xmath150 @xmath151 the outline of this paper is the following . in section 2 , we shall review the necessary background on symmetric functions needed to prove theorems [ thm : d ] , [ thm : eax ] , and [ thm : oax ] . then in section 3
, we shall give the proofs of theorems [ thm : d ] , [ thm : eax ] , and [ thm : oax ] .
in section 4 , we shall give the proof of theorem [ thm : n ] . in section 5 ,
we shall find formulas for @xmath152 where @xmath68 are subsets of the standard tableaux of shape @xmath78 and we shall completely classify all standard tableaux , @xmath153 , such that @xmath154 for all @xmath93 . in addition , we shall briefly outline some methods that we have used to compute @xmath155 for certain standard tableaux @xmath156 where @xmath157 .
in this section we give the necessary background on symmetric functions needed for our proofs of theorems [ thm : d ] , [ thm : eax ] , and [ thm : oax ] .
we shall consider the ring of symmetric functions , @xmath115 , over infinitely many variables @xmath158 .
the homogeneous symmetric functions , @xmath159 , and elementary symmetric functions , @xmath160 , are defined by the generating functions @xmath161 the @xmath42-th power symmetric function , @xmath162 , is defined as @xmath163 let @xmath164 be an integer partition ; that is , @xmath165 is a finite sequence of weakly increasing non - negative integers .
let @xmath166 denote the number of nonzero integers in @xmath165 .
if the sum of these integers is @xmath42 , we say that @xmath165 is a partition of @xmath42 and write @xmath167 . for any partition @xmath168 ,
define @xmath169 , @xmath170 , and @xmath171 .
the well - known fundamental theorem of symmetric functions , see @xcite , says that @xmath172 is a basis for @xmath173 , the space of symmetric functions which are homogeneous of degree @xmath42 .
equivalently , the fundamental theorem of symmetric functions states that @xmath174 is an algebraically independent set of generators for the ring @xmath115 .
it follows that one can completely specify a ring homomorphism @xmath175 from @xmath115 into a ring @xmath176 by giving the values of @xmath177 for @xmath178 .
next we give combinatorial interpretations to the expansion of @xmath179 in terms of the elementary symmetric functions .
given partitions @xmath180 and
@xmath181 , a @xmath21-brick tabloid of shape @xmath182 is a filling of the ferrers diagram of shape @xmath182 with bricks of size @xmath183 such that each brick lies in single row and no two bricks overlap .
for example , figure [ figure : labrick ] shows all the @xmath165-brick tabloids of shape @xmath182 where @xmath184 and @xmath185 .
let @xmath186 denote the set of all @xmath165-brick tabloids of shape @xmath182 and let @xmath187 .
eeciolu and remmel proved in @xcite that @xmath188 if @xmath189 is a brick tabloid of shape @xmath190 such that the lengths of the bricks , reading from left to right , are @xmath191 , then we shall write @xmath192 .
for example , the brick tabloid @xmath193 is pictured in figure [ figure : btab ] .
next we define a symmetric function @xmath194 whose relationship with @xmath195 , which is very similar to the relationship between @xmath196 and @xmath195 , was first introduced in @xcite and @xcite .
let @xmath197 be a function which maps the set of non - negative integers into a field @xmath7 .
recursively define @xmath198 by setting @xmath199 and letting @xmath200 for all @xmath93 . by multiplying series
, this means that @xmath201 where the last equality follows from the definition of @xmath194 .
therefore , @xmath202 or , equivalently , @xmath203 if @xmath204 for all @xmath93 , then becomes @xmath205 which implies @xmath206 .
other special cases for @xmath197 give well - known generating functions .
for example , if @xmath207 for @xmath208 , then @xmath194 is the power symmetric function @xmath209 . by taking @xmath210 for some @xmath211 , @xmath212 is the schur function corresponding to the partition @xmath213 . here for any statement @xmath214 , we let @xmath215 if @xmath214 is true and @xmath216 if @xmath214 is false .
this definition of @xmath194 is desirable because of its expansion in terms of elementary symmetric functions .
the coefficient of @xmath217 in @xmath194 has a nice combinatorial interpretation similar to that of the homogeneous symmetric functions .
suppose @xmath189 is a brick tabloid of shape @xmath190 and type @xmath165 and that the final brick in @xmath189 has length @xmath218 .
define the weight of a brick tabloid @xmath219 to be @xmath220 and let @xmath221 when @xmath204 for @xmath93 , @xmath222 and @xmath223 are the same .
then mendes and remmel @xcite proved that @xmath224
in this section , we shall give the proofs of theorems [ thm : d ] , [ thm : eax ] , and [ thm : oax ] .
we start with the proof of theorem [ thm : d ] .
+ * proof of theorem [ thm : d ] .
* + our goal is to prove that for all @xmath112 , @xmath225 define a ring homomorphism @xmath226 $ ] , where @xmath227 $ ] is the polynomial ring over the rationals , by setting @xmath228 and @xmath229 for @xmath93
. then we claim that @xmath230 for all @xmath93 .
that is , @xmath231 next we want to give a combinatorial interpretation to ( [ ris2 ] ) .
first we interpret @xmath232 as the number of ways to pick a set partition of @xmath233 into sets @xmath234 where @xmath235 for @xmath236 .
for each set @xmath237 , we interpret @xmath238 as the number of ways to arrange the numbers in @xmath237 into a @xmath239 array such that there is an @xmath68-match starting at positions @xmath240 .
finally , we interpret @xmath241 as all ways of picking labels for all the columns of each brick , except the final column , with either an @xmath242 or a @xmath243 . for completeness , we label the final column of each brick with @xmath244 . we shall call all such objects created in this way _ filled labeled brick tabloids _ and let @xmath245 denote the set of all filled labeled brick tabloids that arise in this way .
thus a @xmath246 consists of a brick tabloid , @xmath189 , a filling , @xmath5 , and a labeling , @xmath247 , of the columns of @xmath189 with elements from @xmath248 such that 1 .
@xmath7 has an @xmath68-match starting at each column that is not a final column a brick , 2 .
the final column of each brick is labeled with @xmath244 , and 3 .
each column that is not a final column of a brick is labeled with @xmath242 or @xmath243 .
we then define the weight of @xmath249 , @xmath250 , to be the product of all the @xmath242 labels in @xmath247 and the sign of @xmath249 , @xmath251 , to be the product of all the @xmath243 labels in @xmath247 .
for example , if @xmath252 , @xmath253 , @xmath254 , and @xmath255 , then figure [ figure : fris1 ] demonstrates such a composite object @xmath256 where , @xmath257 and @xmath258 .
thus @xmath259 next we define a weight - preserving sign - reversing involution @xmath260 . to define @xmath261 , we scan the columns of @xmath262 from left to right looking for the leftmost column , @xmath263 , such that either ( i ) @xmath263 is labeled with @xmath243 or ( ii ) @xmath263 is at the end of a brick , @xmath264 , and the brick immediately following @xmath264 , namely @xmath265 , has the property that @xmath7 has an @xmath68-match starting in each column of @xmath264 and @xmath265 except , possibly , the last cell of @xmath265 . in case ( i ) , @xmath266 where @xmath267 is the result of replacing the brick @xmath268 in @xmath189 containing @xmath263 by two bricks , @xmath269 and @xmath270 , where @xmath269 contains the @xmath263-th column plus all the columns in @xmath268 to the left of @xmath263 and @xmath270 contains all the columns of @xmath268 to the right of @xmath263 , @xmath271 , and @xmath272 is the labeling that results from @xmath247 by changing the label of column @xmath263 from @xmath243 to @xmath244 . in case ( ii ) , @xmath273 where @xmath267 is the result of replacing the bricks @xmath264 and @xmath265 in @xmath189 by a single brick @xmath274 @xmath275 , and @xmath272 is the labeling that results from @xmath247 by changing the label of column @xmath263 from @xmath244 to @xmath243
. if neither case ( i ) or case ( ii ) applies , then we let @xmath276 .
for example , if @xmath249 is the element of @xmath277 pictured in figure [ figure : fris1 ] , then @xmath261 is pictured in figure [ figure : fris2 ] .
it is easy to see that @xmath278 for all @xmath279 and that if @xmath280 , then @xmath281 .
hence @xmath282 is a weight - preserving and sign - reversing involution that shows @xmath283 thus , we must examine the fixed points , @xmath284 , of @xmath282 .
first , there can be no @xmath243 labels in @xmath247 so that @xmath285 . moreover , if @xmath264 and @xmath265 are two consecutive bricks in @xmath189 and @xmath263 is the last column of @xmath264 , then it can not be the case that there is an @xmath68-match starting at position @xmath263 in @xmath7 since otherwise we could combine @xmath264 and @xmath265 .
it follows that @xmath286 .
for example , figure [ figure : fris3 ] shows a fixed point of @xmath282 in the case @xmath287 , @xmath253 , and @xmath288 .
vice versa , if @xmath5 , then we can create a fixed point , @xmath289 , by forming bricks in @xmath189 that end at columns that are not the start of an @xmath68-match in @xmath7 , labeling each column that is the start of an @xmath68-match in @xmath7 with @xmath242 , and labeling the remaining columns with @xmath244 .
thus we have shown that @xmath290 as desired .
applying @xmath291 to the identity @xmath292 , we get @xmath293 which proves ( [ eq : fris ] ) .
putting @xmath149 in ( [ eq : fris ] ) immediately yields the following corollary .
let @xmath112 and @xmath104 be the number of @xmath5 that have no @xmath68-matches . then @xmath294 recall that if @xmath295 , then @xmath137 denotes the set of @xmath5 that have @xmath68-matches starting at all the odd - numbered columns and that for @xmath296 , @xmath139 denotes the number of @xmath6 such that @xmath7 has an @xmath68-match starting at position @xmath297 .
next we give the proof of theorem [ thm : eax ] .
+ * proof of theorem [ thm : eax ] .
* + our goal is to prove that if @xmath68 is a subset of @xmath89 , then @xmath298 the proof of theorem [ thm : eax ] is very similar to the proof of theorem [ thm : d ] .
we only have to modify our ring homomorphism sightly .
that is , define a ring homomorphism @xmath299 $ ] by setting @xmath300 , @xmath301 for all @xmath178 , and @xmath302 for @xmath93 .
then we claim that for @xmath93 , @xmath303 and @xmath304 note that @xmath305 now if @xmath42 is odd , then every @xmath306 must have an odd part and hence our definitions force @xmath307 .
thus @xmath308 for all @xmath93 . on the other hand ,
if @xmath42 is even , then we can restrict ourselves to those partitions in ( [ key ] ) that contain only even parts .
for any partition @xmath309 of @xmath42 , we let @xmath310 denote the partition @xmath311 .
then for @xmath93 we have @xmath312 again we can give a combinatorial interpretation to ( [ 2ris2 ] ) .
first we interpret @xmath313 as the number of ways to pick a set partition of @xmath314 into sets @xmath234 where @xmath315 for @xmath236 .
for each set @xmath237 , we interpret @xmath316 as the number of ways to arrange the numbers in @xmath237 into a @xmath317 array such that there is an @xmath68-match starting at positions @xmath318 in the brick @xmath319 .
finally , we interpret @xmath241 as all ways of picking labels for all the even numbered columns of each brick , except the final even numbered column in the brick , with either an @xmath242 or a @xmath243 . for completeness , we label the final even numbered column of each brick with @xmath244 . we shall call all such objects created in this way _ filled labeled brick tabloids _ and let @xmath320 denote the set of all filled labeled brick tabloids that arise in this way .
thus a @xmath321 consists of a brick tabloid , @xmath189 , a filling , @xmath322 , and a labeling , @xmath247 , of the even numbered columns of @xmath189 with elements from @xmath248 such that 1 .
all bricks have even size , 2 .
@xmath7 has an @xmath68-match starting at each column that is not a final column of its brick , 3 .
the final column of each brick is labeled with @xmath244 , and 4 .
each even numbered column that is not a final column of a brick is labeled with @xmath242 or @xmath243 .
we then define the weight of @xmath249 , @xmath250 , to be the product of all the @xmath242 labels in @xmath247 and the sign of @xmath249 , @xmath251 , to be the product of all the @xmath243 labels in @xmath247 .
for example , if @xmath252 , @xmath253 , @xmath254 , and @xmath323 , then figure [ figure:2fris1 ] pictures such a filled labeled brick tabloid @xmath324 where @xmath325 and @xmath258 .
thus @xmath326 next we define a weight - preserving sign - reversing involution @xmath327 essentially in the same way as in theorem [ thm : d ] .
that is , to define @xmath261 , we scan the cells of @xmath262 from left to right looking for the leftmost column , @xmath263 , such that either ( i ) @xmath263 is labeled with @xmath243 or ( ii ) @xmath263 is at the end of a brick , @xmath319 , and the brick immediately following @xmath319 , namely @xmath328 , has the property that @xmath7 has a @xmath68-match starting in each column of @xmath319 and @xmath328 except , possibly , the last cell of @xmath328 . in case ( i ) , @xmath266 where @xmath267 is the result of replacing the brick @xmath329 in @xmath189 containing @xmath263 by two bricks , @xmath330 and @xmath331 , where @xmath330 contains the @xmath263-th column plus all the cells in @xmath329 to the left of @xmath263 and @xmath331 contains all the columns of @xmath329 to the right of @xmath263 , @xmath271 , and @xmath272 is the labeling that results from @xmath247 by changing the label of column @xmath263 from @xmath243 to @xmath244 . in case ( ii ) , @xmath273 where @xmath267 is the result of replacing the bricks @xmath319 and @xmath328 in @xmath189 by a single brick @xmath332 @xmath275 , and @xmath272 is the labeling that results from @xmath247 by changing the label of column @xmath263 from @xmath244 to @xmath243 . if neither case ( i ) or case ( ii ) applies , then we let @xmath276 .
for example , if @xmath249 is the element of @xmath333 pictured in figure [ figure:2fris1 ] , then @xmath261 is pictured in figure [ figure:2fris2 ] .
it is easy to see that the involution @xmath282 is weight - preserving and sign - reversing and hence shows that @xmath334 thus , we must examine the fixed points @xmath284 of @xmath282 .
first , by construction , @xmath7 must have @xmath68-matches starting at all odd positions .
hence @xmath335 . as before
, there can be no @xmath243 labels in @xmath247 so that @xmath285 .
moreover , if @xmath319 and @xmath328 are two consecutive bricks in @xmath189 and @xmath263 is the last column of @xmath319 , then it can not be the case that there is an @xmath68-match starting at position @xmath263 in @xmath7 since otherwise we could combine @xmath319 and @xmath328 .
it follows that @xmath336 .
vice versa , if @xmath335 , then we can create a fixed point , @xmath289 , by forming bricks in @xmath189 that end at even numbered columns that are not the start of an @xmath68-match in @xmath7 , and labeling each even numbered column that is the start of an @xmath68-match in @xmath7 with @xmath242 , and labeling the remaining even numbered columns with @xmath244 .
thus we have shown that @xmath337 as desired .
applying @xmath338 to the identity @xmath292 , we get @xmath339 which proves ( [ eq:2fris ] ) . as noted in the introduction , putting @xmath149 in ( [ eq:2fris ] ) gives us the generating function for the number of @xmath68-alternating elements of @xmath340 .
that is , if @xmath68 is a subset of @xmath89 , then @xmath341 next we give the proof of theorem [ thm : oax ] .
+ * proof of theorem [ thm : oax ] .
* + our goal is to prove that if @xmath68 is a subset of @xmath89 , then @xmath342 let @xmath299 $ ] be the ring homomorphism defined in theorem [ thm : eax ] .
define @xmath343 by setting @xmath344 if @xmath42 is odd and setting @xmath345 for @xmath93 .
we have defined @xmath197 so that for @xmath93 , @xmath346 then we claim that for @xmath178 , @xmath347 and @xmath348 note that @xmath349 now if @xmath42 is odd , then every @xmath306 must have an odd part and hence our definitions force @xmath307 .
thus @xmath350 for all @xmath178 . on the other hand ,
if @xmath42 is even , then we can restrict ourselves to those partitions which contain only even parts in ( [ 3key ] ) .
thus for @xmath93 we have @xmath351 as before , we must give a combinatorial interpretation to ( [ 3ris2 ] ) .
first we interpret + @xmath352 as the number of ways to pick a set partition of @xmath353 into sets @xmath234 where @xmath315 for @xmath354 and @xmath355 .
for each set @xmath237 with @xmath356 , we interpret @xmath316 as the number of ways to arrange the numbers in @xmath237 into a @xmath317 array such that there is an @xmath68-match starting at positions @xmath318 .
we interpret @xmath357 as arranging the numbers in @xmath358 in a @xmath359 array such that there is an @xmath68-match starting at positions @xmath360 . in this case
, we imagine that the array fills the first @xmath361 columns of the brick @xmath362 and the last column is left blank .
finally , we interpret @xmath241 as all ways of picking labels of all the even numbered columns of each brick , except the final even numbered column in the brick , with either an @xmath242 or a @xmath243 .
for completeness , we label the final even numbered column in each brick with @xmath244 .
we call all such objects created in this way _ filled labeled brick tabloids _ and let @xmath363 denote the set of all filled labeled brick tabloids that arise in this way .
thus , a @xmath364 consists of a brick tabloid , @xmath189 , a filling , @xmath365 , and a labeling , @xmath247 , of the even numbered columns of @xmath189 with elements from @xmath248 such that 1 .
all bricks have even size , 2 . for all but the final brick
, @xmath7 has an @xmath68-match starting at each column that is not the final column of its brick , 3 . for the final brick ,
the last column is empty and @xmath7 has an @xmath68-match starting at each column that is not one of the final two columns of that brick , 4 .
the final column of each brick is labeled with @xmath244 , and 5 .
each even numbered column that is not a final even numbered column of a brick is labeled with @xmath242 or @xmath243 .
we then define the weight of @xmath249 , @xmath250 , to be the product of all the @xmath242 labels in @xmath247 and the sign of @xmath249 , @xmath251 , to be the product of all the @xmath243 labels in @xmath247 .
for example , if @xmath252 , @xmath253 , @xmath254 , and @xmath323 , then figure [ figure:3fris1 ] pictures such a filled labeled brick tabloid @xmath366 with @xmath367 and @xmath258 .
thus @xmath368 at this point , we can follow the same set of steps as in theorem theorem [ thm : eax ] to prove @xmath369 that is , we can define the weight - preserving sign - reversing involution @xmath370 in exactly the same way as in theorem [ thm : eax ] and use it to prove ( [ 3fris8 ] ) because the fact that the last cell is empty does not effect any of the arguments that we used in theorem [ thm : eax ] .
applying @xmath338 to the identity ( [ pvu1 ] ) and using ( [ etimesnu ] ) and ( [ 3fris8 ] ) , we obtain @xmath371 dividing the above equation by @xmath263 will then yield @xmath372 which proves ( [ eq:3fris ] ) .
putting @xmath149 in ( [ eq:3fris ] ) gives us the generating function for the number of @xmath68-alternating elements of @xmath373 .
that is , if @xmath68 be a subset of @xmath89 , then @xmath374
in this section we provide arguments similar to those in ( * ? ? ? * sect .
4 ) to prove theorem [ thm : n ] .
that is , suppose that @xmath375 .
recall that @xmath376 and @xmath377 where @xmath378 is the number of @xmath5 with no @xmath68-matches .
our goal is to prove that @xmath379 let @xmath380 denote the set of all @xmath5 such that @xmath7 has exactly one @xmath68-match and it occurs at the end of @xmath7 , i.e. the unique @xmath68-match in @xmath7 starts at position @xmath381 .
we then let @xmath382 and @xmath383 [ lem01 ] suppose that @xmath375
. then @xmath384 .
suppose that @xmath5 .
for any set @xmath385 of size @xmath386 , let @xmath387 be the @xmath388 array so that the last column consists of the elements of @xmath389 in increasing order , reading from bottom to top , and the first @xmath42 columns is a @xmath1 array , @xmath64 , on the elements of @xmath390 such that @xmath391 .
for example , if @xmath392 and @xmath393 pictured on the left in figure [ figure : fs ] , then @xmath387 is pictured on the right .
it is easy to see that for any set @xmath385 of size @xmath386 , if @xmath7 has no @xmath68-matches then either @xmath387 has no @xmath68-matches or @xmath394 .
it follows that @xmath395 if we multiply both sides of ( [ a = a+b ] ) by @xmath396 and sum for @xmath178 , we get that @xmath397 or that @xmath398 suppose that @xmath5 and @xmath399
. one can read any such @xmath7 from left to right making a cut right after the first @xmath68-match .
say the first @xmath68-match ends in column @xmath400 . then starting at column @xmath401 one can continuing reading left to right and make another cut after the first @xmath68-match that one encounters . continuing on in this way , one obtains @xmath6 fillings @xmath402 that have exactly one @xmath68-match and that @xmath68-match occurs at the end of the filling and is possibly followed by a filling , @xmath403 , with no @xmath68-matches . in terms of generating functions
, this says that @xmath404 thus ( [ eq : n2 ] ) immediately follows by substituting the expression for @xmath405 given in lemma [ lem01 ] into ( [ eq : n3 ] ) .
in this section , we show how to compute @xmath406 for various @xmath112 .
in the case @xmath77 , we can compute @xmath406 for all @xmath68 that are subsets of the set of standard tableaux of shape @xmath78 .
we clearly have only 3 choices for @xmath68 in that case , namely , @xmath407 and @xmath408 , and @xmath409 , where @xmath79 and @xmath80 are pictured in figure [ figure:2standard ] . for @xmath410 , it is easy to see that @xmath411 if and only if @xmath412 and @xmath413 for all @xmath6 so that @xmath414 for all @xmath415
. it is also easy to see that @xmath416 if and only if @xmath7 is a standard tableau of shape @xmath417 , so by the hook formula for the number of standard tableaux of shape @xmath78 , @xmath418 where @xmath419 is @xmath42-th catalan number
. we shall give a bijective proof of this theorem by constructing a bijection from @xmath423 , which is the set of dyck paths of length @xmath424 , onto @xmath425 .
a dyck path of length @xmath426 is a path that starts at ( 0,0 ) and ends at ( 2n,0 ) and consists of either _ up - steps _ ( 1,1 ) or _ down - steps _ ( 1,-1 ) in such a way that the path never goes below the @xmath242-axis .
it is well known that the number of dyck paths of length @xmath426 is equal to @xmath42-th catalan number , @xmath125 .
now suppose that @xmath427 .
because there is a @xmath80-match starting in column @xmath6 for @xmath428 , it follows that @xmath429 for @xmath428 .
thus @xmath430 and @xmath431 .
now since @xmath432 and @xmath433 , it follows that we must have that @xmath434 , @xmath435 , @xmath436 , and @xmath437 .
let @xmath438 be the set of fillings , @xmath7 , of the @xmath439 rectangle with the integers @xmath440 such that if the elements of @xmath7 in the first row , reading from left to right , are @xmath441 , and the elements of @xmath7 in the second row , reading from left to right , are @xmath442 , then ( i ) @xmath443 , ( ii ) @xmath444 and @xmath445 , ( iii ) @xmath446 , and ( iv ) @xmath447 and @xmath448 .
let @xmath449 be the set of paths of length @xmath424 that consist of either _ up - steps _ ( 1,1 ) or _ down - steps _ ( 1,-1 ) such that the first step goes from @xmath450 to @xmath451 and the last step goes from @xmath452 to @xmath453 .
here we do not require that the paths in @xmath449 stay above the @xmath242-axis .
there is a simple bijection , @xmath454 , which maps @xmath449 onto @xmath438 . given a path @xmath455 , we label the segments @xmath456 with the numbers @xmath457 , as pictured in figure [ figure : p22bij ] .
we then create a filling , @xmath458 , by processing the elements @xmath459 in order by putting @xmath6 in the first available cell in the first row of @xmath7 if @xmath460 is an up - step and putting @xmath6 in the first available cell in the second row of @xmath7 if @xmath460 is an down - step .
see figure [ figure : p22bij ] for an example .
it is easy to see that @xmath461 will be an element of @xmath438 since there are always @xmath462 up steps and @xmath462 down steps in @xmath463 .
clearly , the process is reversible , so @xmath454 is a bijection .
thus all we have to prove is that @xmath464 if and only if @xmath465 .
now suppose that @xmath466 and @xmath467 .
we know that @xmath468 , @xmath469 and @xmath470 . thus there is a @xmath80-match that starts in the first column of @xmath7 .
now fix @xmath471 .
we shall show that there is a @xmath80-match starting in column @xmath11 of @xmath472 .
suppose that @xmath473 is the @xmath11-th up - step of @xmath3 , reading from left to right .
thus , @xmath474 and the right - hand endpoint of @xmath473 must have height at least @xmath244 , so there can be at most @xmath475 down - steps among @xmath476 .
thus , no more than @xmath475 cells of the second row of @xmath7 are filled with numbers less than @xmath6 , while the first @xmath477 cells of the first row of @xmath7 are filled with numbers that are less than or equal to @xmath6 .
this means that @xmath478 which implies that @xmath479 .
hence there is a @xmath80-match starting in column @xmath11 of @xmath7 .
thus @xmath480 .
next suppose that @xmath481 .
then @xmath3 does not stay above the @xmath242-axis , so there is a smallest @xmath6 such that the right - hand endpoint of @xmath473 is at level @xmath243 .
it follows that for some @xmath482 , there must be @xmath386 down - steps and @xmath483 up - steps in @xmath476 , so @xmath484 .
this means @xmath485 and the elements @xmath486 occupy the first @xmath386 columns of @xmath487 . but
this means that @xmath488 , so there is not a @xmath80-match starting in column @xmath386 of @xmath7 .
our next goal is to classify those standard tableaux , @xmath153 , such that @xmath490 . to this end
, we say that a standard tableau , @xmath153 , is _ degenerate _ if at least one of @xmath491 or @xmath492 holds for each @xmath493 .
then we have the following theorem .
fix @xmath153 . to help us visualize the order relationships within @xmath3 , we form a directed graph @xmath494 on the cells of the @xmath76 rectangle by drawing a directed edge from the position of the number @xmath6 to the position of the number @xmath8 in @xmath3 for @xmath495 . for example , in figure [ figure : unique ] , the @xmath496 pictured on the left , results in the directed graph @xmath494 pictured immediately to its right .
then @xmath494 determines the order relationships between all the cells in @xmath3 since @xmath497 if there is a directed path from cell @xmath498 to cell @xmath499 in @xmath494 .
now suppose that @xmath500 where @xmath501 . because there is a @xmath3-match starting in column @xmath6 , we can superimpose @xmath494 on the cells in columns @xmath6 and @xmath8 to determine the order relations between the elements in the those two columns .
if we do this for every pair of columns , @xmath6 and @xmath8 for @xmath502 , we end up with a directed graph on the cells of the @xmath1 rectangle which we will call @xmath503 .
for example , figure [ figure : unique2 ] , @xmath504 is pictured just to the right of @xmath494 .
it is then easy to see that if @xmath500 and there is a directed path from cell @xmath498 to cell @xmath499 in @xmath503 , then it must be the case that @xmath505 .
note that @xmath503 will always be a directed acyclic graph with no multiple edges .
in general , the problem of computing @xmath155 for a column strict tableau @xmath3 of shape @xmath4 can be reduced to finding the number of linear extensions of a certain poset .
that is , we claim the graph @xmath503 induces a poset @xmath506 on the cells of the @xmath1 rectangle by defining @xmath507 if and only if there is path from @xmath10 to @xmath508 in @xmath503 .
it is easy to see that @xmath509 is transitive .
since @xmath503 is acyclic , we can not have @xmath510 .
thus @xmath511 is poset .
it follows that there is a 1:1 correspondence between the elements of @xmath512 and the linear extensions of @xmath513 .
that is , if @xmath500 , then it is easy to see that @xmath514 where @xmath515 is a linear extension of @xmath513 .
vice versa , if @xmath514 is a linear extension of @xmath513 , then we can define @xmath7 so that @xmath516 and it will automatically be the case that @xmath500 .
since every poset has at least one linear extension , it follows that @xmath517 for all @xmath415 .
thus @xmath518 for all @xmath93 .
our next goal is to apply a simple lemma on directed acyclic geraphs with no multiple edges to replace @xmath503 by a simpler acyclic directed graph which contains the same information about the relative order of the elements in @xmath7 . given a directed acyclic graph @xmath519 with no multiple edges , let @xmath520 equal the set of all pairs @xmath521 such that there is a directed path in @xmath64 from vertex @xmath6 to vertex @xmath11
. then we have the following .
[ lem : remove ] let @xmath519 be a directed acyclic graph with no multiple edges .
let @xmath522 be the subgraph of @xmath64 that results by removing all edges @xmath523 such that there is a directed path from @xmath6 to @xmath11 in @xmath64 that does not involve @xmath524 .
. the proof will be by induction on the number , @xmath386 , of edges @xmath523 such that there is a directed path from @xmath6 to @xmath11 in @xmath64 that does not involve @xmath524 .
if @xmath526 , there is nothing to prove . if @xmath29 , let @xmath523 be an edge such that there is a directed path , @xmath527 from @xmath6 to @xmath11 , in @xmath64
, that does not involve @xmath524 .
let @xmath528 be the directed graph that results from @xmath64 by removing edge @xmath524 .
now suppose that @xmath529 . if there is a directed path from @xmath242 to @xmath530 in @xmath64 that does not involve @xmath524 , then clearly @xmath531
otherwise , suppose that @xmath532 is a directed path from @xmath242 to @xmath530 in @xmath64 , for some @xmath533 , that uses the edge @xmath524 .
thus for some @xmath534 , we have @xmath535 and @xmath536 .
but then @xmath537 is a path in @xmath528 which connects @xmath242 and @xmath530 in @xmath528 .
thus it follows that @xmath528 is a connected acyclic graph with no multiple edges such that @xmath525 .
however , @xmath528 has @xmath483 edges , @xmath538 , such that there is a directed path from @xmath539 to @xmath42 in @xmath528 that does not involve @xmath540 .
thus by induction , we can remove all such edges from @xmath528 to obtain an @xmath522 with @xmath541 .
now suppose that we apply lemma [ lem : remove ] to the graph @xmath504 in figure [ figure : unique ] to produce @xmath542 . we know that for any @xmath543
, it must be the case that @xmath505 for all @xmath544 .
thus since @xmath545 , it must be that case that whenever @xmath546 , @xmath505 for all @xmath544 .
now the directed edge @xmath547 is not in @xmath542 since @xmath504 has the directed path @xmath548 .
also , directed edge @xmath549 is not in @xmath542 because of the directed path @xmath550 .
similarly , the corresponding edges in the third column of @xmath504 will be removed when creating @xmath542 , so @xmath542 is the directed graph on the cells of the @xmath551 rectangle pictured on the far right of figure [ figure : unique ] .
observe that in this case , @xmath542 has the property that outdegree of each vertex is 1 which means that @xmath542 determines a total order on the entries of any @xmath544 .
thus , there is exactly one @xmath544 , which means that @xmath552 in this case .
we should note , however , that it is not always be the case that applying lemma [ lem : remove ] to remove edges from @xmath503 will produce a graph , @xmath553 , where each vertex has outdegree 1 .
for example , consider the standard tableau @xmath554 pictured on the left in figure [ figure : unique2 ] .
the graph @xmath555 is pictured immediately to the right @xmath554 and the graph @xmath556 is pictured immediately to the right of @xmath555 .
the directed edge @xmath547 is not in @xmath557 due to the directed path of length 2 , @xmath558 , in @xmath556 . also , the directed edge @xmath559 is not in @xmath557 since @xmath556 has the directed path @xmath560 .
similarly , the corresponding edges in the third column of @xmath556 are eliminated when creating @xmath557 from @xmath556 .
it is then easy to check that @xmath557 is the directed graph on the cells of the @xmath551 rectangle pictured on the far right of figure [ figure : unique2 ] .
note that in this case , cells @xmath561 and @xmath562 have outdegree 2 .
in fact , it is not difficult to see that there are exactly four @xmath563 so that @xmath564 in this case .
let @xmath83 be a standard tableau .
for all @xmath415 , let @xmath553 be the acyclic directed graph that arises from @xmath503 by removing all edges @xmath565 in @xmath503 such that there is a directed path from @xmath498 to @xmath499 in @xmath503 that does not involve @xmath524 .
by lemma [ lem : remove ] , we know that @xmath553 imposes the same order relations on any @xmath566 as were imposed by @xmath503 .
suppose that @xmath3 is a degenerate standard tableau in @xmath117 .
we claim the outdegree of any vertex in @xmath553 is 1 , which means that @xmath553 imposes a total order on the cells in the @xmath1 rectangle .
thus , there is only one @xmath566 and , hence , @xmath154 for all @xmath93 .
let @xmath567 be the set of directed edges in the graph @xmath494 and @xmath568 be the set of directed edges in @xmath503 .
first observe the out degree of any cell in column 1 or column @xmath42 of @xmath503 is 1 since @xmath494 is superimposed only once in column 1 and column @xmath42 .
however , for any column @xmath569 with @xmath570 , @xmath494 is superimposed on columns @xmath571 and @xmath569 and on columns @xmath569 and @xmath572 so the outdegree of each cell in column @xmath569 could be 2 .
next , observe that since @xmath3 is a standard tableau , we always have that @xmath573 and the elements in the columns of @xmath3 are strictly increasing from bottom to top .
this means that there are three possibilities for an edge in @xmath574 coming out of cell @xmath575 for @xmath576 , namely , it can go to cell @xmath577 , to cell @xmath578 , or to some cell @xmath579 , where @xmath580 .
similarly there are only two possibilities for an edge in @xmath574 coming out of cell @xmath578 for @xmath576 , namely , it can go to cell @xmath581 or it can go to cell @xmath582 , where @xmath583 . since @xmath3 is degenerate , we must have at least one of the edges @xmath584 and @xmath585 in @xmath494 , so in fact , we only have 4 possible cases for pairs of edges that leave cells @xmath575 and @xmath578 , as pictured in figure [ figure : possible ] .
+ * case h. * edges @xmath584 and @xmath585 are in @xmath494 .
+ in this case , the outdegree of cell @xmath586 in column @xmath587 of @xmath553 is 1 because @xmath588 is the only edge out of each @xmath586 in @xmath503 .
+ * case u. * edges @xmath584 and @xmath589 are in @xmath494 for some @xmath590 .
+ in this case , since @xmath591 and @xmath592 , @xmath494 contains the path , @xmath593 , for some @xmath594 , which passes through all the cells in the first column from @xmath582 to @xmath595 .
now consider @xmath503 , as shown in figure [ figure : caseb ] .
for each interior column , @xmath596 , @xmath503 contains the paths @xmath588 and @xmath597 .
therefore , @xmath553 omits the edge @xmath588 in all but its first column , leaving the single edge , @xmath598 , exiting each cell in row @xmath6 .
+ * case x. * edges @xmath599 and @xmath585 are in @xmath494 .
+ this case never occurs .
since @xmath3 increases along rows and columns , this case would require @xmath600 while maintaining @xmath601 , which does not have an integer solution unless @xmath602 , which is impossible in @xmath3 .
+ * case d. * edges @xmath603 and @xmath585 are in @xmath494 for some @xmath580 .
+ in this case , since @xmath604 and @xmath605 , @xmath494 contains the path , @xmath606 , for some @xmath607 , which passes through all the cells in the second column from @xmath579 to @xmath608 .
now consider @xmath503 , as shown in figure [ figure : cased ] .
for each interior column , @xmath596 , @xmath503 contains the paths @xmath588 and @xmath609 .
therefore , @xmath553 omits the edge @xmath588 in all but its last column , leaving a single edge , @xmath610 , exiting each cell in row @xmath6 .
now suppose that @xmath3 is a standard tableau of shape @xmath4 that is not degenerate .
we will show that @xmath611 .
let @xmath612 be the smallest @xmath6 such that neither @xmath613 nor @xmath614 .
thus , @xmath494 has the edges @xmath603 and @xmath589 , where @xmath615 .
fix @xmath501 .
we know @xmath517 so that we fix some @xmath500 . since @xmath7 is column strict , all the elements in the second column below cell @xmath578 are less than @xmath616 and all the elements in the second column above cell @xmath581 are greater than @xmath617 .
that is , we have the situation pictured in figure [ figure : caseii ] .
let @xmath389 be the elements in the first column of @xmath7 which are in the interval @xmath618 and let @xmath189 be the elements of @xmath7 in columns @xmath619 which are in the interval @xmath618 . since @xmath7 has a @xmath3-match in the first two columns , we know that @xmath620 must be an immediate successor of @xmath621 relative to the elements of the first two columns of @xmath7 so that @xmath622 .
similarly , since @xmath7 has a @xmath3-match in columns 2 and 3 , we know that @xmath623 must be an immediate successor of @xmath621 relative to the elements of in columns 2 and 3 of @xmath7 so that @xmath624 .
thus we know that @xmath389 and @xmath189 are non - empty .
let @xmath625 and @xmath626 .
it follows that the interval @xmath618 consists of @xmath627 elements in the first column and @xmath263 elements in columns @xmath628 of @xmath7 .
next suppose that we pick any set of @xmath629 of @xmath627 elements of from @xmath630 .
we can then we modify the positions of the elements in the interval @xmath618 in @xmath7 by placing the elements of @xmath629 in column 1 in the positions that were occupied by @xmath389 in column 1 of @xmath7 in the same relative order as the elements of @xmath389 in column 1 of @xmath7 and placing the elements of @xmath631 in the positions that were occupied by @xmath189 in columns @xmath619 of @xmath7 in the same relative order as the elements of @xmath189 in columns @xmath619 of @xmath7 . call this modified filling @xmath632
because we only changed the positions of elements in the interval @xmath618 in @xmath7 , it is easy to see that @xmath632 will be an element of @xmath512
. thus , @xmath512 contains at least @xmath633 elements .
since @xmath634 , @xmath635 and , hence , @xmath611 for all @xmath501 . now that we have classified the standard tableaux , @xmath3 , of shape @xmath4 such that @xmath154 for all @xmath93 , we will determine the number of such column strict tableaux .
let @xmath636 be the set degenerate column strict tableaux of shape @xmath4 .
a motzkin path of length @xmath42 is a lattice path on the integer lattice in the plane that runs from @xmath450 to @xmath637 consisting of up - steps @xmath638 , down - steps @xmath639 , and horizontal - steps @xmath640 so that the path never passes below the @xmath242-axis , we let @xmath641 denote the set of motzkin paths of length @xmath42 .
if we are allowed to color the horizontal - steps of the paths @xmath642 with one of @xmath386 colors , then we obtain the set of @xmath386-colored motzkin paths of length @xmath42 .
we let @xmath643 denote the set of @xmath386-colored motzkin paths . for @xmath644
, we shall denote the 2-colored horizontal - steps as @xmath522 and @xmath645 .
it was proved in @xcite , @xcite that @xmath646 where @xmath419 is the @xmath42-th catalan number . by the hook formula , we know that @xmath647 .
thus there exists a bijection from the set of standard tableaux of shape @xmath648 onto the set of @xmath649-colored motzkin paths of length @xmath650 .
we claim that we can define such a bijection @xmath651 for @xmath415 as follows . given a column strict tableau @xmath652 ,
define @xmath653 to be the path @xmath654 such that @xmath655 equals @xmath656 note that we have not used the edge , @xmath524 , out of cell @xmath657 in this definition .
we must first check that @xmath658 is a 2-colored motzkin path .
this is equivalent to showing that ( i ) there are an equal number of @xmath659 s and @xmath660 s , i.e. , the path ends at height @xmath661 , and ( ii ) we do nt encounter more @xmath660 s than @xmath659 s in @xmath662 for any @xmath6 . in order to verify condition ( i ) , we notice that if @xmath663 , both edges out of the cells in row @xmath6 are directed toward the first column in @xmath494 , whereas if @xmath664 , both edges out of the cells in row @xmath6 are directed toward the second column in @xmath494 .
if @xmath655 equals @xmath522 or @xmath645 , then there is an edge directed toward each column out of the cells of row @xmath6 in @xmath494 . thus if there are @xmath50 @xmath659 s , @xmath268 @xmath660 s , and @xmath569 @xmath522 s and @xmath665 @xmath645 s in @xmath658 , then these account for @xmath666 edges directed toward the first column in @xmath494 and @xmath667 edges directed toward the second column in @xmath494 . the total in - degree of the first column is @xmath650 since every entry except @xmath668 has an edge of @xmath494 coming into it .
similarly , the total in - degree of the second column is @xmath42 .
since @xmath658 accounts for every edge in @xmath669 except @xmath524 , which is directed towards the second column , it must be the case that @xmath670 .
hence @xmath671 and there are an equal number of @xmath659 s and @xmath660 s in @xmath658
. thus @xmath658 starts at ( 0,0 ) and ends at @xmath637 .
if condition ( ii ) fails , then let @xmath6 be the position of the first step to end beneath the @xmath242-axis , i.e. @xmath664 and there are an equal number of @xmath659 s and @xmath660 s among @xmath672 . by our previous argument , there must be @xmath673 edges directed into second column from the cells in rows @xmath674 . since edges that go from the first column to the second column
go weakly downward and edges from the second column to the second column can go up by at most one , it follows that the edges that hit the second column from cells in rows @xmath675 must all end somewhere in the first @xmath6 rows .
now suppose that @xmath579 , with @xmath676 , is a cell that is not the endpoint of a directed edge from cells in rows @xmath674 .
it can not be that @xmath677 since this would force @xmath494 to have edges @xmath599 and @xmath585 .
but this means that @xmath678 and @xmath679 which is impossible since we must have have @xmath680 .
thus we must assume that @xmath580 .
since @xmath655 is a down - edge and the edge that goes into cell @xmath578 can only have come from cell @xmath681 , it must be the case that @xmath494 has edge @xmath603 . in this case
, we must have the situation pictured in figure [ figure : deg1 ] .
that is , because @xmath3 is a column strict tableau , arrows that go from column 1 to column 2 can not cross each other .
hence all the cells of the form @xmath608 where @xmath682 must have been hit by vertical arrows . but
this is impossible because then @xmath683 would be a set of consecutive elements so it would be impossible to have @xmath680 , as is required for @xmath3 to be a column strict tableau . since we know
that @xmath684 , to prove that @xmath291 is bijection , we need only show that @xmath291 is 1:1 .
let @xmath685 and @xmath686 be two distinct column strict tableaux of shape @xmath648 .
let @xmath627 be the largest element such that @xmath687 are in the same cells in both @xmath685 and @xmath686 and suppose that @xmath627 is in cell @xmath688 .
then without loss of generality , we can assume that @xmath689 is in column 1 in @xmath685 and in column 2 in @xmath686 .
but this means that the arrow out of @xmath688 goes to column 1 in @xmath690 and to column 2 in @xmath691 , which implies that the @xmath6-th step in @xmath692 is not equal to the @xmath6-th step in @xmath693 , so @xmath694 .
thus we have proved the following theorem .
we say that two column strict tableaux @xmath699 are _ consecutive - wilf equivalent _
( @xmath569-wilf equivalent ) if @xmath700 and we say that @xmath3 and @xmath554 are _ strongly @xmath569-wilf equivalent _ if @xmath701 .
clearly , it follows from theorem [ thm : d ] that if @xmath3 and @xmath554 are degenerate column strict tableaux of shape @xmath4 , then @xmath3 and @xmath554 are strongly @xmath569-wilf equivalent .
we end this section with a few remarks about how we have derived formulas for @xmath155 for certain column strict tableaux @xmath3 .
first there are some simple equivalences which help us reduce the number of @xmath156 that we have to consider . given any @xmath5
, we define the generalized complement of @xmath7 , @xmath702 as a composition of three operations on @xmath7 , @xmath569 for complement , @xmath703 for flipping the diagram about the @xmath530-axis , and @xmath704 for flipping the diagram about the @xmath242-axis . here
the operation of complement on @xmath7 is the result of replacing each element @xmath6 in @xmath7 by @xmath705 .
the operation @xmath703 is the operation of taking a filling of the @xmath1 rectangle with the numbers @xmath2 , and flipping the diagram about the @xmath530-axis and the operation @xmath704 is the operation of taking a filling of the @xmath1 rectangle with the numbers @xmath2 , and flipping the diagram about the @xmath242-axis .
we then define @xmath706 .
this process is pictured in figure [ figure : flip ] .
it is easy to see that @xmath5 if and only if @xmath707 and @xmath156 if and only if @xmath708 . moreover for any column strict tableau @xmath3 of shape @xmath4 , @xmath5 has a @xmath3-match in columns @xmath6 and @xmath8 if and only if @xmath702 has a @xmath709-match in columns @xmath710 and @xmath711 .
it follows that @xmath500 if and only if @xmath712 so that @xmath713 .
thus @xmath3 and @xmath709 are c - wilf equivalent . now in @xmath14 , the only non - degenerate column strict tableaux that are not complements of each other are the tableaux @xmath3 and @xmath554 pictured in figure [ figure : pq ] .
recall that in the proof of theorem [ degenerate ] , we observed that for any standard tableaux @xmath156 , the graph @xmath503 induces a poset @xmath506 on the cells of the @xmath1 rectangle by defining @xmath507 if and only if there is path from @xmath10 to @xmath508 in @xmath503 and that there is a one - to - one correspondence between the linear extensions of @xmath513 and the elements of @xmath512 .
now in the case of the standard tableaux @xmath3 and @xmath554 in figure [ figure : pq ] , we have pictured @xmath714 just to the right of @xmath3 and @xmath715 just to the right of @xmath554 . in the case of @xmath3 , if we eliminate the vertices corresponding to ( 1,1 ) and ( 1,2 ) , which must always be the first two elements in any linear extension of @xmath716 , and we eliminate the vertices in cells ( 3,4 ) and ( 3,5 ) , which must always be the last two elements in any linear extension of @xmath717 , we are left with the hasse diagram pictured to its right . in the case of @xmath554 , if we eliminate the vertices corresponding to cells ( 1,1 ) , ( 2,1 ) , ( 1,2 ) and ( 2,2 ) , which must always be the first four elements in any linear extension of @xmath718 , and we eliminate the vertices in cells ( 3,4 ) and ( 3,5 ) , which must always be the last two elements in any linear extension of @xmath718 , we are left with the hasse diagram pictured to its right . we have developed general recursions of the number of linear extension of posets whose hasse diagrams are similar to the one pictured for @xmath554 in figure [ figure : pq ] which have allowed us to show that @xmath719 .
this work will appear in subsequent paper .
however , we have not been able to find a formula for @xmath155 .
langley , alternative transition matrices for brenti s q - symmetric functions and a class of @xmath722-symmetric functions on the hyperoctahedral group , _ proceedings of the 2002 conference on formal power series and algebraic combinatorics _ , melbourne australia .
a. mendes , j.b .
remmel , a. riehl , permutations with @xmath386-regular descent patterns , _ permutation patterns _ , ( s. linton , n. ruskuc , v. vatter , eds . ) , london .
lecture notes 376 , ( 2010 ) , 259 - 286 . | in this paper , we study pattern matching in the set @xmath0 of fillings of the @xmath1 rectangle with the integers @xmath2 such that the elements in any column increase from bottom to top .
let @xmath3 be a column strict tableau of shape @xmath4 .
we say that a filling @xmath5 has @xmath3-match starting at @xmath6 if the elements of @xmath7 in columns @xmath6 and @xmath8 have the same relative order as the elements of @xmath3 .
we compute the generating functions for the distribution of @xmath3-matches and nonoverlapping @xmath3-matches for various classes of standard tableaux of shape @xmath4 .
we say that a filling @xmath5 is _ @xmath3-alternating _ if there are @xmath3-matches of @xmath7 starting at all odd positions but there are no @xmath3-matches of @xmath7 starting at even positions .
we also compute the generating functions for @xmath3-alternating elements of @xmath0 for various classes of standard tableaux of shape @xmath4 . |
the search for physics beyond the standard model ( sm ) is a topic of considerable interest in high - energy particle physics .
concurrently , efforts are also underway to uncover signatures of new physics at low- and medium - energies using atomic and nuclear processes . in this respect , there exist tantalizing hints of new physics in results of neutron and superallowed nuclear @xmath9-decays , which imply a value for @xmath10 differing from the sm unitarity requirement by two or more @xmath11 @xcite .
similarly , the weak charge of the cesium atom , @xmath12 , measured in atomic parity - violation ( apv ) by the boulder group , has been found to differ from the sm prediction by @xmath13 @xcite ( see , however , ref .
if conventional many - body atomic and nuclear effects can be ruled out as the source of these deviations , the @xmath9-decay and apv results imply the existence of new physics at the 1 - 10 tev scale@xcite .
this possibility has motivated a variety of additional atomic and nuclear new physics searches , including new measurements of the neutron @xmath9-decay parameters , apv observables along a chain of isotopes , and parity - violating ( pv ) electron - electron and electron - proton scattering .
one possible manifestation of new physics not probed by the aforementioned experiments would be the existence of new low - energy interactions involving a single generation of fermions which violate time - reversal invariance ( t ) but conserve parity invariance ( p ) .
such interactions are allowed in the sm when quarks of different generations participate .
recently , the first non - zero result for a @xmath14 t - violating , p - conserving ( tvpc ) observable has been reported by the cplear collaboration , which measured the @xmath15 decay asymmetry @xcite .
the results are consistent with the value expected from the measured cp - violating parameter @xmath16 and the cpt theorem .
no new physics is required to explain this result . in the @xmath17 sector , a variety of direct searches for tvpc effects have been carried out .
these efforts include studies of detailed balance in nuclear reactions@xcite , @xmath18-ray correlations in nuclear @xmath18-decay@xcite , five - fold correlations ( fc ) in the scattering of epithermal neutrons from aligned nuclear targets@xcite , charge symmetry breaking ( csb ) in @xmath19 scattering @xcite , and the @xmath20 in neutron @xmath9-decay@xcite .
thus far , all studies have yielded null results .
the limits from the purely hadronic reactions imply @xmath21 , where @xmath22 gives the ratio of tvpc nuclear matrix elements to those of the residual strong interaction .
limits on @xmath23 tvpc interactions involving light quarks may also be derived indirectly from results for atomic , neutron , and electron electric dipole moments ( edm s ) . as observed in ref .
@xcite , the presence of both a new tvpc interaction and a conventional pv interaction ( _ e.g. _ , in the standard model ) could conspire to generate a non - zero edm , whose interaction with an external field violates both p and t. to the extent that pv radiative corrections to possible new tvpc interactions can be calculated , one can derive limits on new tvpc interactions from edm results .
attempts to do so were first reported in ref .
the calculation involved two external elementary fermions ( _ e.g. _ , two valence quarks in the neutron ) and a one - loop @xmath24-boson radiative correction to dimension seven , four - fermion , tvpc operators .
two - loop effects , involving a single external fermion , for the electron edm , @xmath25 , and neutron edm , @xmath26 , were later studied in ref .
navely , one might expect the most stringent bounds on new tvpc interactions to be derived at one - loop order from the experimental limit on the atomic electric dipole moment of mercury , @xmath27 , since the latter is nearly two orders of magnitude more severe than the bound on @xmath26 .
it was argued in ref .
@xcite , however , that the two - loop effects in @xmath25 and @xmath26 generate considerably more stringent bounds than do the results of @xmath28 .
subsequently , the authors of ref .
@xcite recast the analysis of refs .
@xcite into the framework of low - energy effective field theory ( eft ) .
it was argued in ref .
@xcite that the results of ref .
@xcite imply bounds on new tvpc interactions in excess of those presently achievable with direct tvpc searches by several orders of magnitude .
these conclusions have had a discouraging effect on further direct tvpc searches .
recently , it was argued that the conclusions of ref .
@xcite are inconsistent with the separation of scales underlying eft @xcite . in brief
, the argument is as follows@xcite .
let @xmath29 denote the mass scale below which use of an eft involving nonrenormalizable tvpc operators makes sense .
one may expand the edm of an elementary particle , neutron , or atom as @xmath30 where the @xmath31 denote the set of _ a priori _ uknown coefficients of dimension @xmath32 nonrenormalizable operators in the effective lagrangian , the @xmath33 are calculable quantities arising from loops or many - body matrix elements , and @xmath34 is a mass scale associated with the appropriate dynamical degree of freedom in the eft .
the @xmath31 parameterize one s ignorance about the short - distance ( @xmath35 ) dynamics of the new time - reversal violating physics .
the first contributions from new tvpc interactions appear in the @xmath36 .
one may now consider eq .
( [ eq : edmexpand ] ) under two scenarios : * scenario ( a ) * parity symmetry is restored at some scale @xmath37 . in this case , all of the coefficients @xmath38 and @xmath39 must vanish at tree - level in the eft since parity invariance holds at short distances . consequently , the first contributions to the edm arise from loops involving the tvpc @xmath36 operators .
since @xmath40 , these contributions presumably dominate the remaining terms in the series .
hence , one may use experimental edm limits to constrain @xmath41 . as shown below
, the limits obtained from edm s under this scenario vastly exceed those obtainable from direct searches .
* scenario ( b ) * parity symmetry is restored at @xmath42 . in this case , the @xmath38 and @xmath39 do not , in general , vanish at tree - level in the eft since both pv and tv interactions take place at short distance .
there exists no reason to assume they fail to conspire in generating the lowest dimension tvpv effective interactions .
consequently , the tvpc interactions do not generate the leading contribution to the edm as in scenario ( a ) . without independent information on the @xmath43 one can not use the edm as a direct handle on the tvpc @xmath36 terms .
the latter may be more or less suppressed relative to the lower dimension contributions depending on the size of @xmath44 . since one
has no _ a priori _ information on @xmath44 , one can say very little about the importance of tvpc contributions . for the sake of argument
, one might assume @xmath45 so that the first term in eq .
( [ eq : edmexpand ] ) dominates . in this case
, the low - energy effects of tvpc interactions would be negligible . in the more general situation ,
however , one would have to use direct tvpc searches to constrain the new tvpc interactions under this scenario .
the analysis of refs .
@xcite implicitly assumes scenario ( a ) .
the edm calculations performed by these authors , however , do not display the proper @xmath46 scaling behavior which follows from eft .
the reasons for this failure are discussed in ref .
@xcite and summarized below . it was also shown in ref .
@xcite that there exist additional tvpc operators , not considered in refs .
@xcite , which contribute to the elementary fermion edm at one - loop order . under scenario ( a ) , these one loop effects yield the most stringent constraints on the size of tvpc effects . in
what follows , we extend the analysis of ref .
@xcite to include many - quark tvpc contributions to the neutron edm first studied in ref .
@xcite and to atomic edm s .
we concentrate on scenario ( a ) , since under scenario ( b ) one can not use edm s to derive unambiguous information about tvpc new physics . in the case of the neutron edm , we complete the one - loop analysis of ref .
@xcite , including additional diagrams required by electromagnetic gauge invariance .
we show that the impact of these new diagrams is as large as the one - loop effects considered previously in ref .
we also compute tree - level contributions arising from dimension seven tvpc operators not considered in ref .
@xcite . in the case of atomic edm s
, we consider the situation in which they arise from purely hadronic tvpc interactions in the nucleus .
traditionally , the effects of non - leptonic t - violation in nuclear and atomic processes have been analyzed using collective degrees of freedom ( mesons and baryons ) , rather than fundamental quark - quark or quark - gluon interactions .
the t - violating effects are characterized by hadronic coupling constants which may be related to the underlying quark and gluon t - violating interactions using standard hadron structure techniques . in this context ,
two hadronic effects are of interest : ( a ) the presence of a purely tvpc meson - nucleon interaction and ( b ) the presence of a tvpv meson - nucleon interaction .
the leading long - range " tvpc effect arises from @xmath47-meson exchange , where the tvpc @xmath48 vertex is characterized by a coupling strength @xmath49 and an interaction @xcite @xmath50 where @xmath51 and @xmath52 .
a time - reversal violating parity violating ( tvpv ) nuclear effect arises when the second vertex in the exchange is parity - violating .
alternately , a tvpv atomic moment can be generated by a tvpc nuclear @xmath47-exchange and the pv exchange of a @xmath24-boson between the nucleus and atomic electrons ( see fig . [
fig : atomic ] ) .
similarly , the longest - range tvpv effects generally arise from @xmath53-exchange . in this case , the relevant tvpv @xmath54 couplings are @xmath55 , where the superscript denotes the isospin channel and correspond to the interactions @xmath56 n\end{aligned}\ ] ] non - zero values of @xmath55 may arise from either new tvpc interactions plus weak radiative corrections , or from more conventional tvpv interactions , such as the @xmath57-term in the qcd lagrangian .
the latter have been considered extensively elsewhere @xcite , and we concentrate on the former .
upper bounds on @xmath58 have been derived from a variety of t - violating , p - conserving experiments , including the studies of detailed balance in nuclear reactions@xcite , neutron transmission through an aligned @xmath59ho target@xcite , and csb terms in the @xmath19 scattering cross section@xcite .
in addition , measurements of @xmath26 and @xmath28 yield bounds on @xmath49 , when pv and tvpc interactions conspire to generate an edm . under scenario ( a ) ,
the edm limits on @xmath49 may be stronger than those obtained from tvpc nuclear processes@xcite . regarding the @xmath55 , experimental edm limits yield the most stringent bounds .
as discussed in detail in ref .
@xcite , @xmath26 and @xmath27 are sensitive to different linear combinations of the @xmath55 . roughly speaking ,
the edm upper bounds on the @xmath60 are of the order of a few @xmath61 . in this paper , we relate @xmath49 and the @xmath55 to the underlying tvpc quark and gluon operators , and use limits on the hadronic couplings to infer limits on these underlying interactions .
we then compare these limits as well as those obtained from the one - loop many - quark contributions to @xmath26 with those obtained from the edm s of elementary fermions as in ref .
we find : 1 . under scenario ( a ) , the one - loop elementary fermion edm studied in ref . @xcite produces the most stringent limits on new tvpc interactions when applied to the electron and neutron .
many - quark effects @xmath62 or @xmath63 are suppressed by at least five orders of magnitude .
2 . by making certain naturalness assumptions
, one may use experimental edm limits to derive lower bounds on @xmath29 . under scenario ( a ) ,
one infers from @xmath25 ( single - quark @xmath26 ) limits that @xmath64 tev ( @xmath65 tev ) if the new tvpc physics is strong .
the corresponding bounds derived from many - quark effects in @xmath26 and @xmath66 are at roughly 1000 times weaker . in order for @xmath67 to provide competitive limits
, the precision of the atomic edm experiments would need to improve by roughly nine orders of magnitude would improve tighten the bounds on @xmath26 . ] .
3 . under scenario ( a )
, one expects @xmath68 .
4 . for scenario ( b ) , experimental limits on @xmath49 derived from fc in neutron transmission and csb in @xmath19 scattering produce the strongest limits on tvpc new physics . in terms of mass scales ,
these limits give @xmath69 gev for new strong tvpc physics .
an improvement of six orders of magnitude in experimental precision would bring this lower bound up to the weak scale .
the analysis leading to these conclusions is presented in the remainder of the paper as follows . in section 2
, we review the framework of eft for new low - energy tvpc and tvpv interactions . in section 3 we illustrate the application of this framework by considering the edm of an elementary fermion , as discussed in ref . @xcite . in section 4 ,
we discuss the renormalization of effective tvpv four - fermion operators arising from pv radiative corrections to effective tvpc interactions . here , we take particular care to implement electromagnetic gauge invariance . the latter implies the existence of additional contributions to the edm s of composite systems not considered in ref . @xcite . in section 5 , we relate the effective tvpc and tvpv operators to @xmath49 and the @xmath55 , respectively , using the quark model , factorization , and current algebra .
we also compute new many - quark contributions to @xmath26 generated by new operators including those required by gauge invariance not considered in ref . @xcite . in section 6 , we compare the implications of the many - quark contributions for the scale of new tvpc interactions with those obtained from the study of ref .
we also consider the limits on @xmath49 and the @xmath55 obtained from atomic edm limits and direct searches and the corresponding implications for new tvpc interactions under the two scenarios outlined above .
section 6 summarizes our conclusions .
herczeg _ et al _ have shown that tvpc interactions between quarks cannnot arise from tree - level boson exchange in renormalizable gauge theories @xcite .
such interactions could , however , be generated by higher - order or non - perturbative effects .
whatever new physics produces p - conserving t - violation among light quarks and gluons must be characterized by some heavy mass scale , @xmath29 . given that the underlying renormalizable gauge theory for p - conserving t - violation is not known , it is natural to consider the low - energy consequences of such a theory in the context of an eft valid below the scale @xmath29 . letting @xmath70 denote the effective , low - energy lagrangian for new physics , we follow ref . @xcite and expand in inverse powers of @xmath29 : @xmath71 where the subscripts denote the dimension of the operators appearing in each term and where @xmath72 with the sum running over @xmath73 complete set of dimension @xmath32 operators @xmath74 . the lowest dimension t - violating operators appear in @xmath75 .
these operators are tvpv only . of particular interest here
are the electric dipole fermion - gauge boson interactions : @xmath76 where @xmath77 , @xmath78 , and @xmath79 denote the photon , @xmath24-boson , and gluon field strength tensors , respectively , and the @xmath80 are the gell - mann matrices .
the term @xmath81 contains the lowest - order tvpv four - fermion operators , which include @xmath82 and so forth .
the lowest - dimension tvpc interactions arise in @xmath83 . here
, we consider the following three : @xmath84 the operator @xmath85 was first considered in refs .
@xcite , while @xmath86 was introduced in ref .
the interaction @xmath87 was subsequently considered in ref .
@xcite .
the @xmath88 lagrangian also contains several tvpv operators . among those relevant to us
are @xmath89 the specific forms of other @xmath90 tvpv operators and @xmath91 tvpc operators are not relevant to the present discussion , so we do not list them explicitly .
a key ingredient underlying the expansion of eq . ( [ eq : lnew ] ) is a separation of scales and an associated power - counting scheme .
specifically , the contribution to a t - violating observable from any physics associated with scales @xmath42 is contained in the operator coefficients , @xmath31 .
these short - distance contributions are not calculable since the underlying theory ( _ e.g. _ , renormalizable gauge theory ) responsible for tvpc effects is not known ) in the first place ] .
consequently , the @xmath31 can only be determined from experiment .
contributions from physics having @xmath92 live in loops and many - body ( _ e.g. _ , hadronic ) matrix elements containing the non - renormalizable operators @xmath93 and physical degrees of freedom having masses and momenta less than the scale @xmath29 .
only these long - distance " contributions can be computed using the eft . as a consequence of this scale separation ,
one obtains a systematic power - counting scheme by which to organize contributions to any low - energy t - violating observable . if @xmath32 is the lowest dimension of an effective operator which contributes to such an observable @xmath94 , then @xmath95 where @xmath96 denotes a mass or momentum smaller than @xmath29 . to the extent that @xmath97 , contributions from higher - dimension operators will be suppressed relative to those from @xmath93 by powers of @xmath98 . in general , one may thus truncate the expansion of @xmath94 at the first or first few orders in @xmath98 . in our analysis of the edm
, we find @xmath96 is one of the following : elementary fermion mass , @xmath99 ; weak gauge boson mass , @xmath100 ; qcd scale , @xmath101 ; inverse hadron size , @xmath102 ; and long wavelength photon momentum , @xmath103 .
any renormalization of the @xmath93 by loops involving any one of the @xmath93 and , _
e.g. _ , degrees of freedom appearing in @xmath104 must be carried out in a manner which preserves the eft scale separation . in particular , only intermediate states with momenta and energies below @xmath29 must contribute to loop integrals .
higher - momentum ( @xmath35 ) states have been effectively integrated out in arriving at the expansion in eq .
( [ eq : lnew ] ) .
consequently , we regulate all loop integrals using dimensional regularization ( dr ) , which preserves the separation of scales .
since the subtraction scale @xmath105 arising in dr only appears logarithmically in any regulated amplitude and never as a power , the use of dr does not alter the eft power counting described above .
we emphasize that the cut - off regulator used in refs .
@xcite does not preserve the eft scale separation . in those analyses ,
loop integrals were cut off at momenta @xmath106 .
consequently , in refs .
@xcite the renormalization of the edm @xmath107 due to loops involving any @xmath108 operator scales as @xmath109 and not as @xmath110 as implied by eft power counting .
as we show below , this loss of power counting prevents one from deriving any information about the @xmath88 operators from experimental edm limits . in this respect ,
our conclusions differ dramatically from those of refs .
we illustrate this point in the following section .
the application of eft for new tvpc interactions to the edm of an elementary fermion was considered in ref . @xcite . in
what follows , we summarize the arguments of that analysis , as they illustrate the general principles of eft to be used in the remainder of this study . to that end , we first observe that if , as in scenario b , both new tvpc interactions and pv interactions ( _ e.g. _ , in the standard model ) exist at momentum scales @xmath111 , then there exists no reason to assume that the coefficients of the tvpv effective operators in eq .
( [ eq : lnew ] ) vanish at tree level .
although we have no detailed knowledge of the dynamics of short - distance tvpc and tcpv interactions , nothing prevents their conspiring to generate non - vanishing low - energy tvpv interactions .
in particular , the coefficients of the @xmath112 electric dipole operators , @xmath113 should be non - vanishing at tree - level unless some fortuitous fine - tuning of the short - distance tvpc and tcpv interactions occurs . the situation here is analogous to the chiral expansion of the octet baryon magnetic moments . in the latter case , the leading order contribution occurs at tree level from the @xmath112 magnetic moment lagrangian @xcite : @xmath114\right)\ f^{\mu\nu } \ \ \ , \end{aligned}\ ] ] where the @xmath115 are the octet baryon fields for states of velocity @xmath116 , @xmath117 is the spin operator , and @xmath118 is the scale of chiral symmetry breaking .
the tree level relationship between a baryon magnetic moment and the low - energy constants @xmath119 is @xmath120 where @xmath121 is the mass of the baryon , @xmath122 " denotes its su(3 ) indices , and @xmath123 is the appropriate combination of the @xmath119 . since the baryon magnetic moments are typically of order unity , the tree - level relation implies that the @xmath119 are also of order unity .
alternatively , one may use eq .
( [ eq : mmoments ] ) to estimate the scale @xmath118 . if the low - energy constants in eq .
( [ eq : lmmoments ] ) are of order unity , the one must have @xmath124 gev .
since the chiral symmetry of pionic interactions implies that @xmath125 gev , the tree - level magnetic moment relation produces a self - consistent value for the scale of chiral symmetry breaking when the leading low - energy constants are chosen to be of order one . in a similar way
, one may use the tree - level relation between the edm and the coefficients of the appropriate @xmath112 operators to estimate the scale @xmath29 .
this relation is @xmath126 we follow a standard convention for parameterizing the strength of new physics interactions and take @xmath127 . using the present limit for the edm of the electron @xmath128 e - cm@xcite , one obtains from eq .
( [ eq : edmtree ] ) the limit @xmath129 gev .
thus , if the new tvpc physics is strong " ( @xmath130 ) , one obtains a tremendously large value for the corresponding mass scale . for both eft s in eqs .
( [ eq : lnew ] , [ eq : lmmoments ] ) , loop corrections involving light , dynamical degrees of freedom modify the tree - level relations in eqs .
( [ eq : mmoments],[eq : edmtree ] ) .
single pion loop amplitudes , for example , generate corrections to isovector magnetic moments of @xmath131 relative to the tree - level contribution .
the @xmath53 loops are quadratically divergent , yet generate a finite contribution to the @xmath112 magnetic moment operator .
power - counting implies the appearance of one additional mass factor in this finite contribution .
when dr is used to regulate the integral , this mass factor becomes @xmath132 , resulting in the @xmath133 suppression relative to the tree level relation .
this scaling behavior of the loop contributions , which follows from the eft separation of scales , provides for a chiral expansion in powers of @xmath134 ( where @xmath135 ) which may be reasonably truncated at any order .
similarly , if the eft of eq .
( [ eq : lnew ] ) is well - behaved , one would expect the loop corrections to the relation in eq .
( [ eq : edmtree ] ) to be suppressed by powers of @xmath136 , where @xmath137 is a mass scale associated with the dynamical degrees of freedom in the @xmath138 . for purposes of illustration
, we assume that the only dynamical degrees of freedom operative below @xmath29 are those appearing in the standard model . under scenario
a , additional degrees of freedom such as the right - handed neutral gauge boson in left - right symmetric theories would also generate loop contributions .
these additional degrees of freedom would be required in order for parity symmetry to be restored for @xmath139 . as noted below
, however , the study of only the standard model contributions yields conservative upper bounds on the @xmath88 tvpc operators . in order to avoide introducing model - dependence associated with parity restoration scenarios ,
we restrict our attention to these sm effects . in this case , the leading loop corrections are generated by pv standard model radiative corrections to the @xmath88 operators @xmath140 .
these corrections have been computed in ref .
@xcite for @xmath141 using the diagrams in figs .
[ fig : oneloop_2quark ] and [ fig : twoloop ] .
all of the loop amplitudes corresponding to figs .
[ fig : oneloop_2quark ] and [ fig : twoloop ] are superficially quadratically divergent .
the loops are regulated using dr and the poles removed by the corresponding counterterm in the renormalized @xmath142 in the @xmath143 subtraction scheme .
the closed fermion loop subgraph in fig .
[ fig : twoloop]a is nominally linearly divergent and corresponds to the adler - bell - jackiw anomaly diagram . in this case , the vector current insertions arise from the @xmath18 and @xmath24 couplings to the internal fermion , while the axial vector insertion arises from @xmath144 .
since the edm operator is linear in the photon momentum @xmath145 , we follow ref .
@xcite and retain only the terms linear in @xmath145 arising from this sub - graph . denoting its amplitude @xmath146 , we choose the loop momentum routing to satisfy @xmath147 , where @xmath145 and @xmath148 are the photon and @xmath24-boson momenta , respectively .
the result is the usual anomalous term in @xmath149 . to linear order in @xmath103 ,
there exist three structures which satisfy these vector current conservation conditions : @xmath150 the loop integrals for @xmath146 are nominally linearly divergent . as a check on the calculation
, we regulate the integrals using two different regulators pauli villars and dr and obtain identical results in each case .
the result is finite : @xmath151\int_0 ^ 1\ dx\ { x^2(1-x)\over m_{f'}^2+x(1-x)k^2}\ \ \ .\end{aligned}\ ] ] it is straightforward to verify that when the corresponding term linear in @xmath73 and quadratic in @xmath103 , the photon momentum ) by @xmath152 , @xmath153 , is added to the expression in eq .
( [ abj ] ) and the divergence @xmath149 computed , one obtains the finite , textbook result for @xmath154 @xcite . the closed fermion loop of fig .
[ fig : twoloop]b contains axial vector insertions from @xmath144 and from the coupling of the @xmath24-boson to the internal fermion .
the external fermion line contains the vector current @xmath24-fermion coupling .
this sub - graph diverges quadratically and must be renormalized by the appropriate @xmath155 subtraction before the remaining loop integration is performed . in all cases ,
the amplitudes are infrared - finite .
consequently , we follow ref . @xcite and neglect the fermion mass dependence entering the loops . at leading - log order ,
the results are @xmath156 for the one - loop contribution in fig .
[ fig : oneloop_2quark ] , which contains @xmath157 @xcite .
the loops in fig .
[ fig : twoloop ] , which contain the four - fermion operator @xmath144 , yield @xmath158 for the amplitude of fig .
[ fig : twoloop]a and @xmath159 for the amplitude of fig .
[ fig : twoloop]b . here ,
@xmath160 and @xmath161 denote the sine and cosine of the weinberg angle , respectively , @xmath162 and @xmath163 denote the vector and axial vector couplings of the @xmath24-boson to fermion @xmath164 , and @xmath165 is the corresponding fermion electromagnetic ( em ) charge . since they contribute to the edm at two - loop order , however , we expect them to be no more important than the two - loop contributions involving @xmath85 . ] as expected from general considerations , the results in eqs .
( [ eq : oneloop]-[eq : twoloop ] ) display the @xmath166 suppression relative to the tree - level relation in eq .
( [ eq : edmtree ] ) , where in this case @xmath167 . thus , if the @xmath36 have natural size @xmath168 and if @xmath29 is determined from the experimental limit on @xmath169 via eq .
( [ eq : edmtree ] ) , then the loop corrections to the tree - level edm will be suppressed by more than 25 orders of magnitude for new strong interactions ( @xmath130 ) . in this case , tree - level dominance of the edm implies that edm limits do not provide direct bounds on the @xmath88 operators appearing in eq .
( [ eq : lnew ] ) .
the situation here is analogous to that of the baryon magnetic moments .
since the latter are generally dominated by the tree - level term , they can not be used to determine the meson - baryon couplings which enter the one - loop , sub - leading contributions .
instead , the meson - baryon interaction must be determined directly from , _
e.g. _ , @xmath170 scattering and the results used as input into the chiral loop corrections .
the experimental edm limits could be used to constrain the @xmath88 tvpc operators directly if as in scenario ( a )
@xmath142 vanishes at tree - level in the eft .
such a situation could arise under scenarios , such as left - right symmetric gauge theories , in which parity symmetry is restored at some scale well below @xmath29 . in this case , there would exist no short - distance pv interactions to conspire with the new tvpc physics in generating a tree - level @xmath142 .
the leading contribution to an elementary fermion edm would then be given by the results in eqs .
( [ eq : oneloop]-[eq : compton ] ) , with @xmath100 replaced by , _
e.g. _ , the scale of parity - restoration ( such as the mass of a right - handed gauge boson ) , and with the appropriate combinations of couplings . since the low - energy scale ( _ e.g. _ , @xmath100 ) enters quadratically , a conservative upper bound on the @xmath41 for this scenario can be obtained by using the standard model results given above .
the most stringent constraint results from the one - loop amplitude in eq .
( [ eq : oneloop ] ) applied to the electron edm . using the parameterization of eq .
( [ eq : naturalb ] ) , we obtain @xmath171 tev .
the corresponding two - loop limits are about a factor of five @xmath172 weaker .
the one - loop constraints from the neutron edm are also slightly weaker , given the somewhat less stringent experimental limits on @xmath62 @xcite .
the results of the foregoing analysis have important implications for low - energy direct tvpc searches in light quark systems .
these implications are most transparent under scenario a. in this case , edm limits constrain the ratio @xmath41 via the one- and two - loop results of eqs .
( [ eq : oneloop]-[eq : compton ] ) . as noted in ref .
@xcite , one expects the ratio @xmath22 to scale as @xmath173 where @xmath96 is a momentum characteristic of low - energy hadronic interactions . taking @xmath174 gev@xmath175 and using the one - loop electron edm results , one obtains a limit of @xmath176 . by comparison ,
the present direct tvpc search limits are considerably weaker : @xmath177 . in short , under the parity - restoration scenario ( a ) , the edm results provide the most stringent bounds by many orders of magnitude . for scenario ( b ) in which @xmath142 does not vanish at tree - level ( _ e.g. _ , pv persists at short distances ) , the situation is more subtle . in this case
the edm results do not provide direct constraints on the @xmath88 operators .
nevertheless , one might argue from the edm limits that the low - energy effects of @xmath88 operators should be considerably smaller than the present sensitivity of direct tvpc searches .
comparing eqs .
( [ eq : oneloop]-[eq : compton ] ) to eq .
( [ eq : alphat ] ) , we infer that the low - energy tvpc effects of the @xmath88 operators should be suppressed relative to the corresponding contributions to an edm by @xmath178 if , in addition , the @xmath88 loop contributions are already suppressed relative to the tree - level edm by many orders of magnitude , one would conclude the that corresponding effects in low - energy ( @xmath179 gev@xmath175 ) tvpc processes would be even smaller certainly well below the present direct search sensitivity . nevertheless , this line of reasoning is not airtight . if @xmath180 , it is conceivable that the tree - level and loop contributions to the edm can be comparable in magnitude and that , due to possible cancellations , the magnitude of either term can be considerably larger than the edm limit itself . an analogous situation arises , for example , in the chiral expansion of the isoscalar nucleon magnetic moment . in this case , the leading corrections to the tree - level relation in eq .
( [ eq : mmoments ] ) arise from kaon loops .
since @xmath181 and @xmath118 do not differ appreciably , the loop corrections are consirably larger than the isoscalar magnetic moment . a similarly large short - distance ( tree - level ) contribution
is needed to cancel the loop effect and obtain the small isoscalar magnetic moment . in this case , the use of the isoscalar magnetic moment , together with power counting and nave dimensional analysis , to infer either the size of @xmath118 via the tree - level relation eq .
( [ eq : mmoments ] ) or the size of the one - loop effects would lead to erroneous conclusions .
an independent determination of the strength of the loop contribution ( _ e.g. _ , of the kaon - baryon interaction ) is needed .
should a similar situation obtain for the edm , then direct tvpc searches would still be needed to ascertain the scale of the @xmath88 contributions . before concluding our discussion of the elementary fermion edm , we emphasize the differences between our analysis and that of ref .
@xcite . in that work , a calculation of the amplitudes in fig .
3 was used to try and estimate the size of the short - distance contributions .
this estimate was implemented by regulating the two - loop integrals corresponding to fig .
[ fig : twoloop ] with a form factor of the type @xmath182 the use of this regulator causes the loop integrals to be dominated by contributions from intermediate states having @xmath106 , thereby blurring the separation of scales crucial to the eft expansion of eq .
( [ eq : lnew ] ) .
consequently , the two - loop results of ref .
@xcite scale , incorrectly , as @xmath109 rather than as @xmath46 .
the corresponding implications for low - energy direct tvpc observables are , therefore , erroneous . more generally , as noted in ref .
@xcite , the use of a cut - off regulator destroys the power - counting which justifies truncation of the edm analysis at @xmath88 .
this loss of a systematic expansion can be seen by considering the tower of operators @xmath183 where @xmath184 .
the two - loop calculation of ref .
@xcite may be repeated by replacing the insertion of @xmath85 by each of the operators in eq .
( [ eq : tower ] ) . to regulate the divergences , one may , as the calculation of ref .
@xcite , regulate the integrals with a form factor @xmath185 the corresponding loop integrals will be the same as those in ref . @xcite but with additional factors of @xmath186\\ \label{eq : binomialc } & = & 1+n(p^2/\lamtvs)^{n-1}(p^2/\lamtvs-1)^{-n } - \cdots\ \ \\end{aligned}\ ] ] appearing in the integrand . the first term ( @xmath187 ) on the rhs of eq .
( [ eq : binomialc ] ) will yield the same leading - log contribution as obtained in the calculation of ref .
the remaining terms will generate sub - leading contributions , finite as @xmath188 .
thus , at leading - log order , each operator in the tower will generate the _ same _ contribution , apart from the operator coefficient @xmath189 , so that the edm will be proportional to @xmath190 in this case , there exists no reason to isolate the effects of the @xmath88 operators from those of any other operator in the tower .
all contribute with equal weight .
it would be erroneous , therefore , to truncate the series at @xmath88 , as was done in ref .
@xcite , and to argue that the edm limits constrain the magnitude of only one term in this infinite series .
as this example illustrates , the preservation of the scale separation is crucial to the power counting arguments which justify truncation of the expansion of the edm at a given order .
when dr and @xmath143 subtraction is used , for example , the contributions of each operator in the tower ( [ eq : tower ] ) will be supressed by successive powers of @xmath191 . to the extent that @xmath192 , truncation at @xmath88 makes for a reasonable approximation . in the remainder of this study , we therefore to work with dr and @xmath143 subtraction in treating loop effects .
the previous discussion considered the edm of an elementary fermion . for a composite system such as a neutron , for example , one must also consider many - body contributions involving more than one quark degree of freedom . a generic contribution of this type
is shown in fig .
[ fig : hadronic ] .
the operators which describe these many - quark effects include the tvpv @xmath193 operators listed in eqs .
( [ eq : d6tvpv ] , [ eq : d7tvpv ] ) . as in the case of the single fermion edm @xmath194
, the @xmath195 operators may exist at tree - level in the eft if parity is violated at sufficiently high scales ( scenario b ) .
similarly , these operators may be renormalized by pv radiative corrections to the @xmath88 tvpc operators . as with @xmath194 , the @xmath196
will be dominated by these loop effects if parity symmetry is restored for @xmath197 ( scenario a ) . in what follows ,
we compute the relevant loop effects .
the leading corrections to the @xmath196 are generated by the set of graphs illustrated in fig .
[ fig : oneloop_4quark ] , where the operator inserted is @xmath85 . the diagrams of fig .
[ fig : oneloop_4quark]a were considered previously in the study of ref .
the diagrams for fig .
[ fig : oneloop_4quark]b , which are required by electromagnetic ( em ) gauge invariance , were not included in that analysis .
the inclusion of these graphs is needed in order to obtain the pieces of the @xmath88 tvpv operators containing the photon field .
as we discuss in section 6 , the contributions from these @xmath18-insertion diagrams to the neutron edm are numerically as large as the contributions arising from the graphs of fig .
[ fig : oneloop_4quark]a .
the reason for this situation is relatively straightforward .
the diagrams in fig .
[ fig : oneloop_4quark]a yield the pieces of the @xmath88 operators containing a derivative , as well as contributions to the @xmath198 operators proportional to one power of quark mass operators do not preserve chirality , they can not be induced dynamically by massless quarks . ] .
when the derivative operator acts on quarks inside the hadron , the result is of order @xmath199 .
the resulting contribution to the neutron edm , as in fig .
[ fig : hadronic]a for example , will therefore go as @xmath200 , where @xmath201 is the mass difference between the neutron and one of its excited states ( _ e.g. _ , an unbound @xmath202 pair ) .
the @xmath18-insertion diagrams of fig .
[ fig : hadronic]b do not produce such derivative operators , and their corresponding neutron edm contributions contain no @xmath200 factors . to the extent that @xmath200 is of order one , the magnitude of the two sets of contributions should be comparable .
the amplitudes , @xmath203 , of fig .
[ fig : oneloop_4quark ] are logarithmically divergent .
as before , we regulate the loops using dr and define the finite results using @xmath143 subtraction . at leading - log order , the sum of all twenty six diagrams yields the following linear combination of the @xmath204 : @xmath205 \nonumber \\
+ m_{f^\prime}g^{f^\prime}_a g^{f}_v { \cal o}_{6c}^{f^\prime f } + \left ( g^{f^\prime}_a g^{f}_v - 2 g^{f^\prime}_a g^{f^\prime}_v - 2g^{f}_a g^{f^\prime}_v \right){\cal o}_{7f}^{f^\prime f } \nonumber \\
-{3 \over 2 } g^{f}_a g^{f^\prime}_v { \cal o}_{7h}^{f^\prime f } + { 1 \over 2 } g^{f}_a g^{f^\prime}_v { \cal o}_{7j}^{f^\prime f } \bigr\}\ \ \ .\end{aligned}\ ] ] the expression in eq .
( [ four_quark_symbolyc ] ) is obtained by keeping the external fermion lines off shell .
doing so allows us to identify uniquely the contributions of the @xmath88 tvpv derivative operators and verify the gauge invariance of the overall result . in order to compare our result with the calculation of ref .
@xcite , we use the equations of motion and let the quarks go on shell and convert to momentum space .
we obtain @xmath206 \ \ \ , \end{aligned}\ ] ] where @xmath207 and @xmath208 are the spinors for incoming and outgoing fermions @xmath164 , respectively .
the first three of the terms on the rhs of eq .
( [ on_shell_four_qark_i ] ) are identical to those appearing in @xcite . our coefficient of the fourth term , however , differs from the corresponding expression in @xcite , and the fifth term does not appear in that work at all . we trace part of the difference on the fourth term to diagrams in fig .
5a where the @xmath209 boson connects to initial and final quarks of the same species .
it is unclear from the discussion of ref .
@xcite whether those authors included this class of diagrams . given that our sum of the amplitudes for figs . 5a and 5b satisfies a gauge invariance self - consistency check , we are confident in our result .
the operators obtained in the previous sections can be used to compute contributions to the edm of the neutron and of neutral atoms . doing so requires that one calculate various hadronic matrix elements of two- and four - quark operators .
a first principles treatment of these matrix elements in qcd goes beyond the scope of the present study .
moreover , since we seek only to derive order of magnitude constraints on new tvpc interactions and not to obtain definitive numerical results , it suffices to draw upon various approximation methods . to that end , we turn to the quark model @xcite , factorization , and chiral symmetry .
below , we estimate a number of different matrix elements relevant to the neutron edm and couplings @xmath49 and @xmath210 : 1 . the contribution to @xmath62 from quark edms ( fig .
the relationship between @xmath49 and @xmath88 tvpc operators ( fig .
7a ) and the relationship between @xmath49 and @xmath62 ( fig .
3 . the contribution to @xmath62 from the four - quark / photon tvpv operators appearing in eq .
( [ four_quark_symbolyc ] ) ( fig .
4 . the contribution to the @xmath210 from the purely hadronic terms in eq . ( [ four_quark_symbolyc ] ) ( fig.8a ) and the relationship between the @xmath210 and @xmath62 ( fig .
8b ) . 5 .
the tree - level contribution from @xmath211 ( fig .
* a. two - quark tvpv operators * relating the edm of a constituent quark to that of the neutron using the quark model is a straightforward exercise . as shown in ref .
@xcite , this relationship is given by @xmath212\int\ d^3x\ \left(u^2+{1\over 3 } \ell^2\right)\ \ \ , \ ] ] where @xmath213 and @xmath214 are the upper and lower component quark model radial wave functions , respectively .
the integral in eq .
( [ eq : qmrel ] ) can be estimated using the wave function normalization condition @xmath215 and expression for the axial vector charge @xmath216 where @xmath217 . from eqs .
( [ eq : wavenorm]-[eq : axial ] ) one obtains @xmath218 * b. four - quark tvpc operators * deriving the relationship between the effective hadronic coupling @xmath49 and the four - quark @xmath88 tvpc operators ( fig .
7a ) requires more thought than in the case of evaluating two - quark matrix elements . for simplicity , we focus on the four - quark operator @xmath219 .
we make a simple estimate using factorization . doing so requires use of the fierz transformed version of this operator , since the interaction in eq .
( [ eq : lrhonn ] ) involves only @xmath220 .
the fierzed form of the operator is @xmath221 in the factorization approximation , one makes the replacement @xmath222 where @xmath223 and @xmath224 denote nucleons , @xmath225 is the field for a quark of flavor @xmath122 and color @xmath226 , and @xmath227 denotes any of the products of quark bilinears appearing in eq .
( [ eq : fierzed1 ] ) .
note that since hadrons are color singlets , one has @xmath228 where repeated indices are summed over .
hence , each factorization contribution contains a factor of @xmath229 since any pieces of the interaction above which involve the @xmath230 will not give rise to a @xmath47 meson - vacuum matrix element , we are concerned with only the first eight terms of the fierzed interaction .
we may also reduce the number of terms to be evaluated using the equations of motion . of the resulting operators ,
we keep only those containing no powers of the quark mass , since the latter generate significant suppression factors . in the case where @xmath164 represents an up quark and
@xmath231 represents a down quark , the remaining structures are @xmath232 in the factorization approximation , the matrix elements required are @xmath233 plus the corresponding matrix elements for @xmath234 ( note that in eq .
( [ eq : rhome ] ) the color indices have been suppressed for simplicity ) .
a detailed evaluation of these matrix elements appears in appendix a. the resulting tvpc @xmath48 lagrangian is @xmath235 ( \tau^- \rho^+_\mu \ - \tau^+ \rho_\mu^- ) n\end{aligned}\ ] ] it is customary to write the standard rho - nucleon lagrangian and the tvpc rho - nucleon lagrangian in the respective forms ( see , _
e.g. _ eq . ( [ eq : lrhonn ] ) , @xmath236 @xmath237 where @xmath238 is the anamolous isovector magnetic moment , @xmath239 is the mass of the nucleon and @xmath240 is the @xmath241 coupling constant . by redefining the phase of the rho meson , we can eliminate the dirac structure @xmath242 in eq .
( [ eq : todd_rho_lagrang ] ) .
we begin by writing eqs .
( [ eq : todd_rho_lagrang ] ) and ( [ eq : smrhonuc ] ) together as @xmath243 \tau^- \rho^+_\mu n \nonumber \\ + \sqrt{2 } f_\rho \bar{n } \left [ \gamma^\mu ( 1 - { ia \over f_\rho } ) + i { \kappa_v \over 2 m_n } ( 1 - ib { 2 m_n \over f_\rho \kappa } ) \sigma^{\mu \lambda } q_\lambda \right ] \tau^+ \rho^-_\mu n\end{aligned}\ ] ] where @xmath244\ ] ] @xmath245 we then observe that since @xmath246 and @xmath247 $ ] , the phases of the rho mesons can be redefined as , @xmath248 and @xmath249 .
this allows us to rewrite the total lagrangian as @xmath250 ( \tau^- { \tilde \rho}^+_\mu + \tau^+ { \tilde \rho}_\mu^- ) n \nonumber \\ - \sqrt{2 } f_\rho \bar{n } \left({b 2 m_n \over f_\rho \kappa_v } - { a \over f_\rho } \right ) { \kappa_v \over 2 m_n } \sigma^{\mu \nu } q_\nu ( \tau^- { \tilde \rho}^+_\mu - \tau^+ { \tilde \rho}_\mu^- ) n\end{aligned}\ ] ] the quantity @xmath251 is then @xmath252\end{aligned}\ ] ] where @xmath253 as calculated in ref .
@xcite , @xmath49 may contribute to the neutron edm via the loop diagrams of fig .
7(b ) as well as to the atomic edm via process like those shown in figs .
1(a , b ) . the @xmath62 calculation of ref .
@xcite included the introduction of form factors at the hadronic vertices in order to render the loop integrals finite .
the result is @xmath254 where @xmath255 and the pv pion - nucleon coupling @xmath256 is constrained by the pv @xmath18-decay of @xmath257f to lie in the range : @xmath258 , where @xmath259 characterizes the strength of the charged current @xmath23 hadronic interaction .
the function @xmath260 depends on the masses appearing in the loop integral as well as the form factor cut - off parameter , @xmath261 .
we note that the use of eq .
( [ eq : grhobardn ] ) and the experimental limits on @xmath26 to derive bounds on tvpc interactions entails several ambiguities .
first , the value of @xmath262 measured in nuclei such as @xmath257f may differ from the value appropriate to the single nucleon or few - nucleon systems .
many - body nuclear effects may renormalize the long range pv @xmath263 interaction in such a way as to shift the value of the effective pv @xmath54 coupling from the value appropriate for eq .
( [ eq : grhobardn ] ) .
this ambiguity may be resolved by future experiments , such as the measurement of the @xmath264 asymmetry planned at lansce@xcite .
second , the use of a form factor to render the loop integral finite can introduce considerable ambiguity . the form factor chosen in ref .
@xcite was taken from the bonn potential , with @xmath265 gev .
one may just as well have chosen a cut - off given by the inverse size of the hadron .
the variations due to this spread of choices can be significant @xcite . moreover , as argued in section 3 , the use of cut - off regulators can render one s effective theory devoid of any systematic power counting , leaving it poorly defined . for these reasons
, we will not use eq .
( [ eq : grhobardn ] ) to derive limits on the @xmath88 tvpc operators . * c. four - quark tvpv operators * we consider the relationship between the @xmath198 and @xmath88 tvpv operators and the neutron edm .
we specify here to scenario ( a ) , in which case the tvpv operators arise entirely from pv radiative corrections to the tvpc operators , as in fig .
first , we estimate the contribution from the pieces of the @xmath88 operators containing the photon field .
these contributions can be understood diagrammatically as shown in fig 4(b ) .
starting from expression in eq .
( [ four_quark_symbolyc ] ) , we see that the tvpv @xmath18-four - quark interaction can be written as @xmath266 { \bar\psi}_{f } \sigma_{\lambda \mu } { \psi}_{f } { \bar\psi}_{f^\prime } \gamma_5 \gamma_\mu{\psi}_{f^\prime } \nonumber \\ & & + g^{f^\prime}_v g^{f}_a 3i{\bar\psi}_{f } \gamma_5 { \psi}_{f } { \bar\psi}_{f^\prime } \gamma_\lambda { \psi}_{f^\prime } \biggr ] a^\lambda\ \ \ .\end{aligned}\ ] ] for simplicity , we consider the case where @xmath164 is an up quark and @xmath231 denotes a down quark . as in our estimate of @xmath49
, we can find neutron matrix elements using the quark model .
we identify the appropriate quark model expression for the electric dipole moment , @xmath267 here , @xmath268 is a neutron in the @xmath269 state , the @xmath270 are used in the wave packet description of the momentum eigenstates as discussed in appendix a , and the @xmath271 are the neutron spinors .
choosing @xmath272 we have @xmath273 we evaluate the two matrix elements needed and find @xmath274 @xmath275 using these matrix elements , we obtain the following expression for the neutron edm : @xmath276\end{aligned}\ ] ] since @xmath277 in the standard model , this contribution to the neutron electric dipole moment is approximately proportional to @xmath278 .
the second way tvpv operators contribute to @xmath62 is by mixing states of opposite parity into the neutron ground state ( fig .
the lightest state which may contribute is the @xmath279 s - wave . in the chiral limit
, its contribution is dominated by the loop diagrams of fig .
the tvpv @xmath280 couplings are just the @xmath55 , generated from the tvpv quark operators as in fig .
the @xmath55 also contribute to the atomic edm via the meson - exchange interaction of fig .
in what follows , we relate the @xmath55 to the purely hadronic parts of the operators in eq .
( [ four_quark_symbolyc ] ) . following ref .
@xcite , we carry out the calculation in the factorization approximation while using the on - shell form of the tvpv operators given in eq .
( [ on_shell_four_qark_i ] ) .
first , we consider the tvpv @xmath281 coupling .
the only dirac structures which give rise to pion - vacuum matrix elements are @xmath230 , and @xmath282 .
therefore the structure in the third line does not contribute .
the last two structures in eq .
( [ on_shell_four_qark_i ] ) do not contribute in the factorization approximation .
this conclusion follows from symmetry arguments , @xmath283 , and the equations of motion @xmath284 .
for the remaining structures the pcac relation @xmath285 ) and using the consequences of isospin symmetry @xmath286 yields the following two matrix elements needed in the calculation : @xmath287 @xmath288 in addition , we require the nucleon matrix elements of the light quark scalar densities . from the quark model and @xmath170 sigma term ,
one obtains @xmath289 @xmath290 using these results , we obtain for the tvpv neutral pion nucleon couplings @xmath291 and @xmath292 , the strength of the couplings are : @xmath293\ ] ] @xmath294\ ] ] here again , since @xmath295 at tree level in the standard model , the terms proportional to @xmath296 make the largest contribution . in order to determine the couplings of the charged pion with the nucleon
, we fierz transform eq .
( [ on_shell_four_qark_i ] ) .
the result is listed in appendix b. as discussed in appndix b , we estimate the strength of the @xmath297 coupling using the first two terms in the fierz transformed expression
. the remaining terms do not contribute in the factorization approximation .
in this case , we require the following matrix elements : @xmath298 @xmath299 the coefficient can then be written as , @xmath300 the last piece @xmath301 comes from applying equations of motion to the derivative terms , as discussed in appendix b. expressed in terms of a , b and c , the pvtv pion - nucleon couplings from eq .
( [ eq : lpinn ] ) are @xmath302 @xmath303 @xmath304 in the chiral expansion of the edm , the dominant contribution to @xmath62 from the tvpv @xmath54 interaction arises from the loops in fig .
8(b ) , where the intermediate state contains a @xmath305 . in this case , only the constant @xmath306 contributes .
the loop calculation , first performed in ref .
@xcite , yields @xmath307 written explicitly in terms of the tvpc scale , @xmath308 , the neutron electric dipole moment takes the form , @xmath309 where we have kept only the leading , non - analytic loop contribution proportional to @xmath310 and where @xmath311 is the strong @xmath54 coupling .
we note that an evaluation of the analytic contributions has been performed using sidewise dispersion relations in ref .
in addition , loop contributions involving the @xmath312 intermediate state have been considered in refs .
@xcite .
* d. tree - level contributions * there exists one way in which the @xmath88 tvpc operators may contribute to @xmath62 without the consideration of loop effects . as shown in fig . 9
, the operator @xmath87 generates a tree - level contribution when the @xmath209 is exchanged and couples to the second quark s axial vector current .
navely , one might expect this contribution to compete with the one - loop quark edm generated by @xmath87 in fig .
2 . the calculation of the process in fig .
9 is tedious but straightforward . using the quark model , we obtain @xmath313\left(\mz\over\lamtv \right)^2\left[{0.34\over m_{{\scriptscriptstyle z}}^3\mw r^4_n}\right]\\ \nonumber & \times&\int_0^{2.04}\ x^2dx\left[\j_0(x)^2j_0'(x)j_1(x ) -j_0(x)j_1(x)^2j_1'(x)-j_0(x)j_1(x)^3/x\right]\\ \nonumber & \times&\sum_{m , m ' , \alpha,\ldots}\sigma^3_{mm}\bra{n1/2 } u^{\dag}(m',\alpha ' )
u(m',\alpha ) d^{\dag}(m,\beta ' ) d(m,\beta)\ket{n 1/2}\ \ \ , \end{aligned}\ ] ] where the sum runs over all spin ( @xmath314 ) and color @xmath315 indices . evaluating the integrals and quark model contractions
yields @xmath316\left(\mz\over\lamtv \right)^2\left[{0.162\over m_{{\scriptscriptstyle z}}^3\mw r^4_n}\right]\ \ \ .\ ] ] comparing the result in eq .
( [ eq : dntree ] ) with one - loop amplitude in eq .
( [ eq : oneloop ] ) , we observe that the tree - level contribution contains the suppression factor @xmath317 the difference results from the contributions of high - momentum ( @xmath318 ) intermediate states in the loops of fig .
2 . the tree - level process , in contrast ,
is dominated by states with momenta @xmath319 . since the tvpc operator is has dimension seven , and since the momentum transferred through @xmath209-boson propagator in fig .
9 is @xmath199 , one would expect a suppression factor of @xmath320 for the process of fig .
the foregoing analysis allows us to make a quantitative connection between various observables and effective coupling constants on the one side and the eft of tvpc interactions of quarks and gauge bosons on the other . in this section
, we use those relationships to derive bounds on the tvpc interactions .
it is also instructive to interpret these bounds in terms of the mass scale , @xmath29 . to that end , we adopt the parameterization of eqs .
( [ eq : naturala ] , [ eq : naturalb ] ) and present constraints in terms of @xmath29 and @xmath321 . in
what follows , we use experimental @xmath25 and @xmath62 limits directly as well as the limits on @xmath49 and the @xmath210 derived from several sources .
the corresponding limits on these quantities are given in tables 1 and 2 .
it is also useful to convert our expressions for @xmath49 and the @xmath210 into numerical form . in the case of
@xmath49 we have @xmath322 similarly , the combinations of the @xmath210 relevant to @xmath62 and @xmath66 are , respectively , @xmath323 for @xmath26 and @xmath324 for @xmath66 . * scenario a. * under this scenario , the first two terms in eq .
( [ eq : edmexpand ] ) vanish , so that the leading terms in the expansion of the edm are the @xmath325 contributions from the tvpc operators . as in the case of the effective hadronic couplings ,
it is useful to express our relationships between the edm s and the tvpc interactions in numerical form . for the elementary fermion edm s arising from the one - loop graphs containing @xmath87 ( fig .
2 ) , we require a choice for @xmath105 . since the typical momentum of a quark in the neutron is @xmath326 , we take @xmath327 .
the appropriate choice for @xmath169 is probably smaller , on the order of the typical momentum of an electron bound in a heavy atom .
since the @xmath105-dependence of @xmath328 is only logarithmic , however , the precise choice of scale is not decisive . to be conservative
, we also use @xmath327 for @xmath169 . the corresponding results are @xmath329 and @xmath330 where we have used the naturalness relation of eq .
( [ eq : naturalb ] ) to express @xmath331 in terms of @xmath332 and where the 0.88 factor in eq .
( [ eq : dnnuma ] ) is the value of the bag model integral in eqs .
( [ eq : qmrel ] ) and ( [ eq : qmintegral ] ) .
since the two - loop effects involving @xmath85 are suppressed relative to the one - loop @xmath87 contributions , we do not give explicit numerical formulae for the former .
for the contribution to @xmath26 from the tvpv @xmath18-four quark interaction of fig .
4b , generated by the loops of fig .
5b , we find @xmath333 while the process of fig .
9 yields @xmath334 where we have again used eq .
( [ eq : naturalb ] ) .
we now apply these expressions to the experimental limits given in tables 1 and 2 .
the results are listed in table 3 .
clearly , the one - loop elementary fermion edm s generated by radiative corrections to @xmath87 yield the most stringent bounds on @xmath29 . whether one uses @xmath26 or @xmath169 ,
the lower bounds generated by this mechanism are roughly three orders of magnitude larger than those obtained from the many - quark effects .
the reason for this difference is clear .
the one - loop graphs are dominated by high - momentum ( @xmath318 ) intermediate states , whereas the many - quark matrix elements are governed by physics at scales @xmath335 .
since the dimension of the tvpc operators is greater than that of the edm by two , the edm must contain at least two powers of the relevant mass scale .
hence , the one - loop @xmath87 effects should be at least @xmath336 larger than the many - quark effects in @xmath26 .
we emphasize the comparison between the one - loop elementary fermion edm limits and those obtained from @xmath66 . in the latter case ,
the strongest bounds are derived from the combination of the @xmath210 appearing in eq .
( [ eq : gpibarnumb ] ) . since these couplings scale as @xmath46
, one would require an improvement of nine orders of magnitude in the experimental limits on @xmath66 in order for the atomic edm to compete with @xmath169 and @xmath26 . ] .
we also note that were it not for the one - loop elementary fermion limits , the diagrams of fig .
5b required by gauge invariance but omitted in ref .
@xcite would yield the most stringent constraints on new tvpc interactions under scenario ( a ) . by far the weakest bounds follow from the extraction of @xmath49 from @xmath66 .
the limits in table 3 can be translated into an expectation for @xmath22 under scenario ( a ) . using the general scaling arguments of ref .
@xcite which imply @xmath337 and using the naturalness assumption of eq .
( [ eq : naturalb ] ) we have @xmath338 for a low energy tvpc process with @xmath339 gev , the scenario ( a ) edm limits of table 3 imply that the size of the effect should be roughly @xmath340 .
current direct search limits are at the @xmath341 level for @xmath22 .
* scenario b. * as argued above , experimental edm limits do not apply to tvpc interactions when pv persists at short distances . in this case , one must rely on direct tvpc tests , from which limits on @xmath49 may be extracted . the most precise determinations of @xmath49 from tvpc observables are obtained from charge symmetry breaking ( csb ) terms in @xmath19 scattering cross sections @xcite and fivefold correlations ( fc ) in the transmission of polarized neutrons through spin aligned holmium @xcite .
the limits on @xmath49 obtained from these observables are given in table 2 .
in addition , a proton - deuteron transmission experiment proposed for cosy would be sensitive to @xmath49 at the @xmath341 level @xcite .
further improvements in the @xmath19 csb limits may also be possible @xcite .
the bounds on @xmath29 from direct tvpc searches are given in table 4 .
at present , the data imply new tvpc physics could arise at scales as light as a few gev .
ideally , future experiments would push these bounds closer to the weak scale .
such a sensitivity would be comparable to the present and anticipated new physics sensitivities of atomic pv and pv electron scattering experiments @xcite .
achieving such sensitivity , however , would be a daunting task , requiring at least six orders of magnitude improvement in precision beyond the present state of the art .
measurements of edm s provide one of the most powerful probes of possible new physics . thus far
, null results for @xmath169 , @xmath26 , and @xmath28 lead to tight constraints on a variety of new physics scenarios . in this paper , we have studied the implications of edm s for tvpc new physics .
we have also developed relationships between effective tvpc and tvpv hadronic couplings , on the one hand , and tvpc interactions involving quarks and gauge bosons , on the other .
this relationship has not been systematically delineated in the past .
consequently , some confusion about the respective implications of edm s and the values of hadronic couplings extracted from direct tvpc searches has ensued .
we believe that our analysis has helped clarify these implications . while the possible origin of tvpc interactions in the context of a renormalizable gauge theory has yet to be delineated ,
one may , nevertheless , address the issue using the framework of eft .
this framework affords a systematic method for treating nonrenormalizable interactions and for carrying out the associated phenomenology .
since the operators we have treated here have @xmath342 , eft is the natural framework for performing our analysis .
as emphasized above , a key ingredient in the application of eft to this problem is the maintenance of a scale separation .
the short - distance ( @xmath35 ) tvpc physics about which we are ignorant is parameterized by the _ a priori _ unknown coefficients of the nonrenormalizable effective interactions . long distance ( @xmath97 ) physics may be treated explicitly in the guise of loops and many - body matrix elem ents involving degrees of freedom having masses and momenta below @xmath29 . as we illustrated for the example of the cut - off regulator used in the loop calculations of refs .
@xcite , a failure to adhere to this scale separation can lead to disastrous results .
indeed , in that example , the blurring of scale separation implies ( erroneously ) that the edm is proportional to an infinite series of unknown coefficients of nonrenormalizable operators of arbitrarily high dimension . without the truncation scheme implied by the eft scale separation ,
one is unable to learn anything from experiment about a given tvpc interaction ( or even finite set of such interactions ) . as argued at the outset of our paper
, the eft scale separation implies one must consider the low - energy consequences of tvpc new physics under two scenarios : parity - restoration below @xmath29 ( scenario a ) and parity - restoration above @xmath29 ( scenario b ) . in the case of scenario
a , the edm is dominated by pv radiative corrections to @xmath88 tvpc operators .
the most significant impact arises from the edm of elementary fermions , generated by a one - loop graph involving @xmath87 .
interpreted in terms of a mass scale , this mechanism , taken with the experimental limits on @xmath169 and @xmath26 , imply that @xmath343 100 - 250 tev when the tvpc new physics is strong " ( @xmath344 ) .
the impact of many - quark effects which contribute both directly to @xmath26 as well as to both @xmath26 and @xmath66 via the couplings @xmath49 and @xmath55 is considerably smaller than that of the one - loop elementary fermion edm .
indeed , for the many - quark mechanism to compete with that of the elementary fermion edm , one would require an improvement in the limits on @xmath66 of nine orders of magnitude .
the impact of direct search limits is even weaker under this scenario .
we conclude , then , that @xmath169 and @xmath26 provide the most powerful probes of new tvpc physics when parity symmetry is restored at short distances , while the atomic edm measurements are better suited to constraining other new physics scenarios . under scenario b , the implications of edm s are ambiguous at best .
the existence of parity - violation at short distances implies that @xmath345 tvpv operators contribute to the edm along with the radiatively corrected @xmath91 tvpc operators .
consequently , no single tvpv observable constrains the latter unless one invokes additional assumptions .
in contrast , the relationship between tvpc observables such as csb @xmath19 scattering cross sections or fc in neutron transmission is not clouded by the additional unknown constants which enter the edm . in this case , however , the limits obtained from direct tvpc searches are rather weak : @xmath343 1 gev ( for @xmath344 ) . as a long term benchmark , one would ideally search for new tvpc physics at least up to the weak scale . achieving this goal
would require an improvement in the precision of direct tvpc searches by six orders of magnitude .
while achieving such improvement would seem formidable , any intermediate progress would constitute a welcome step .
it is a pleasure to thank d. demille , r. springer , and m. savage , and wim van oers for useful discussions .
this work was supported in part under u.s .
department of energy contracts # de - ac05 - 84er40150 and # de - fg06 - 90er40561 , and # de - fg02 - 00er41146 and a national science foundation young investigator award .
99 the particle data group , euro .
journal * c15 * , 1 ( 2000 ) . s.c . bennett and c.e .
wieman , phys .
rev . lett . * 82 * , 2484 ( 1999 ) ; c.s .
et al . _ , science * 275 * , 1759 ( 1997 ) .
a. derevianko , phys .
* 85 * , 1618 ( 2000 ) .
ramsey - musolf , phys . rev . *
c60 * , 015501 ( 1999 ) .
ramsey - musolf , phys . rev . *
d62 * , 056009 ( 2000 ) .
a. angelopoulos _
( cplear collaboration ) , phys .
b444 * , 43 ( 1998 ) .
e. blanke _
et al . _ ,
lett . * 51 * , 355 ( 1983 ) .
f. boehm in _ fundamental interactions in nuclei _
, w. haxton and e. haxton , eds .
( world scientific , singapore , 1995 ) p. 67 .
p. herczeg , j. kambor , m. simonius , parity conserving time reversal violation in flavor conserving quark - quark interactions " , los alamos theory division report ( unpublished ) ; p. herczeg , hyp .
interact . *
75 * ( 1992 ) 127 ; p. herczeg in _ fundamental interactions in nuclei _ ,
w. haxton and e. haxton , eds .
( world scientific , singapore , 1995 ) p. 89 .
et al . _ ,
phys . rev .
* c55 * , 2684 ( 1997 ) .
et al . _ ,
b267 * , 23 ( 1991 ) .
m. simonius , pys .
lett . * 78 * , 4161 ( 1997 ) .
cosy proposal , p.d .
eversheim , spokesperson ; f. hinterberger ( for the tri collaboration at cosy ) , contribution to the fifth annual wein symposium : a conference on physics beyond the standard model , nucl - ex/9810013 .
van oers , hep - ph/9912454 , to appear in int .
phys . * e*. l.j .
( emit collaboration ) , phys .
rev . * c62 * , 055501 ( 2000 ) . i.b .
khriplovich , nucl .
b352 * ( 1991 ) 385 .
r.s . conti and i.b .
khriplovich , phys .
* 68 * ( 1992 ) 3262 .
j. engel , p.h .
frampton , r.p .
springer , phys . rev .
* d53 * ( 1996 ) 5112 .
p. herczeg , time reversal violation in nuclear processes " , in _ symmetries and fundamental interactions in nuclei _ ,
haxton and e.m .
henley , eds .
, world scientific , singapore , pp .
89 - 126 .
ramsey - musolf , phys .
lett . * 83 * , 3997 ( 1999 ) ; ( e ) _ ibid _ * 84 * 5681 ( 2000 )
ramsey - musolf , hep - ph/0010023 , to appear in proceedings of the workshop on fundamental physics with pulsed neutron beams , durham , n.c .
, may , 2000 .
ramsey - musolf and h. ito , phys
* c55 * , ( 3066 ) . c. itzykson and j .- b .
zuber , quantum field theory , mcgraw - hill , new york , 1980 .
note that the result in eq .
( [ eq : oneloop ] ) is a factor of six larger than eq .
( 6 ) of ref .
@xcite , which contains an error .
donoghue , e. golowich , and b. holstein , phys .
* 131 * , 319 ( 1986 ) .
sakuri , j. j. , currents and mesons , university of chicago press , chicago , 1969 .
w. c. haxton , a. horing , and m. j. ramsey - musolf , phys rev * d50 * 3422 .
w.m . snow _
et al . _ ,
* a440 * , 729 ( 2000 ) .
k. f. smith _
_ phys . lett . *
b234 * ( 1990 ) 191 , i. s. altev _
b276 * ( 1992 ) 242 , p. g. harris _
* 82 * ( 1999 ) 904 .
crewther , p. di vecchia , g. veneziano , and e. witten , phys .
b88 * , 123 ( 1979 ) ; _ ibid _ * 91 * , 487 ( e ) ( 1980 ) .
g. barton and e.g. white , phys . rev . * 184 * , 1660 ( 1969 ) .
g. valencia , phys . rev .
* d41 * , 1562 ( 1990 ) .
he and b. mckellar , phys . rev .
* d47 * , 4055 ( 1993 ) . e.d .
et al . _ ,
phys . rev . * a50 * , 2960 ( 1994 ) .
b. desplanques , j.f .
donoghue , and b. holstein , ann .
( ny ) * 124 * , 449 ( 1980 ) .
phys . rev . * a52 * , 3521 ( 1995 ) .
here , we evaluate the matrix elements appearing in eq . ( [ eq : rhome ] ) . we define the following quantities : @xmath346 , @xmath16 is the @xmath47 polarisation . since the tvpc @xmath48 lagrangian is linear in the @xmath47 momentum operator , we retain only those terms in the matrix element products linear in @xmath103 .
the fourth matrix element product in eq . ( [ eq : rhome ] ) is quadratic in @xmath103 since the nucleon matrix element vanishes to first order in @xmath103 .
thus , the last producet does not contribute to @xmath49 .
for the remaining matrix element products in eq .
( [ eq : rhome ] ) , we first require the @xmath47 to vacuum matrix element , parameterized in the standard way as @xmath347 where @xmath348 .
the rho matrix element with the derivative may be evaluated using wave packets in the quark model , using the approach of donoghue and johnson @xcite .
we begin by assuming the following structure for the matrix element , @xmath349 where @xmath350 is the momentum of the rho - meson and @xmath351 is the polarization vector .
we solve for the coefficients , @xmath122 and @xmath352 .
the quark model matrix element may be written using wave packets as @xmath353 multiplying each side of the equation by @xmath354 , results in @xmath355 the expression for @xmath270 may be obtained from the vector current matrix element eq .
( [ eq : vc ] ) , @xmath356 we use the @xmath357 component of the above equation to find an expression for @xmath270 .
this , taken together with the vector current matrix element produces , @xmath358 upon evaluating both quark model matrix elements , we find @xmath359 , while , @xmath360 r^2 dr / \int ( u^2 - l^2/3 ) r^2 dr}.\ ] ] here , @xmath213 and @xmath361 are the upper and lower components of the quark model wave functions , respectively .
the final expression is @xmath362 in this expression , @xmath363 is the radius of the rho meson in the bag model . in the following
we shall also use @xmath364 for the bag model radius of the nucleon .
the nucleon matrix elements also may be evaluated with the quark model .
@xmath365 @xmath366 where @xmath367 is quark energy inside the bag . for the third matrix element product ,
we obtain using similar methods and eqs .
( [ eq : wavenorm ] , [ eq : axial ] ) @xmath368 where the @xmath369 denote terms higher order in @xmath103 .
in calculating the tvpv @xmath370-nucleon couplings , we require the fierz transformed form of the amplitude in eq .
( [ on_shell_four_qark_i ] ) .
the transformed amplitude is only the first two terms on the rhs of eq .
( [ eq : fierz_fourquark1 ] ) contribute in the factorization approximation .
the terms on the second line , such as @xmath372 will not contribute in this approximation .
the reason is that the pion - to - vacuum matrix element will be proportional to the pion momentum , @xmath103 , which equals the change of the momentum of the nucleon .
when contracted with the nucleon matrix element of the vector current , the result vanishes by current conservation ( neglecting the small @xmath373-@xmath96 mass difference ) .
the terms on the third line will not contribute for symmetry reasons .
the remaining terms involve derivatives on the quark fields . in each case
the derivative pair is split , so one derivative acts on a quark field in one quark bilinear and the second derivative acts on the second .
some of these terms may be rewritten using equations of motion for the on shell quarks .
doing so yields new contributions to the first four lines of the rhs of eq .
( [ eq : fierz_fourquark1 ] ) , requiring an adjustment in the coefficients .
the remaining terms involve derivative operators , such as @xmath374 factorization matrix elements of these operators vanish by either symmetry or current conservation . | we construct the relationship between nonrenormalizable , effective , time - reversal violating ( tv ) parity - conserving ( pc ) interactions of quarks and gauge bosons and various low - energy tvpc and tv parity - violating ( pv ) observables . using effective field theory methods , we delineate the scenarious under which experimental limits on permanent electric dipole moments ( edm s ) of the electron , neutron , and neutral atoms as well as limits on tvpc observables provide the most stringent bounds on new tvpc interactions . under scenarios in which parity invariance is restored at short distances ,
the one - loop edm of elementary fermions generate the most severe constraints .
the limits derived from the atomic edm of @xmath0hg are considerably weaker .
when parity symmetry remains broken at short distances , direct tvpc search limits provide the least ambiguous bounds .
the direct limits follow from tvpc interactions between two quarks .
# 1#1 # 1#1 # 1#2 # 1@xmath1 # 1 # 1#1 # 1^#1 # 1#1 # 1#1 # 1#1 # 1_phys .
rev . _ * # 1 * # 1_phys .
rev .
_ * c#1 * # 1_phys . rev .
_ * d#1 * # 1_phys .
rev .
lett . _ * # 1 * # 1_nucl .
phys . _ * a#1 * # 1_nucl .
phys . _ * b#1 * # 1_ann . of phys . _ * # 1 * # 1_phys .
reports _ * # 1 * # 1_phys .
lett . _ * b#1 * # 1_z .
fr phys .
_ * a#1 * # 1_z .
fr phys .
_ * c#1 * 0q_w^0 a. kurylov@xmath2 , g. c. mclaughlin@xmath3 , and m.j .
ramsey - musolf@xmath4 @xmath5 department of physics , university of connecticut , storrs , ct 06269 usa + @xmath6 department of physics and astronomy , state university of new york , stony brook , ny 11794 usa + @xmath7 theory group , thomas jefferson national accelerator facility , newport news , va 23606 usa + @xmath8 previous address : triumf , 4004 wesbrook mall , vancouver , b.c .
canada v6t2a3 + |
the nuclear matter symmetry energy , which is defined as the difference in energy per nucleon between the pure neutron matter and the symmetric nuclear matter , is an important quantity that determines the properties of objects such as the atomic nucleus and the neutron star @xcite .
the study of symmetry energy and its dependence on nuclear density and temperature is currently a subject of great interest @xcite .
theoretically , the symmetry energy can be determined from microscopic calculations such as the brueckner - hartree - fock ( bhf ) and the dirac - brueckner - hartree - fock ( dbhf ) calculations , or the phenomenological calculations such as the skyrme hartree - fock ( shf ) and the relativistic mean field ( rmf ) calculations @xcite .
these calculations currently predict wide range of symmetry energies for densities below and above normal nuclear density , @xmath0 = 0.16 @xmath1 ( fig .
1 ) . experimentally , the symmetry energy is not a directly measurable quantity and has to be extracted indirectly from observables that are related to the symmetry energy .
the experimental determination of the symmetry energy is therefore dependent on how reliable the model that describes the experimental observable is .
so far there has been very few experimental determination of the symmetry energy .
these experimental studies are of two types : 1 . where , a certain form of the density dependence of the symmetry energy is assumed in the theoretical calculation and the experimental observable reproduced using the dependence that best explains the data .
such studies @xcite often make use of dynamical models such as , the isospin boltzmann uehling uhlenbech ( ibuu04 ) @xcite , the improved quantum molecular dynamics ( imqmd ) @xcite , the boltzmann nordheim vlasov ( bnv ) transport calculation @xcite and the anti symmetrized molecular dynamics ( amd ) @xcite , to relate the symmetry energy to the experimental observable .
the disadvantage in such studies is that they assume a single form of the symmetry energy for all densities , which may not be true .
such studies provide only a gross dependence of the symmetry energy without much insight into its evolution with the density . also , these studies are highly model dependent as will be discussed in a later section .
where , the form of the density dependence of the symmetry energy is not known a priori and the symmetry energy is studied by mapping its value at each density .
such studies @xcite require a detailed understanding of the relation between the symmetry energy , excitation energy , density and temperature .
they make use of the statistical multifragmentation model ( smm ) @xcite to relate the symmetry energy to the experimental observable .
the advantage in such studies is that they provide a direct " means of studying the symmetry energy with density and temperature that can then be compared with any theoretical predictions .
the difficulty in such studies is often the theoretical interpretation of the symmetry energy .
.,width=5,height=2 ] in this article , we review recent experimental studies on the symmetry energy and its density dependence obtained using the above two approaches .
the symmetry energy below saturation density has been studied in heavy ion multifragmentation reactions using fragments with a @xmath6 4 .
observables such as the fragment yield , isoscaling parameter , isospin diffusion transport ratio , double neutron - proton ratio , pre - equilibrium emission , and @xmath7n@xmath8/z has been used to extract the symmetry energy .
the symmetry energy in this density region has been studied by comparing the experimental observables with the dynamical model calculations .
@xcite , compared the nscl - msu isospin diffusion data from @xmath9sn + @xmath10sn reaction at 50 mev / a with the dynamical ibuu04 @xcite calculation , and obtained a symmetry energy of the form , e@xmath11 = 31.6(@xmath5/@xmath0)@xmath12 with @xmath13 = 1.05 .
the nscl - msu isospin data was also compared with the ibuu04 by li _
@xcite , that included the isospin dependence of the in - medium nucleon - nucleon cross - section , giving symmetry energy of the form , e@xmath11 = 31.6(@xmath5/@xmath0)@xmath12 with @xmath13 = 0.69 .
@xcite , extracted the symmetry energy by comparing the isoscaling parameters from @xmath14ar , @xmath14ca + @xmath15fe , @xmath15ni and @xmath15fe , @xmath15ni + @xmath15fe , @xmath15ni reactions with the dynamical amd model calculation @xcite , resulting in symmetry energy , e@xmath11 = 31.6(@xmath5/@xmath0)@xmath12 with @xmath13 = 0.69 .
@xcite , studied the symmetry energy by comparing the experimental double neutron to proton ratio in @xmath10sn + @xmath10sn and @xmath9sn + @xmath9sn reactions with the buu97 calculation @xcite and obtained a symmetry energy , e@xmath11 = 32(@xmath5/@xmath0)@xmath12 with @xmath13 = 0.5 .
@xcite , studied the symmetry energy by comparing the isospin diffusion and pre - equilibrium emission data in @xmath15ni + @xmath15ni , @xmath16au reactions at 52 - 74 mev / a with the bnv calculation , and obtained a symmetry energy that increase linearly with density .
more recently , tsang _ et al . _
@xcite , compared the isospin diffusion and the neutron to proton double ratio for @xmath9sn + @xmath10sn reaction at 50 mev / a with the imqmd calculation @xcite and obtained a symmetry energy of the form , e@xmath11 = 12.5(@xmath5/@xmath0)@xmath17 + 17.6(@xmath5/@xmath0)@xmath12 with @xmath13 = 0.4 - 1.05 .
the functional dependence of the symmetry energy obtained from all these comparisons are very similar to each other .
it must be mentioned that while both the ibuu04 and imqmd comparisons lead to similar form of the density dependence of the symmetry energy for the isospin diffusion observable , the ibuu04 fails to adaquately explain the neutron to proton ratio @xcite .
furthermore , these comparisons lead to very different results for densities above saturation density @xcite ( as discussed in section 5 ) .
the determination of the symmetry energy from the dynamical model comparison is therefore highly model dependent , and the assumption of a single form of the symmetry energy at different densities questionable .
the symmetry energy has also been studied by comparing the experimental observables with the statistical multifragmentation model ( smm ) calculation @xcite in this density region . in multifragmentation a nucleus expands with increasing excitation energy and its equilibrium density is reduced from the ground state density .
such a decrease should essentially map the density dependence of the symmetry energy .
a detailed understanding of the relationship between excitation energy , density , and temperature is however required . an attempt has recently been made by shetty _
@xcite from the study of isoscaling parameters in @xmath15fe , @xmath15ni + @xmath15fe , @xmath15ni reactions at 30 , 40 and 47 mev / a . in this work ,
the symmetry energy was studied by correlating the experimentally observed decrease in the isoscaling parameter , density , and the flattening of temperature ( caloric curve ) , with the increase in excitation energy .
the decrease in the symmetry energy with increasing excitation energy was also observed in the fragmentation of excited target residues following @xmath18c + @xmath9sn , @xmath10sn reactions at 300 and 600 mev / a by le fevre _ et al .
it was also observed in @xmath14ar , @xmath14ca + @xmath15fe , @xmath15ni reactions at 25 , 33 , 45 and 53 mev / nucleon by iglio _
@xcite , and by shetty _
although in these work , a detailed understanding of the relation between temperature , excitation energy and nuclear density was not undertaken .
recently , ogul _
@xcite , compared the msu experimental isoscaling data ( @xmath19 = 0.36 and @xmath20 = -0.39 ) from @xmath10sn , @xmath9sn + @xmath10sn reaction with the smm calculation , and observed that a significant reduction of the symmetry energy is necessary to reproduce the experimental data .
@xcite , studied @xmath9sn + @xmath21ni and @xmath10sn + @xmath15ni reactions at 35 mev / a and observed a similar decrease in the symmetry energy with increasing excitation energy .
the reduction in the symmetry energy has also been observed in the projectile fragmentation of @xmath21ni + @xmath21ni , @xmath22kr + @xmath21ni , @xmath9sn , @xmath23pb reactions at 25 mev / a using the observable @xmath7z / a@xmath8 by souliotis _
more recently , hudan _ et al . _
@xcite , studied the projectile fragmentation in @xmath9xe + @xmath10sn reaction at e / a = 30 mev and measured the @xmath7n@xmath8/z and isotope distribution of fragments with z @xmath6 6 .
they observed that a reduced value of the symmetry energy from 25 mev to 14 mev is essential for explaining the data .
they further showed that fragments with z @xmath6 6 are even more sensitive to the variation in symmetry energy and can be an important probe for studying symmetry energy .
a decrease in the symmetry energy was also implied by wuenschel _
@xcite , in the study of @xmath24kr + @xmath25ni reactions at 35 mev / a .
it has been argued @xcite that the decrease in the symmetry energy to reproduce the experimental observables in these studies may not be necessary if the surface corrections to the symmetry energy are included in the mass parameterization used in the smm model calculation .
the possibility of modification of the symmetry energy and the surface energy coefficient of nuclear matter was studied recently by ogul _
they found that the isoscaling parameters are affected very little by the surface energy variation and are very sensitive to the symmetry energy .
a decrease in the symmetry energy is therefore essential to explain the experimental data .
w. ye _ et al . _
@xcite , also studied the influence of the surface entropy on the isoscaling using the extended compound nucleus ( ecn ) model and found that although the surface entropy increases the numerical values of isoscaling parameters it has only minor effect on the extracted symmetry energy .
a clear theoretical interpretation of the symmetry energy in statistical model is therefore important .
the nuclear symmetry energy has also been studied from observables other than those studied using heavy ion reactions .
@xcite , have studied the symmetry energy using the experimental neutron skin measured in 26 antiprotonic atoms .
@xcite , have studied the symmetry energy from correlation between the symmetry pressure and the symmetry energy using the properties of pygmy dipole resonance ( pdr ) in @xmath23pb nuclei .
@xcite , have studied the symmetry energy from the correlation between the symmetry energy and the experimental centroid energy of the giant dipole resonance ( gdr ) in @xmath23pb nuclei .
@xcite , have studied the symmetry energy from the folding model analysis of the charge exchange @xmath26(@xmath27he , @xmath27li)@xmath28 reaction measured at 41.6 mev / a .
danielewicz @xcite have studied the symmetry energy by constraining the binding energy , neutron skin and the isospin analogue state of finite nuclei .
the symmetry energy has also been studied by fitting the binding energy of 1654 nuclei using the thomas - fermi model of myers and swiatecki @xcite .
the results of these independent studies are shown in fig . 2 and 3 and discussed in section 4 .
symmetry energy at very low densities ( 0.01 - 0.05 ) @xmath3 has been studied by kowalski _
et al._@xcite , in heavy ion collision of @xmath21zn + @xmath30mo , @xmath16au reactions at 35 mev / a . it was shown that isoscaling analysis of the light clusters ( a @xmath4 4 ) can be used to study the symmetry energy at such low densities .
they observed that the experimental symmetry energy is somewhat higher than those expected from the mean field calculations . at such low densities cluster formation
becomes important and mean field calculations do not take into account such effect .
recently , natowitz _ et al . _
@xcite , have shown that the symmetry energy at such low densities can be explained by quantum statistical calculation that includes cluster correlation in nuclear medium . at densities
higher than 0.2 - 0.3 @xmath3 the many body correlation disappears and the symmetry energy follows the dependence predicted by the mean field calculations .
the symmetry energy obtained from the studies discussed above are as shown in fig .
2 . the orange box on the extreme left of fig . 2
correspond to the values of symmetry energy extracted by kowalski _
the green and the red solid points are the symmetry energies extracted from the correlation between temperature , density , isoscaling parameter and the excitation energy by shetty _
the blue solid point in the figure correspond to the symmetry energy obtained by constraining the experimental energy of giant dipole resonance ( gdr ) in @xmath23pb by trippa _
the square solid point in the figure correspond to the symmetry energy obtained by fitting the experimental differential cross - section data in a charge exchange reaction using the isospin dependent interaction of the optical potential by khoa _
the green solid curve in the figure is the one obtained from the dynamical amd model comparison of the experimental isoscaling data assuming the sequential decay effect to be small @xcite .
the symmetry energy extracted from the ibuu04 comparison of isosopin diffusion results by li _
@xcite , and the imqmd comparison by tsang _
@xcite , are similar to the green curve . the solid black curve is the symmetry energy extracted from the double neutron to proton ratio by famiano _
@xcite , using the buu97 model calculation .
the red dashed curve corresponds to the one obtained from an accurately calibrated relativistic mean field calculation by todd@xmath31rutel and piekarewicz @xcite for describing the giant monopole resonance ( gmr ) in @xmath32zr and @xmath23pb , and the isovector giant dipole resonance ( ivgdr ) in @xmath23pb .
the shaded region in the figure corresponds to those obtained by constraining the binding energy , neutron skin thickness and isospin analogue state in finite nuclei using the mass formula by danielewicz @xcite .
the yellow curve correspond to the parametrization adopted in the studies of neutron star @xcite .
current studies in the region , 0.3 @xmath3 @xmath4 @xmath5 @xmath4 @xmath3 , can therefore be parameterized by a stiff " form of the symmetry energy , e@xmath11 = 31.6@xmath33 , with @xmath13 = 0.5 - 0.7 .
this parameterization is in good agreement with the mean field calculation . in the density region , @xmath5 @xmath29 @xmath2 @xmath3 , the experimental symmetry energy is observed to deviate from the mean field calculation due to the importance of cluster formation , and is in good agreement with the quantum statistic calculation @xcite .
the slope parameter , l = 3@xmath3 ( @xmath34 ) , is an alternate way of constraining the symmetry energy near saturation density , @xmath3 .
3 shows the slope parameter l , obtained from various studies discussed above .
these comparison results in a constraint of 30 @xmath29 l @xmath29 80 mev for the symmetry energy slope at saturation density @xcite .
at densities higher than the normal nuclear matter density the theoretically determined symmetry energy as a function of density is largely unconstrained
. experimental determination of the symmetry energy at such densities have been very few and they are all from dynamical model comparison .
the results are highly model dependent and contradicting .
one observable that has recently been studied is the @xmath35@xmath36@xmath37 ratio in @xmath14ca + @xmath14ca , @xmath38ru + @xmath38ru , @xmath38zr + @xmath38zr and @xmath16au + @xmath16au reactions at gsi by the fopi collaboration @xcite .
a comparison @xcite of this data with the transport model ibuu04 calculation shows that it can be reproduced with a soft " ( x = 1 ) form of the density dependence of the symmetry energy .
this is in contrast to those obtained from the low density studies of the symmetry energy where a stiff " ( x = 0 ) form of the symmetry energy reproduced the data well .
more recently , the @xmath35@xmath36@xmath37 gsi data was also compared with the imqmd calculation by feng _ et al . _ @xcite .
the comparison favored a much stiffer form of the symmetry energy than those obtained from the low density studies , and in direct contradiction to those obtained from the ibuu04 model comparison .
the symmetry energy results at such high densities is therefore currently controversial .
another observable that has been suggested for probing the high density behavior of the symmetry energy is the relative and differential collective flow between triton and @xmath39he particles @xcite .
preliminary calculation by yong _
@xcite , suggests that a stiff " ( x = 0 ) dependence of the symmetry energy would lead to a linear behavior for the relative flow , whereas thesoft " ( x = 1 ) dependence would give a non linear behavior with appreciable difference between the two .
so it is expected that a measurement of this observable should provide important information on the high density behavior of the symmetry energy .
the theoretical determination of the symmetry energy at high density is a challenge .
it is not known how reliable the dynamical model calculations can be at such high densities as it is not known how the quantum many body effects can be treated in these calculations , or how the effects of spin@xmath31isospin for three body force can be included . also , the nuclear matter at such densities is not in a pure nucleonic state at thermodynamical equilibrium .
the above comparison with the experimental data is therefore only a `` circumstantial '' evidence @xcite and much theoretical work needs to be done before a definitive conclusion can be reached .
the symmetry energy of finite nuclei at saturation density is often extracted by fitting ground state masses with various versions of the liquid drop mass formula . for a finite nuclei
it is important to decompose the symmetry term of liquid into bulk ( volume ) and surface contributions , and identify the volume symmetry energy coefficient as the symmetry energy derived from the nuclear matter at saturation density . using the constraint obtained from the above studies on nuclear matter symmetry energy , the symmetry energy of a finite nucleus of mass @xmath40 , @xmath41 ,
can be written as @xcite , @xmath42}\ ] ] where , @xmath19 = 31 - 33 mev , @xmath13 = 0.5 - 0.7 and @xmath43 = 2.6 - 3.0 .
the quantities @xmath19 and @xmath20 are the volume and the surface symmetry energy at normal nuclear density . presently , the values of @xmath19 , @xmath13 and @xmath43 remain unconstrained .
the ratio of the volume symmetry energy to the surface symmetry energy ( @xmath43 ) , is closely related to the neutron skin thickness @xcite . depending upon how the nuclear surface and the coulomb contribution is treated , two different correlations between the volume and the surface symmetry energy have been predicted @xcite from fits to nuclear masses .
experimental masses and neutron skin thickness measurements for nuclei with @xmath44 @xmath6 1 should provide tighter constraint on the above parameters . the above relation for the symmetry energy of finite nuclei has been compared @xcite with the thomas - fermi calculation of samaddar _ et al .
it was observed that this empirical relation compares very well with the thomas - fermi calculation for wide range of nuclei .
future measurements of symmetry energy as a function of excitation energy for very light and heavy nuclei should provide further insight into the validity of this relation .
currently , studies on the density dependence of the symmetry energy are being carried out using beams of stable nuclei . in order to study large isospin dependency beams of neutron - rich nuclei
are needed . a number of observables sensitive to the symmetry energy that can be experimentally tested has been sugggested by di toro _ et al _ @xcite .
it is hoped that future facilities for rare isotope beams such as frib , fair and spiral ii should provide increased precision in the measurement of symmetry energy using these observables over a wide range of density .
these facilities will allow studies to be carried out at 2 - 3 times the normal nuclear density .
new detectors such as the time projection chamber ( tpc ) at nscl@xmath36msu and the time projection chamber , samurai at riken@xmath36japan are therefore being planned to study the symmetry energy and the nuclear equation of state .
these detectors will significantly enhance our understanding of the nuclear symmetry energy and their relation to atomic nuclei and neutron stars .
in summary , the experimental determination of the nuclear symmetry energy is reviewed .
it is observed that despite the model dependent ways in which the symmetry energy is extracted significant progress has been made .
these studies are the first step in an effort to constrain the symmetry energy that is important for studying the structure of neutrons stars and exotic nuclei . in the future
, it is hoped that more measurements will be carried out where the symmetry energy is extracted for each densities and compared with theoretical calculations .
these measurements using beams of neutron rich nuclei and robust theoretical interpretation of the symmetry energy will help refine some of these results .
furthermore , they will help constrain the symmetry energy at densities above normal nuclear density where experimental results are scarce .
chen , and c.m .
ko , phys .
464 , 113 ( 2008 ) .
v. baran , m. colonna , v. greco , and m. di toro , phys .
410 , 335 ( 2005 ) .
d.v . shetty , s.j .
yennello , a.s .
botvina , g.a .
souliotis , m. jandel , e. bell , a. keksis , s. soisson , b. stein and j. iglio , phys .
c 70 , 011601 ( 2004 ) .
l.w . chen , c.m .
ko and b.a .
li , phys .
94 , 032701 ( 2005 ) .
li and l.w .
chen , phys .
c 72 , 064611 ( 2005 ) .
shetty , s.j .
yennello and g.a . souliotis , phys . rev .
c 75 , 034602 ( 2007 ) .
shetty , s.j .
yennello and g.a . souliotis , phys .
c 76 , 024606 ( 2007 ) .
et al . , _ phys .
97 , 052701 ( 2006 ) .
e. galichet , m. colonna , b. borderie and m.f .
rivet , phys .
c 79 , 064615 ( 2009 ) .
et al . , _ phys . rev .
102 , 122701 ( 2009 ) .
das , s. dasgupta and c. gale , phys .
c 69 , 011603 ( 2004 ) .
y. zhang , p. danielewicz , m. famiano , z. li , w.g .
lynch and m.b .
tsang , phys .
b 664 , 145 ( 2008 ) .
a. ono , p. daneielewicz , w.a .
friedman , w.g .
lynch and m.b .
tsang , phys .
c 68 , 051601 ( 2003 ) .
shetty , g.a .
souliotis , s. galanopoulos , z. kohley , s.n .
soisson , b.c .
stein , s. wuenschel and s.j .
yennello , j. phys .
g 36 , 075103 ( 2009 ) .
s. kowalski _
et al . , _ phys .
c 75 , 014601 ( 2007 ) .
botvina , o.v .
lozhkin and w. trautmann , phys .
c 65 , 044610 ( 2002 ) .
et al . , _ phys .
86 , 5023 ( 2001 ) .
ko and z. ren , phys .
78 , 1644 ( 1997 ) .
et al . , _ phys .
b 634 , 378 ( 2006 ) .
z. xiao , b.a .
chen , g.c .
yong and m. zhang , phys . rev .
102 , 062502 ( 2009 ) .
z. feng _ et al .
b 683 , 140 ( 2010 ) .
et al . , _ phys .
94 , 162701 ( 2005 ) .
j. iglio , d.v .
shetty , s.j .
yennello , g.a .
souliotis , m. jandel , a.l .
keksis , s.n .
soisson , b.c .
stein , s. wuenschel and a.s .
botvina , phys .
c 74 , 024605 ( 2006 ) .
shetty , a.s .
botvina , s.j .
yennello , g.a .
souliotis , e. bell and a. keksis , phys .
c 71 , 024602 ( 2005 ) .
r. ogul , u. atav , f. bulut , n. buyukcizmeci , m. erdogan , h. imal , a.s .
botvina and i , mishustin , j. phys .
g 36 , 115106 ( 2009 ) .
e. geraci _
et al . , _ eur .
j. 150 , 21 ( 2007 ) .
souliotis , d.v .
shetty , a. keksis , e. bell , m. jandel , m. veselsky and s.j .
yennello , phys . rev .
c 73 , 024606 ( 2006 ) .
souliotis , a.s .
botvina , d.v .
shetty , a.l .
keksis , m. jandel , m. veselsky and s.j .
yennello , phys . rev .
c 75 , 011601 ( 2007 ) . s. hudan _ et al . , _ phys .
c 80 , 064611 ( 2009 ) .
s. wuenschel _
et al . , _ phys .
c 79 , 061602 ( 2009 ) .
souza , m.b .
tsang , r. donangelo , w.g .
lynch and a.w .
steiner , phys .
c 78 , 014605 ( 2008 ) .
w. ye , j. toke , w.u .
schroder , phys .
b 676 , 193 ( 2009 ) .
m. centelles , x. roca - maza , x. vinas and m. wards , phys .
102 , 122502 ( 2009 ) ; m. warda , x. vinas , x. roca - maza and m. centelles , phys .
c 80 , 024316 ( 2009 ) .
a. klimkiewicz _
et al . , _ phys .
c 76 , 051603 ( 2007 ) .
l. trippa , g. colo and e. vigezzi , phys .
c 77 , 061304 ( 2008 ) .
d.t . khoa and h.s .
than , phys .
c 71 , 044601 ( 2005 ) .
khoa , h.s . than and d.c .
cuong , phys .
c 76 , 014603 ( 2007 ) .
p. danielewicz , nucl .
a727 , 233 ( 2003 ) .
w.d . myers and w.j .
swiatecki , nucl .
a601 , 141 ( 1996 ) .
w.d . myers and w.j .
swiatecki , phys .
c 57 , 3020 ( 1998 ) .
et al . , _ arxiv:1001.1102 ( 2010 ) .
todd - rutel and j. piekarewicz , phys
95 , 122501 ( 2005 ) ; piekarewicz and m. centelles , phys .
c 79 , 054311 ( 2009 ) .
h. heiselberg and h. hjorth - jensen , phys . rep .
328 , 237 ( 2000 ) .
p. roy chowdhury , d.n .
basu , and c. samanta , phys .
c 80 , 011305 ( 2009 ) ; d.n .
basu , p. roy chowdhury , and c. samanta , phys . rev .
c 80 , 057304 ( 2009 ) .
w. reisdorf _
et al . , _ nucl .
phys . a 781 , 459 ( 2007 ) .
yong , b.a .
chen , x.c .
zhang , phys .
c 80 , 044608 ( 2009 ) .
steiner , m. prakash , j.m .
lattimer and p.j .
ellis , phys .
411 , 325 ( 2005 )
samaddar , j.n .
de , x. vinas and m. centelles , phys .
c 78 , 034607 ( 2008 ) .
m. di toro _ et al .
e17 , 1799 ( 2008 ) . | the nuclear symmetry energy is a fundamental quantity important for studying the structure of systems as diverse as the atomic nucleus and the neutron star .
considerable efforts are being made to experimentally extract the symmetry energy and its dependence on nuclear density and temperature . in this article
, we review experimental studies carried out up - to - date and their current status .
symmetry energy , density dependence , nuclear matter 25.70.mn ; 25.70.pq ; 26.50.+x |
the study of single - domain magnetic particles has been an active field of research since the pioneer work of stoner and wohlfarth @xcite , who studied the hysteretic rotation of the magnetisation over the magnetic - anisotropy energy barrier under the influence of applied fields , and nel @xcite who predicted that at non - zero temperature the magnetisation can surmount the energy barrier as a result of thermal agitation .
important progress has been made since brown @xcite derived the fokker
plank equation for the probability distribution of spin orientations , starting from the stochastic landau
lifshitz equation , and calculated the relaxation time for uniaxial particles in a longitudinal field .
recent work on spins with non - axially symmetric potentials revealed @xcite a large dependence of the relaxation time on the damping coefficient @xmath0 in the medium - to - weak damping regime ( @xmath0 measures the relative importance of the precession and the damping in the dynamics ) .
experiments on individual nanoparticles @xcite analysed with accurate asymptotes of the relaxation time @xcite , gave damping coefficients in that regime : @xmath1@xmath2 . turning our attention from independent particles to systems of interacting particles ,
the complexity of the problem increases drastically , as it becomes a many - body problem with a long - ranged and reduced symmetry interaction mechanism the dipolar interaction .
the approach to study the relaxation time @xmath3 has so far been based on how the energy barriers of the spins are modified by the interactions @xcite . with the @xmath3 so obtained
, one can study the effects of the interaction on the superparamagnetic blocking ( the maximum in the dynamical response at the temperature where @xmath3 becomes of the order of the observation time ) .
however , the barrier - based approach to determine @xmath3 corresponds to assuming @xmath4 , so that dynamical features such as the precession of the spins are disregarded , as incisively noted by berkov and gorn @xcite .
indeed , numerical integration of the stochastic landau lifshitz equation for weakly interacting systems revealed @xcite non - trivial effects of finite damping on the superparamagnetic blocking , such as enhanced shifts of the temperature of the maximum of the dynamical response and non - monotonic behaviour of its height with the interaction strength . in this article
we apply thermodynamic perturbation theory ( treating the anisotropy energy exactly and the dipolar energy perturbatively ) to calculate the averages of the dipolar field produced at the position of a given spin .
these averages are then introduced in a rigorous low - field expansion of the single - spin relaxation rate , which can be combined with a recently derived perturbative formula for the equilibrium susceptibility @xcite to provide a model for the dynamical susceptibility of weakly interacting spins . in the low temperature range , where the superparamagnetic blocking takes place
, our model recreates the non - trivial damping dependence of the blocking found in ref .
the analytical treatment permits us to ascribe the observed features to the deviations from uniaxial symmetry of the potential due to the transversal components of the dipolar field , deviations which render the relaxation time very sensitive to the damping .
in weak fields , we can handle the field dependence of the relaxation time @xmath3 by expanding the relaxation rate @xmath5 in powers of the field components ( for convenience we use the field in temperature units @xmath6 , where @xmath7 is the magnetic moment ) .
as the spins have inversion symmetry in the absence of the field , the relaxation rate will not change if the field is reversed ( since @xmath8 accounts for jumps over the energy barrier in both directions ) .
therefore , @xmath8 should be an even function of @xmath9 , and hence all odd order coefficients in the expansion must be identically zero . for spins with uniaxial anisotropy ,
the second order corrections can be written in terms of the longitudinal @xmath10 and transverse @xmath11 components of the field with respect to the anisotropy axis .
using , for instance , the invariance of the relaxation rate upon field reflection through the barrier plane , where @xmath11 does not change but @xmath10 is reversed , we see that no quadratic term of the form @xmath12 can occur .
this leads to the following generic form for the relaxation rate of uniaxial spins in the presence of a weak , arbitrarily directed , field @xmath13 where @xmath14 is the zero - field relaxation rate and the expansion is actually valid to third order .
the coefficients @xmath15 and @xmath16 will be determined by choosing special configurations in which they are known ( strictly longitudinal and transverse fields ) .
we shall restrict our attention to low temperatures where the superparamagnetic blocking takes place .
then , the zero - field relaxation rate is given by brown s result @xcite @xmath17 where @xmath18 ( @xmath19 ) is the relaxation time of isotropic spins , @xmath20 is the anisotropy barrier in temperature units , and corrections of order @xmath21 are disregarded due to the low @xmath22 assumption .
note that in this case ( and also in a longitudinal field ) the damping parameter @xmath0 only enters through @xmath18 and hence it only matters to establish a global time scale . in other words , the results for different @xmath0 presented in units of @xmath18
show complete dynamical scaling , and in this sense the @xmath0 dependence is said to be trivial .
expanding the corresponding expression for @xmath8 in the presence of a longitudinal field @xmath10 @xcite ( disregarding again @xmath21 terms ) , we find @xmath23 comparison with the general expansion ( [ gamma : expansion : uni : gen ] ) gives the longitudinal coefficient @xmath24
. there is no general expression for the relaxation time in the presence of a non - zero transverse field valid for all values of the relevant parameters .
nevertheless , garanin _ et al .
_ @xcite have derived a low temperature formula valid for weak transverse fields , which is perfectly suited for our purpose of determining @xmath16 , namely @xmath25 \ ; , \qquad f(\alpha ) = 1 + 2 ( 2\alpha^{2}e)^{1/(2\alpha^{2 } ) } \gamma \left ( 1+\frac{1}{2\alpha^{2 } } \,,\ , \frac{1}{2\alpha^{2 } } \right ) \;.\ ] ] here @xmath26 and @xmath27 is the incomplete gamma function . comparing with the expansion ( [ gamma : expansion : uni : gen ] ) one gets the transverse coefficient @xmath28 .
on gathering these results we finally get the desired expression for the low temperature relaxation rate in weak fields @xmath29 \;,\ ] ] which constitutes a straightforward generalisation of the formula of garanin _ et al .
_ @xcite to an arbitrary field orientation .
the function @xmath30 takes into account without further approximations the effects of the precession .
@xmath30 decreases towards @xmath31 for strong damping ( where @xmath15 and @xmath16 are of the same order of magnitude ) , while @xmath30 grows as @xmath32 for weak damping , where the relaxation time turns to be very sensitive to the damping ( or to transverse fields ) .
let us compare the rigorous eq .
( [ gamma : expansion : uni ] ) with the corresponding result of mrup and co - workers ( see , for instance , eq . ( 6 ) in ref .
@xcite ) , which for @xmath33 can be written as @xmath34 this expression has the undesirable feature of a term linear in the field , which can be attributed to accounting for the escape from one of the potential wells only . actually , if we average over both wells the linear term cancels out and the corrected formula equals the overdamped @xmath4 limit of eq .
( [ gamma : expansion : uni ] ) ( then @xmath35 ) .
this is natural , since in ref .
@xcite the effects of the field were only considered via barrier changes , so that no gyromagnetic effects were included , and their result can only be correct when the precession can be neglected ( @xmath4 ) .
besides , as their linear term disappeared upon random anisotropy averaging , their results will correspond to the limits overdamped _ plus _ random anisotropy of those derived here from eq.([gamma : expansion : uni ] ) .
let us consider a system of magnetoanisotropic spins coupled via the dipole - dipole interaction . the dipolar field at the position @xmath36 of the spin @xmath37 created by all other spins is given by @xmath38 where the term with @xmath39 is omitted
, @xmath7 is the magnitude of the magnetic moment , and in the dipolar tensor @xmath40 and @xmath41 .
all lengths are measured in units of the characteristic length @xmath42 , which is defined in such a way that @xmath43 is the mean volume occupied around each spin . in a simple cubic arrangement
@xmath42 is the lattice constant and for nanoparticles of volume @xmath44 the volume concentration of particles is @xmath45 . for notational simplicity
we are assuming that the parameters characterising the different spins are identical ( it is immediate to generalise the expressions for different anisotropy constants , magnetic moments , volumes , etc . ) . for spins with uniaxial anisotropy the total energy of the system can be written as @xmath46 where @xmath47 , @xmath48 is the unit vector along the anisotropy axis of the @xmath49th spin
, @xmath50 is the anisotropy barrier divided by the thermal energy , and @xmath51 the quantity @xmath52 is a convenient temperature independent measure of the interaction strength ( @xmath53 concentration ) equal to the magnitude of the field , measured in units of the maximum anisotropy field @xmath54 , produced at a given position by a spin located at a distance @xmath42 . using thermodynamic perturbation theory ( ref.@xcite , 32 ) to expand the boltzmann distribution @xmath55 in powers of @xmath56
, one gets an expression of the form @xmath57 \;. \label{wexp}\ ] ] here @xmath58 is the boltzmann distribution of the non interacting ensemble ( in our case @xmath59 includes the magnetic anisotropy ) and @xmath60 is linear in the dipolar energy ( and hence quadratic in the spins ) .
the calculation of the observables reduces to computing averages weighted by the non interacting probability distribution @xmath61 of low grade powers of the spin variables . in this way ,
the results are exact in the magnetic anisotropy and only perturbational in the dipolar interaction . the averages of terms linear and quadratic in @xmath62 , weighted by the noninteracting distribution , can be calculated by means of the following algorithms " @xcite @xmath63 where @xmath64 and @xmath65 are arbitrary constant vectors .
the first average vanishes ( actually any odd power ) since the magnetic anisotropy has inversion symmetry [ @xmath66 and there is no external bias field .
the quantity @xmath67 is the average of the second order legendre polynomial @xmath68\rangle_{\rm a}$ ] , and it can be written in terms of the one - spin partition function @xmath69 , as @xmath70 in the equations for the relaxation rate the field enters squared [ eq .
( [ gamma : expansion : uni ] ) ] .
therefore , we do not calculate the statistical mechanical average of @xmath71 [ eq .
( [ dipolar ] ) ] and plug it into @xmath8 ( which would be a sort of mean - field approach ) but we average instead the combinations of field variables as they enter in the expression for @xmath8 . since @xmath72 , we average , by means of the algorithms ( [ alg ] ) , the square of the field and the square of the projection along the local anisotropy axis @xmath73 , getting ( @xmath74 ) @xmath75 \ ; , \\ \label{h : perp } \big\langle \xi_{i,\perp}^{2 } \big\rangle & = & \frac{{\xi_{\rm d}}^{2}}{3 } \sum_{j } \big [ 6r_{ij}^{-6 } + 3s_{2 } \ , r_{ij}^{-3 } ( { \vec{n}_{j}}\cdot{{\bf g}_{ij}}\cdot{\vec{n}_{j } } ) \nonumber\\ & & \qquad \quad { } -\left(1-s_{2}\right ) \ , ( { \vec{n}_{i}}\cdot{{\bf g}_{ij}}\cdot{{\bf g}_{ij}}\cdot{\vec{n}_{i } } ) -3s_{2 } \ , ( { \vec{n}_{i}}\cdot{{\bf g}_{ij}}\cdot{\vec{n}_{j}})^{2}\big ] \;.\end{aligned}\ ] ] the averages have been calculated to order @xmath76 ( so we just need eq .
( [ boltzmann ] ) to zero order ; @xmath77 only enters in the third order corrections ) .
note that the averaged fields depend on @xmath22 via @xmath78 , reflecting the fact that the field created at a given position can be different if , for example , the source spins are almost freely rotating ( high @xmath22 , @xmath79 ) or are almost parallel to their anisotropy axes ( low @xmath22 , @xmath80 )
. one may wonder about the validity of these field averages below the superparamagnetic blocking , where the spins are not in complete equilibrium .
however , since at those temperatures the spins are still in quasi - equilibrium confined to one of the two wells , we can repeat the derivation of the algorithms ( [ alg ] ) restricting the phase space for integration to one well . in this case , averages of the form @xmath81 do not vanish , and should be considered together with @xmath82 , which being even in @xmath62 is not modified .
the extra terms associated with @xmath83 , however , vanish if the overall state is demagnetised , and we recover eqs .
( [ h : para])([h : perp ] )
. this can be interpreted as if two nearby blocked spins , one in its upper well and the other in the lower , create a net field at the position of a third spin similar to that created by any of them when in equilibrium , since each of the spins approximately compensates for the restricted phase space of the other .
the general expressions for the longitudinal and transversal fields are notably simplified in some important situations . for a system with parallel anisotropy axes ( e.g. , in a single crystal of magnetic molecular clusters , or ferrofluids frozen in a strong magnetic field ) we equate all the @xmath84 to @xmath85 . for a system with randomly distributed anisotropy axes
we replace expressions involving @xmath86 by integrals @xmath87 , and use @xmath88 . in both cases
the final expressions involve some rapidly convergent sums over the lattice .
let us concentrate in cases of sufficiently isotropic " lattices , in the sense of fulfilling @xmath89 , e.g. , cubic and completely random lattices ( incidentally , the type of arrangements for which in the classical lorentz cavity field calculation the contribution of the dipoles inside the small " sphere vanishes ) .
similarly we consider large enough systems , so that all spins have approximately equivalent surroundings ( then the index @xmath49 on the different quantities can be dropped ) . under these circumstances the lattice sums " involved are @xmath90 where the terms with @xmath39 are of course omitted . for a system with aligned anisotropy axes the averaged fields are given in terms of the lattice sums ( [ r&t ] ) by the compact expressions @xmath91 , \qquad \big\langle \xi_{\perp}^{2 } \big\rangle = \frac{{\xi_{\rm d}}^{2}}{3 } \left [ ( 2+s_{2 } ) { \cal r } - 3s_{2}\,{\cal t } \right ] \;,\ ] ] while for randomly distributed anisotropy axes they read @xmath92 in both cases @xmath93 is simply given by the temperature independent result @xmath94 .
this independence also holds for the field components after the random anisotropy average [ eq . ( [ h : para : perp : random ] ) ] , but not for parallel axes [ eq .
( [ h : para : perp : aligned ] ) ] where the field components remain @xmath22 dependent .
the dependence of the features of the superparamagnetic blocking on the interaction strength @xmath95 has been a subject of some controversy . two main approaches @xcite addressed this problem on the basis of the modifications of the energy barriers by the interactions .
however , this type of approach overlooks the fact that not only the energy landscape is important , but also how the spin evolves in it , depending on if the spin is precessing almost freely ( weak damping ) or strongly damped .
indeed , berkov and gorn @xcite have shown with rigorous langevin dynamics simulations that for weak interactions the position of the blocking temperature @xmath96 ( where @xmath97 reaches its maximum ) of strongly damped spins is hardly affected by the interaction strength , whereas for weak damping @xmath96 significantly decreases with @xmath52 and the peak height behaves non - monotonically .
the apparent discrepancy with some experimental results , in which @xmath96 increases with @xmath52 , was attributed in ref.@xcite to the different behaviour of systems with weak anisotropy ( or moderate - to - high interactions ) , where the energy barriers are mostly due to the interactions and hence grow with @xmath52 . in order to assess the features of the superparamagnetic blocking emerging from our model , we make use of a rigorous perturbative expansion of the equilibrium susceptibility @xmath98 in powers of @xmath56 ( to second order , see ref .
this expression , together with the relaxation rate @xmath8 obtained when the averaged fields ( [ h : para : perp : aligned])([h : para : perp : random ] ) are introduced in eq .
( [ gamma : expansion : uni ] ) , can be combined in a simple debye - type formula @xmath99 naturally , at low enough temperatures @xcite the results will become invalid by the very nature of the approximations involved , although this does not affect the characteristics of the superparamagnetic blocking for weak interactions . the dynamical response for a large spherical sample with parallel anisotropy axes and simple cubic lattice structure is shown in fig .
[ fig : fig1 ] ( the lattice sums required are @xmath100 and @xmath101 ) . in the overdamped case , @xmath96 is not noticeably affected by the dipolar interaction and the height of the susceptibility peak decreases monotonically with @xmath52 .
this corresponds to the effect of a slight decrease of @xmath96 found in ref .
@xcite for random anisotropy , which is observable only at very high frequencies ( e.g. , mssbauer ) . for weak
damping , however , fig .
[ fig : fig1 ] shows that the blocking temperature significantly decreases as the interaction strength increases and , in addition , the peak height of @xmath97 initially rises for small values of @xmath52 and then decreases for larger values .
the same behaviour can be seen in @xmath102 with the appropriate choice of parameters .
the features shown in fig . [ fig : fig1 ] are in complete agreement with the results obtained by berkov and gorn @xcite .
the analytical treatment employed here , however , allows us to readily trace back the origin of the results obtained and to interpret them in terms of the different @xmath52 dependences of @xmath8 and @xmath103 . for overdamped systems the coefficients @xmath24 and @xmath104 in the expression ( [ gamma : expansion : uni ] ) for the relaxation rate are of order unity and lead to a slight increase of @xmath8 ( decrease of @xmath105 ) with @xmath52 , while the entire curve is lowered by the reduction of @xmath98 with @xmath52 . for weak damping , however , @xmath106 is large and makes @xmath8 very sensitive to the interaction strength , moving the peak quickly towards low @xmath22 as soon as @xmath52 departs from zero .
then , as the coefficients in @xmath98 are not as large as @xmath16 , the initial decrease of the equilibrium susceptibility is smaller than the increase associated to the quick shift of the blocking to low @xmath22 , where @xmath98 is higher ( roughly @xmath107 ) , producing the rise of the peak .
if @xmath52 is further increased the decrease of @xmath98 starts to be competitive and the peaks are reduced as they shift to lower temperatures .
the different behaviours are related to the presence of transverse components of the local fields , which create a saddle point in the uniaxial potential barrier of the spins , turning the relaxation rate sensitive to @xmath0 @xcite .
we can picture the underlying physical mechanism as follows @xcite .
consider one spin that after a favourable " sequence of fluctuations , reaches a point close to the top of the barrier but does not surmount it .
in the subsequent spiralling down back to the bottom of the potential well , a strongly damped spin descends almost immediately , whereas a weakly damped spin executes several rotations ( @xmath108 ) about the anisotropy axis .
this allows the latter spin to pass close to the saddle area , where it will have additional opportunities , not available for the damped spin , to cross the barrier . as we see
, this mechanism duly combines the characteristics of the potential _ and _ the dynamical evolution of the spins in the potential .
note finally that the transverse components are non - zero even for parallel anisotropy axes
( [ h : para : perp : aligned ] ) ] , so there is no need , as in the case of non - interacting particles , to appeal to oblique fields , applied @xcite or probing @xcite , to find a large sensitivity to the damping in interacting systems .
we have proposed a model for the relaxation time of weakly interacting superparamagnets .
the single - spin relaxation time at low fields is generalised for an arbitrary directed field , and the components of the local field calculated by thermodynamical perturbation theory .
the non - monotonic behaviour of the height of the dynamical susceptibility peak with the interaction strength and the increased lowering of the blocking temperature with decreasing damping , discovered numerically by berkov and gorn , are captured by our model .
these features are interpreted in terms of the different sensitivity , depending on the damping strength , of the relaxation rate to transverse fields , which for interacting spins are provided by the dipolar interaction . | the relaxation time of weakly interacting classical spins is calculated by introducing the averages of the local dipolar field , obtained by thermodynamic perturbation theory , in a rigorous expression for the single - spin thermoactivation rate in a weak but arbitrarily oriented field . at low temperatures the non - trivial dependence of the superparamagnetic blocking on the damping coefficient , numerically found by berkov and gorn ,
is reproduced by our model and interpreted in terms of the deviations from uniaxial anisotropy associated to the transversal component of the dipolar field acting on each spin . |
analyses of high - throughput genomic data often produce ranked lists of genomic loci .
two examples are lists of differentially expressed genes from an rna - seq or microarray experiments @xcite and lists of transcription factor ( tf ) binding peaks from chip - seq data @xcite . in each list , loci are ranked based on scores such as p - values , false discovery rates ( fdr ) @xcite or other summary statistics .
when two such lists are available , a common problem is to characterize the degree of concordance between them .
below are two examples . *
_ characterizing co - binding of two transcription factors _ :
chip - seq data are collected for two different tfs . for each tf ,
an initial data analysis yields a list of peaks along the genome representing its putative binding regions . in order to characterize whether the two tfs collaborate and how they interact with each other ,
one wants to compare the two peak lists to answer the following questions : ( 1 ) what proportion of the true binding sites are shared by the two tfs ? ( 2 ) how does this proportion change as one moves from high quality peaks to low quality ones ? * _ assessing reproducibility of scientific findings _ : gene expression data for the same biological system are collected independently by two different laboratories .
each lab collects the data using its own platform and protocol .
the data from each lab contain gene expression profiles for two biological conditions , each with multiple replicate samples .
each lab analyzes its own data to generate a list of differentially expressed genes .
one wants to compare the differential gene lists from the two labs to determine which differential genes are likely to be reproducible by other labs . in both these scenarios ,
perhaps the best way to compare two datasets is to model it at the raw data level .
whenever possible , directly comparing or modeling the raw data may allow one to keep most of the information . however , this is not always easy or feasible .
for instance , sometimes genomic rank lists are published without releasing the raw data to protect confidentiality of research subjects .
sometimes , one may want to compare his / her own data with thousands of other datasets in public repositories such as encode @xcite , modencode @xcite , and gene expression omnibus ( geo ) @xcite . analyzing all the raw data in these databases
is a huge undertaking that requires significant amount of resources .
this is often beyond the capacity of an individual investigator , and it may not be justified based on the return . in those situations , comparing two datasets based on the readily available rank lists may be preferred . sometimes , this may be the only solution .
this article considers analysis issues in this scenario . as an exploratory tool ,
the simple venn diagram approach is widely used to show the overlap between two genomic loci lists . however
, this approach does not consider the concordance or correlation of ranks between the two lists .
a feature commonly seen in genomic rank lists is that the top ranked loci are more likely to be true signals .
signals are more likely to be reproduced in independent studies than noise ; therefore , they tend to be correlated between different datasets .
because of this , the concordance between the two rank list is a function that changes with the rank of the loci .
this information is not reflected in a venn diagram . to address this limitation , li et al .
recently proposed a method to measure the concordance of two rank lists as a function of rank .
they developed a gaussian copula mixture model to assign a reproducibility index , irreproducible discovery rate ( idr ) , to each locus .
the idr analysis produces a concordance curve rather than a scalar number to measure the overlap between two lists .
this approach is semiparametric and invariant to monotone transformations of the scores used for ranking @xcite . in principle , idr is a model based version for one minus the correspondence at the top ( cat ) plot proposed by irizarry et al.@xcite .
the original authors of idr demonstrated their method using an application where they evaluated the reproducibility of different chip - seq peak callers by comparing the peak calling results from two replicate experiments . although the idr approach represents a significant advance compared to the simple venn diagram analysis , it also has limitations .
importantly , the gaussian copula mixture model in the original idr approach requires one to know the ranks of each locus in both lists . however , many loci occur only in one list . as a result , to perform the idr analysis , li et al .
first filtered out all loci that were reported in only one rank list .
loci are included in the idr analysis only if they appear in both lists .
as li et al . reported , for the real data they analyzed ( which are peak lists from replicate chip - seq experiments ) , only `` 23 - 78% of peaks are retained for this analysis '' @xcite . as such , the original idr analysis only characterizes the signal concordance for a subset of loci that are reported in both lists .
attempting to interpret the resulting idr as a reproducibility measure for the whole dataset could be misleading .
it is possible that the two original loci lists have little overlap and , therefore , low reproducibility , but the loci in their overlapping set ( i.e. , the loci shared by both lists ) are highly correlated in terms of their relative ranking .
in such a situation , the idr computed using only the overlapping loci may misleadingly suggest high reproducibility of the two datasets .
this is a limitation caused by ignoring list - specific loci , and can only be addressed by bringing them back into the analysis . here
we propose a survival copula mixture model , scop , to tackle the general problem of comparing two genomic rank lists .
this new approach allows one to include the list - specific loci in the analysis when evaluating the signal concordance between the two datasets . for loci that occur only in one list , we treat the scores used for ranking ( e.g. , p - values or fdrs ) in the other list as censored data . in this way
, we translate the problem into a bivariate survival problem .
although many works have been done in the area of estimating correlation structure of bivariate failure times in survival analysis @xcite , none of them considered the issue specific to genomic data . in genomic applications , the higher ranked loci are more likely to be true signals .
thus , adopting the traditional survival analysis terminology , earlier failure time is of higher interest . built upon li et al , our survival copula mixture model attempts to borrow strength from both the copula mixture model and the survival analysis .
the benefit is that it can better characterize the overlap and concordance between two rank lists .
the article is organized as follows . in section 2
, we introduce the survival copula mixture model and discuss its connection to survival analysis .
section 3 uses simulations to demonstrate our method and compare it with the other alternatives .
we apply our method to two real chip - seq datasets example in section 4 .
we then conclude the article with discussions in section 5 .
consider two genomic loci lists such as lists of differentially expressed genes from two rna - seq experiments or lists of transcription factor binding regions from two chip - seq experiments . in each list @xmath0 ( @xmath1 ) , loci are rank ordered based on a score such as a p - value or an fdr .
let @xmath2 be the score for locus @xmath3 in list @xmath0 .
without loss of generality , we assume that smaller score ( e.g. , smaller fdr ) represents a higher significance .
often , a locus is reported in list @xmath0 only when its score passes a cutoff @xmath4 .
thus , all loci in list 1 satisfy @xmath5 , and any locus with @xmath6 is not reported .
similarly , list 2 contains loci for which @xmath7 .
a locus may be reported in both lists , in one list only , or in none of the lists .
each list may contain a certain amount of noise or false positives in addition to signals . by comparing the two lists ,
the goal is to characterize the degree of concordance of the signals from the two datasets , and how the concordance varies as one moves from the top ranked loci to those lower ranked . to analyze these data
, we borrow the idea of idr .
however , instead of excluding loci that occur in only one list from the analysis , we retain all loci that occur in any of the two lists . if a locus does not appear in one list , its score in that list is labeled as missing .
this creates missing data , but the data here are not missing completely at random .
for example , if rank list 1 uses an fdr cutoff of 0.1 , then we know that for any loci in list 2 but not in list 1 , their missing fdr in list 1 are indeed greater than 0.1 . in other words , the data we observe are right truncated .
this naturally translates the problem into a survival problem with right censoring data .
figure 1(a ) shows a numerical example .
the figure displays two chip - seq peak lists ranked according to fdr .
region 2 passes the fdr cutoff for both lists , but region 1 only appears in the peak list for tf a. it is absent in the peak list for tf b since its fdr in that dataset is higher than the 0.1 cutoff . rather than excluding region 1 from the analysis , we retain it and encode the data using `` observed survival time '' and `` censoring indicator '' adopting the terminology in survival analysis .
the `` observed survival time '' is defined as @xmath8 , and the `` censoring indicator '' is defined as @xmath9 . in this example , the observed survival time for region 1 in peak list b is 0.1 , and the censoring indicator is equal to zero indicating that the data is censored .
intuitively , the original idr approach by @xcite only models the red points ( i.e. , cases with complete data ) in figure 1(b ) , whereas our new approach attempts to use information from all data points regions ii , iii , and iv .
later we will show that compared to the original idr calculation which excludes the list - specific loci , including them as censored data in our model will provide more information .
let @xmath10 be the probability density function for @xmath2 .
@xmath11 is the corresponding survival function . for any bivariate random variables
, there exists a copula , which is invariant under monotone transformation for the marginal distribution @xcite .
based on this , we use two latent random variables @xmath12 and @xmath13 to characterize the relationship between @xmath14 and @xmath15 . for each @xmath0 , @xmath16 is assumed to follow a gaussian mixture distribution .
@xmath17 represents the survival function for the latent variable @xmath16 .
the latent variables ( @xmath12 and @xmath13 ) and the observed scores ( @xmath14 and @xmath15 ) are linked through a monotone transformation @xmath18 .
let @xmath19 denote the density function for @xmath12 .
it is assumed that this density is a mixture of a noise component @xmath20 and a signal component @xmath21 , where @xmath22 .
similarly , the density function for @xmath13 , @xmath23 , is assumed to be a mixture of noise @xmath24 and signal @xmath25 where @xmath26 .
the data are assumed to be generated as below ( see figure 1(c ) for a cartoon illustration ) : 1 . a random indicator @xmath27 is first assigned to each locus @xmath3 . * if @xmath28 , then locus @xmath3 is noise in both lists . * if @xmath29 , then locus @xmath3 is signal in list 1 but noise in list 2 . * if @xmath30 , then locus @xmath3 is signal in list 2 but noise in list 1 . * if @xmath31 , then locus @xmath3 is signal in both lists .
+ thus , @xmath27 represents the co - existing pattern of signals .
it is usually called `` frailty '' in survival analysis .
the @xmath27 is assumed to be assigned according to the probability vector @xmath32 , where @xmath33 .
2 . given @xmath27 , latent variables @xmath34 and
@xmath35 are generated according to @xmath19 and @xmath36 , respectively .
3 . @xmath34 and @xmath35 are truncated using @xmath37 and @xmath38 as cutoffs .
4 . the truncated pseudo data @xmath39 and @xmath40 are monotone transformed to observed data @xmath41 and @xmath42 based on @xmath43 , which yields @xmath44 .
correspondingly , @xmath45 . also note that @xmath46 where and @xmath47 .
since @xmath2 is truncated at @xmath48 and the censoring time @xmath48 is a constant , the censoring time is independent of the underlying true failure time @xmath2 and contains no information about @xmath10 and @xmath49 . as a result , the contribution of each locus @xmath3 to the likelihood can be represented by one of the four formulas below : * @xmath28 : @xmath50 .
* @xmath29 : @xmath51 .
* @xmath30 : @xmath52 .
* @xmath31 : @xmath53 .
collect the data and latent variables into three sets @xmath54 , @xmath55 and @xmath56 , and define @xmath57 .
the full likelihood can be derived as : @xmath58 to fit the model , we use an iterative em algorithm similar to the one proposed by li et al.@xcite to estimate @xmath59 . 1 .
intialize @xmath59 using random values @xmath60 .
2 . use the kaplan - meier @xcite estimator to estimate the marginal survival functions for @xmath61 and @xmath62 .
3 . given the initial @xmath60 , obtain pseudo - data @xmath63 , @xmath64 .
estimate parameters @xmath59 based on the pseduo - data @xmath39 and @xmath40 using an em algorithm @xcite . 5 .
update the @xmath65 and @xmath66 using the newly estimated @xmath59 , and update the pseudo - data @xmath39 and @xmath40 using the new @xmath65 and @xmath66 .
iterate between steps 3 and 4 until the change in log - likelihood between the two nearby iterations is less than a pre - specified threshold .
details of the algorithm are given in the appendix .
once the model parameters are estimated , a coexistence probability ( also called probability for having reproducible signals ) can be computed for each locus @xmath3 as : @xmath67 using these coexistence probabilities , we define two coexistence curves ( cop curves ) as : @xmath68 @xmath69 intuitively , @xmath70 indicates that among the loci whose scores in list 1 are less than @xmath41 , the proportion that are true signals in both lists .
similarly , @xmath71 shows among the loci ranked higher than locus @xmath3 in list 2 , the fraction that represents signals reproducible in both lists . from these two cop curves
, one can see how the co - existence strengths between the two lists change from the most significant loci to the least significant ones . to facilitate the comparison with the idr approach in @xcite
, we also define : @xmath72 @xmath73 @xmath74 represents the fraction of noise or non - concordant ( non - reproducible ) signals among loci whose score in list 1 does not exceed @xmath41 .
@xmath75 can be interpreted similarly .
our model and measures here allows asymmetry of the signals in the two lists .
for instance , if one list is obtained from a poor quality experiment with low signal - to - noise ratio and the other list is from a high - quality experiment with high signal - to - noise ratio , the two cop curves will be different .
in contrast , the original idr approach only produces one idr curve to show the concordance . as a result
, it can not show the difference between two asymmetric datasets .
in this section , we use simulations to illustrate scop and compare it with the venn diagram and idr approach .
case i illustrates why scop is better at characterizing the degree of concordance between two rank lists .
consider two lists , each with 10,000 loci .
since the copula model is invariant to monotone transformation of marginal scores , we generated the simulation data by first generating latent random variables @xmath12 and @xmath13 and then transforming them to p - values , denoted as @xmath2 , from a one - sided @xmath76-test for @xmath77 vs @xmath78 . specifically , @xmath79 where @xmath80 follows the standard normal distribution @xmath81 . for both lists ,
a normal distribution @xmath82 was used as the signal component for the latent variable @xmath16 , and @xmath81 was used as the noise component .
the mixture proportions of the four possible co - existence patterns were @xmath83 .
in other words , for the full lists without truncation , only 10% of the loci represent signals in both lists , and the other 90% of the loci are noise .
all @xmath16 values greater than -1.65 , corresponding to p - value@xmath84 in a one - sided z - test , were truncated .
the p - values @xmath2 were then generated according to the process described in section 2.2 . under this setting ,
the two lists are symmetric in terms of their signal - to - noise ratio . to reflect the scenario in real applications , loci
whose p - values were greater than @xmath84 in both lists were excluded from the analysis
. meanwhile , all the other loci , either censored in only one or neither of the two lists , were retained . as shown in figure 2 ( a ) ,
a total of 1,872 loci passed the p - value cutoff in either list 1 or list 2 . among them , 56.1% ( 1,050 ) were reported in both lists .
nevertheless , the venn diagram does not characterize the rank concordance between the two lists . for case i. ( d ) @xmath85 for case i. ( e)-(f ) the venn diagram , idr by li et al(2011 ) , @xmath86 and @xmath85 for case ii.,scaledwidth=80.0% ] we then applied the idr approach to the 1,050 loci reported in both lists , consistent with how the idr analysis was performed by @xcite .
figure 2 ( b ) shows the corresponding idr curve .
based on the curve , the idr analysis would claim high reproducibility between the two datasets .
however , this is clearly not the case , since figure 2 ( a ) shows that 43.9% of the 1,872 reported loci were not shared by the two lists .
this illustrates why ignoring the list - specific loci in the idr analysis can be misleading .
the high reproducibility that the method reports only describes the degree of concordance among the loci common in both lists .
it does not characterize the concordance or the reproducibility of the whole lists .
this has important implications .
idr is widely used in the encode project to measure the reproducibility of replicate experiments , and to evaluate the performance of data analysis algorithms in terms of how consistent they perform when applied to replicate experiments @xcite .
the example here shows that idr can be very misleading if one wants to measure the global reproducibility of two replicate experiments , or to evaluate if a data analysis algorithm is stable .
this is caused by ignoring the list - specific loci , which is not allowed in the original idr model . finally , we applied scop to the simulated data .
figures 2 ( c ) and ( d ) show the corresponding @xmath86 and @xmath85 curves , together with the underlying truth curve .
the @xmath86 and @xmath85 curves ( red dashed lines ) match the underlying truth curves ( black solid lines ) very well . as a benchmark comparison
, we also counted among the top ranked @xmath87 loci in one list how many of them were absent from the other list , and created the corresponding curves called `` naivevenn '' ( dark green dotted lines ) hereafter . from another perspective , naivevenn curves were constructed by fixing one circle in the venn diagram , varying the other circle with different rank cutoffs , and counting the overlap proportions .
to certain extent , `` naivevenn '' can be viewed as a naive estimation for @xmath86 and @xmath85 .
however , naivevenn underestimates the irreproducibility for loci occurring in only one list , whereas scop is able to borrowing information from all loci , complete or missing in one list , to better estimate the signal and noise proportions in the data .
@xmath86 and @xmath85 curves clearly demonstrate that the fraction of concordant signal in the two lists is not high , and the irreproducible loci consist of 40% of both observed lists .
@xmath86 and @xmath85 curves also show that the signal concordance decreases as one moves from top ranked loci to lower ranked loci , a trend not directly revealed by the venn diagram approach . with these curves , one may adjust the cutoff for calling signals based on the degree of reproducibility between two independent experiments , which is a function not usually provided by the venn diagram approach . in case
ii , each rank list contains 1,000 loci .
the mixture proportions of the co - existence patterns were @xmath88 .
for both lists , the noise and signal components for generating latent random variables @xmath16 were again assumed to follow @xmath81 and @xmath89 respectively .
@xmath16s were truncated at -1.65 as well . among the 1,083 loci that passed the cutoff in either list , 98.8% ( 1,070 ) were not found in both lists . figures 2(k ) and ( l ) show that scop accurately characterizes the degree of signal concordance between the two lists ( compare the red and the black curves ) . comparing figure 2(b ) in case i and figure 2(f ) in case ii , one can see that the original idr approach would claim high reproducibility in both cases .
however , figure 2(g)(h ) clearly demonstrates that these two cases are different . in case
i , 40% of all loci are claimed as noise ( figure 2 ( c)(d ) ) for the observed lists , whereas only 1.5% of all loci in case ii are estimated as noise ( figure 2 ( g)(h ) ) . in summary ,
the two simulations above show that the overlap revealed by venn diagrams does not contain all information about the degree of signal concordance and how it changes when rank changes .
they also show that the idr computed using the loci present in both lists can be misleading for characterizing global concordance or reproducibility . by incorporating the censoring data into the analysis
, scop addresses both issues and can provide a better characterization of concordance or reproducibility . for case iii .
( d ) @xmath85 for case iii.,scaledwidth=80.0% ] unlike the idr approach which only produces one idr curve , scop creates two curves , one for each rank list . using these two curves
, one can explore characteristics specific to each rank list .
for instance , besides measuring the overall concordance of two rank lists , idr is also used to determine where to cut the rank lists to keep only the loci that are likely to be reproducible in independent experiments , that is , it serves a role similar to false discovery rate ( fdr ) . in many real applications
, one rank list is obtained from a high quality experiment , whereas the other list is obtained from a noisy dataset ; thus , the two rank lists may be asymmetric in terms of their signal - to - noise ratio . in such a scenario
, one may want to have a more detailed view of each list .
since the idr approach produces only one idr curve for loci shared by both lists , it does not reveal the asymmetry between the two lists . using this idr curve to choose cutoff forces the same cutoff to be applied to both lists .
this may result in decreased power in detecting reproducible loci .
in contrast , scop allows one to estimate idr separately for each list and to observe the asymmetry of data quality for the two lists . to demonstrate , we generated scores for 10,000 loci and created two rank lists using a procedure similar to case i in section 3.1 .
for both lists , the signal and noise components were @xmath90 and @xmath81 respectively .
the latent variables in both lists were censored at -1.65 . the mixture proportions of the four co - existence patterns were set as @xmath91 . in this case , 70% of loci in the complete list 1 are signals , of which only 20% are also reproducible in complete list 2 , corresponding to signal proportion of 96% and 28% in the observed lists 1 and 2 .
this simulation is referred to as case iii . when the idr approach was applied to analyze the loci present in both lists , the idr estimates were very conservative compared to the true fdr ( figure 3(b ) ) .
this is because the asymmetry of the two lists lead to high variability , inflating the error rate estimates .
in contrast , the @xmath86 and @xmath85 curves produced by scop accurately estimated the proportion of irreproducible signals in each list ( figure 3(c)(d ) ) and indicated that the two lists have asymmetric signals .
thus , a separate analysis rather than a pooled one is needed for these two lists .
moreover , figure 3(d ) once again illustrates that naivevenn can underestimate the irreproducibility .
the reason is that it fails to distinguish between the signals and noise in the overlap part of the venn diagram , and hence count both of them in the calculation .
we downloaded two replicate chip - seq experiments for transcription factor nf - kb in cell line gm10847 from the encode @xcite together with their corresponding input data ( table 1 ) .
peak list a was called using cisgenome @xcite by comparing sample 1 with sample 3 and 4 at cutoff of fdr=0.01 ; similarly , peak list b was called by comparing sample 2 with sample 3 and 4 . for each peak
, we extracted the 150bp window centered at the peak summit to ensure the same length in peaks .
we then compared the two peak lists .
the venn diagram in figure 4(a ) shows that the majority of the loci in these two peak lists were different .
only 39.0% of loci in list a and 20.5% of loci in list b were found in the other list .
nevertheless , the idr analysis of the shared loci gives a low idr estimate of 0.015 , misleadingly suggesting high reproducibility between the two replicates ( figure 4(b ) ) .
in contrast , scop was able to show that the two replicate experiments have low reproducibility and high idrs ( figure 4(c)(d ) ) . for replicate a. ( d ) @xmath85 for replicate b.,scaledwidth=80.0% ]
foxa transcription factors are a key family of tfs that regulate gene activities in liver cancer .
biologists are interested in how members in this tf family interact with each other and whether different members bind to the same genomic loci in liver cancer cells .
the encode project has generated chip - seq data for both foxa1 and foxa2 in a liver cancer cell line hepg2 .
these data can be used to answer the questions raised in section 1 . using cisgenome @xcite
, we called 65,535 binding peaks for foxa1 ( comparing sample 5 and 6 with sample 9 and 10 ) and 48,503 peaks for foxa2 ( comparing sample 7 and 8 with sample 9 and 10 ) , respectively at the fdr=0.01 cutoff . finally , we applied scop to characterize the concordance between the two lists .
figure 5(a ) shows that among the top ranked foxa1 peak regions , about 60% are also bound by foxa2 .
as one moves to the lower ranked foxa1 peaks , a lower percentage are simultaneously bound by foxa2 .
thus , robust foxa1 binding seems to require foxa2 binding at the same location .
in contrast , the very top ranked foxa2 peaks are more likely to be foxa2 specific and less likely to be shared for foxa1 binding .
the middle ranked foxa2 peaks are more often bound by foxa1 .
the co - binding proportion drops again for foxa2 peaks with low ranking which are increasingly more likely to be noise .
this suggests that foxa2 may play its regulatory role in a different mode compared to foxa1 .
this information is not immediately revealed by the venn diagram and idr approach .
in summary , scop offers a new solution for comparing two rank lists .
scop takes into account both the overall proportion of overlap shared by the two lists and the consistency of ranks along them .
this overcomes the shortcomings of the venn diagram and the idr approach , and allows better characterizing of the concordance and global reproducibility between two datasets .
our simulation studies show that drawing conclusions on concordance from venn diagrams may not reveal all the information in the data .
the same degree of overlap may correspond to different signal - to - noise ratio .
idr , on the other hand , is limited in terms of characterizing the global reproducibility between two datasets since it focuses on analyzing loci shared by both lists . in light of these results , the scop curves should provide a better solution to assessing data quality ( e.g. , reproducibility between replicate chip - seq samples ) and computational algorithms ( e.g. , evaluate consistency of the results when a method is applied to two replicate experiments ) in projects such as encode .
our current model considers the problem of comparing two rank lists .
an interesting future research topic is how to extend it to comparing multiple rank lists .
currently , one can apply scop to compare each pair of lists .
however , this pairwise comparison approach does not directly reveal higher order relationships .
for instance , with three datasets , one can also ask how many loci are shared by all three lists in addition to asking how many loci are shared by each pair of lists . for @xmath92 rank lists , there are @xmath93 combinatorial signal coexistence patterns . as @xmath92 increases , the complexity of the problem increases exponentially .
efficient ways to perform the comparison and summarize results , similar to those in @xcite , need to be developed in order to solve this problem .
currently , an r package for scop is available upon request .
the package will soon be submitted to bioconductor .
here we present the details of the iterative algorithms used to estimate @xmath94 . 1 . initialize parameters @xmath95 .
2 . estimate the survival function @xmath96 using the kaplan - meier estimator .
3 . compute the pseudo - data @xmath97 . since @xmath98 does not have a closed form , @xmath99
is first computed on a grid of 5,000 points over the range @xmath100 $ ] .
@xmath101 is then obtained through linear interpolation on the grid .
4 . run em algorithm to search for @xmath102 that maximizes the log - likelihood of pseudo data @xmath103 .
the resulting @xmath102 is denoted as @xmath104 .
iterate between steps 3 and 4 until the change in log - likelihood between the two nearby iterations is less than a pre - specified threshold .
below are details of the em algorithm in step 4 . in the e - step ,
one evaluates the q - function @xmath105 here the expectation is taken with respect to probability distribution @xmath106 .
@xmath107 therefore , @xmath108 in the m - step , one finds @xmath59 that maximize the q - function @xmath109 .
denote them by @xmath110 .
solving @xmath111 we have : @xmath112 recall : @xmath113 and @xmath114 @xmath115 @xmath116 @xmath117 @xmath118 can be computed by replacing @xmath119 with @xmath120 accordingly .
only @xmath121 in equation a.3 involves @xmath122 .
because @xmath123 , the tail probability of a normal distribution , has no close form , we use the r function @xmath124 with the `` l - bfgs - b '' option to obtain the values that maximize @xmath121 .
@xmath125 are searched in a similar fashion to maximize @xmath126 .
the authors thank professor mei - cheng wang and members from the hopkins slam and genomics working group for their helpful discussions and suggestions . 9 barrett t , wilhite se , ledoux p , evangelista c , kim if , tomashevsky m , marshall ka , phillippy kh , sherman pm , holko m , yefanov a , lee h , zhang n , robertson cl , serova n , davis s , soboleva a ( 2013 ) ncbi geo : archive for functional genomics data sets update .
_ nucleic acids research _ * 41 * , ( d1 ) : d991d995 .
benjamini , y and hochberg , y ( 1995 ) controlling the false discovery rate : a practical and powerful approach to multiple testing . _ journal of the royal statistical society , series b ( methodological ) _ * * 57 * * ( 1 ) : 289300 celniker se , dillon la , gerstein mb , gunsalus kc , henikoff s , karpen gh , kellis m , lai ec , lieb jd , macalpine dm , micklem g , piano f , snyder m , stein l , white kp , waterston rh , modencode consortium ( 2009 ) unlocking the secrets of the genome .
_ nature _ * 459 * , 92730 .
irizarry ra , waren d , spencer f , kim if , biswal sh , frank bc , gabrielson e , garcia jgn , geoghegan j , germino g , griffin c , hilmer sc , hoffman e , jedlicka ae , kawasaki e , martinez - murillo f , morsberger l , lee h , petersen d , quackenbush j , scott a , wilson m , yang y , ye sq , yu w ( 2005 ) multiple - laboratory comparison of microarray platforms . _ nature methods_. * 2 * , ( 5 ) 345349 landt sg , marinov gk , kundale a , kheradpour p , pauli f , batzoglou s , bernstein be , bickel p , brown jb , cayting p , chen y , desalvo g , epstein c , fisher - aylor ki , euskirchen g , gerstein m , gertz j , hartemink aj , hoffman mm , iyer vr , jung yl , karmakar s , kellis m , kharchenko pv , li q , liu t , liu xs , ma l , et al.(2012 ) chip - seq guidelines and practices of the encode and modencode consortial . _ genome research_. * 22 * ( 9):181331
. mikkelsen ts , ku m , jaffe db , issac b , lieberman e , giannoukos g , alvarez p , brockman w , kim tk , koche rp , lee w , mendenhall e , odonovan a , presser a , russ c , xie x , meissner a , wernig m , jaenisch r , nusbaum c , lander es , bernstein be ( 2007 ) genome - wide maps of chromatin state in pluripotent and lineage - committed cells .
_ nature _ * 448 * , 55360 nan b , lin xh , lisabeth ld , harlow s ( 2006 ) piecewise constant cross - ratio estimation for association of age at a marker event and age at menopause . _
journal of american statistician association_. * 101 * ( 437):6577 | analyses of high - throughput genomic data often lead to ranked lists of genomic loci .
how to characterize concordant signals between two rank lists is a common problem with many applications .
one example is measuring the reproducibility between two replicate experiments .
another is to characterize the interaction and co - binding between two transcription factors ( tf ) based on the overlap between their binding sites . as an exploratory tool
, the simple venn diagram approach can be used to show the common loci between two lists . however , this approach does not account for changes in overlap with decreasing ranks , which may contain useful information for studying similarities or dissimilarities of the two lists . the recently proposed irreproducible discovery rate ( idr ) approach compares two rank lists using a copula mixture model .
this model considers the rank correlation between two lists .
however , it only analyzes the genomic loci that appear in both lists , thereby only measuring signal concordance in the overlapping set of the two lists .
when two lists have little overlap but loci in their overlapping set have high concordance in terms of rank , the original idr approach may misleadingly claim that the two rank lists are highly reproducible when they are indeed not . in this article
, we propose to address the various issues above by translating the problem into a bivariate survival problem .
a survival copula mixture model is developed to characterize concordant signals in two rank lists .
the effectiveness of this approach is demonstrated using both simulations and real data .
yy wei and hk ji genomic rank lists comparison * keywords * genomics ; high - throughput experiments ; mixture model ; survival copula ; em algorithm ; reproducibility ; co - binding of transcription factors . |
splitting functions are universal quantities in qcd . they govern the collinear evolution in hard scattering processes with hadrons in the initial or final state . in reactions with initial state protons the parton luminosity is parametrized by parton distribution functions whose scale dependence is subject to evolution equations with space - like kinematics ( a space - like hard scale @xmath0 ) and splitting functions @xmath1 , where @xmath2 denotes the parton s momentum fraction of the proton momentum .
similarly , for processes with identified hadrons in the final state the parton to hadron transition is described by the parton fragmentation distributions @xmath3 , where @xmath2 represents the fractional momentum of the final - state parton @xmath4 transferred to the outgoing hadron @xmath5 and @xmath6 is a time - like hard scale .
the scale dependence of the fragmentation distributions is controlled by the so - called time - like splitting functions @xmath7 , and is given by @xmath8 where the summation runs over the number @xmath9 of effectively massless quark flavors and the gluon , @xmath10 for @xmath11 .
the time - like splitting functions @xmath12 can be computed in perturbation theory in powers of the strong coupling @xmath13 , @xmath14 where we normalize the expansion parameter as @xmath15 . at one and two loops , the leading ( lo ) and next - to - leading order ( nlo ) splitting functions @xmath16 and @xmath17
have been obtained with different methods in the past , e.g. , from the collinear singularities in inclusive hadron production in electron - positron annihilation @xcite .
in contrast , the next - to - next - to - leading order ( nnlo ) terms @xmath18 at three loops have been determined by crossing relations from the respective functions @xmath19 for space - like kinematics based on analytic continuation in the scaling variable @xmath2 , i.e. , mapping @xmath20 @xcite .
in fact , the lo space - like and time - like splitting functions are identical ( up to transposition ) , a fact known as the gribov - lipatov relation @xcite . at higher orders such relations between the space - like splitting functions , or their analytic continuations , and their time - like counterparts do not hold in the usual @xmath21 scheme @xcite .
analytic continuations of the corresponding physical evolution kernels in the respective kinematics can , however , be used for a constructive approach to the time - like splitting functions and yield relations , which fix the components of the matrix @xmath18 at three loops @xcite . for the diagonal terms
@xmath22 these results agree with assumptions about the universality of evolution kernels @xcite , while for the off - diagonal terms an uncertainty in the time - like quark - gluon splitting function @xmath23 remains .
this motivates a direct calculation of the time - like splitting functions in perturbative qcd from a physical process .
we choose electron - positron annihilation and consider inclusive hadron production through photon exchange as well as top - quark mediated decay of the higgs boson into hadrons in the effective theory @xcite .
the report is organized as follows . in section [ sec:2 ]
we show how to reconstruct time - like splitting functions from the bare fragmentation functions in the framework of mass factorization , picking the process @xmath24 as an example .
section [ sec:3 ] is dedicated to some technical aspects of the calculation . here
we briefly discuss , in particular , how to perform final - state integration with the help of integration - by - parts ( ibp ) method and how to find master integrals from differential equations .
in addition , to illustrate our approach we provide some examples relevant to the calculation of nlo corrections to time - like splitting functions .
we summarize in section [ sec:4 ] .
let us consider the relevant parton processes in electron - positron annihilation @xmath25 with photon ( @xmath26 ) exchange and higgs ( @xmath27 ) boson exchange in the effective theory and the partons @xmath28 with momentum @xmath29 .
for the photon - exchange process ( [ eq : process ] ) , following the notation in @xcite , the unpolarized differential cross - section in @xmath30 dimensions is given by @xmath31 where @xmath32 denotes an angle between the beam and parton momentum @xmath29 .
the scaling variable @xmath2 is defined as @xmath33 in the center - of - mass frame of the @xmath34 pair @xmath2 can be interpreted as a fraction of the beam energy carried by the parton with momentum @xmath29 . the _ transverse _ , _ longitudinal _ , and _ asymmetric _ fragmentation functions are defined as @xmath35 , @xmath36 , and @xmath37 respectively .
they are the analog quantities of the deep - inelastic structure functions for space - like @xmath38 ( @xmath0 ) . in the context of the _ mass factorization _ formalism , the transverse fragmentation function @xmath39 in the @xmath21 scheme can be written as @xmath40 and explicitly , in terms of coefficients of the qcd @xmath41-function and splitting functions up to nnlo for @xmath42 @xmath43 where summation over the repeated indices @xmath44 , mellin convolution in @xmath2-space , and the normalization @xmath45 is understood . the term in eq .
proportional to @xmath46 for @xmath47 contains the off - diagonal time - like splitting function @xmath48 ( plus some terms of lower orders which are known ) .
the fragmentation functions @xmath49 for the higgs boson decay into hadrons in the effective theory @xcite with the normalization @xmath50 , i.e. @xmath51 , allow the extraction of @xmath52 ( in which we are interested ) from @xmath53 .
the corresponding expressions for the mass factorization can be obtained by replacing @xmath54 in eqs .
( [ eq : ft1][eq : ft3 ] ) . in perturbative qcd
, the transverse fragmentation function @xmath39 can be calculated with standard means as ( see @xcite ) @xmath55 where the hadronic tensor @xmath56 ( for the case of photon exchange ) reads @xmath57 here @xmath58 is @xmath59-particle real phase - space , defined in eq . ,
and @xmath60 describes amplitudes for the process of eq .
and the index @xmath61 corresponds to the polarization of the virtual photon , which is summed over . to generate feynman diagrams for the processes and ( [ eq : process - phi ] ) we use ` qgraf ` @xcite and
process its output further with the help of ` form`@xcite , i.e. , contract indices , calculate dirac traces and @xmath62 color factors , make partial fractioning , etc .
the resulting scalar expression is suitable for the further final - state integration as described in section [ sec:3 ] . in the remaining part of this section
let us discuss which contributions to the nnlo splitting functions are known and which actually need to be calculated .
0.3 ( tagged gluon , photon exchange process ) ; respectively @xmath63 ( tagged quark , higgs boson exchange process ) . , title="fig : " ] 0.3 0.3 as described in ( * ? ? ?
* eq . ( 38 ) ) the uncertainty for @xmath64 is proportional to @xmath65 , which means that only contributions with closed or cut fermion loops should be considered .
for that reason , we do not need to account for the topologies depicted in fig .
, which are sub - leading in @xmath9 . the two - loop contributions in fig .
( [ fig : amp3a ] , [ fig : amp3b ] ) have been recently calculated in @xcite , while the the contributions in fig . can be constructed from one - loop helicity amplitudes calculated in @xcite ( @xmath66 partons ) and @xcite ( @xmath67 partons ) , respectively . for those contributions the final - state integration is of nlo complexity and , hence ,
is considered to be simple . finally , the contributions of fig . are not known and should be computed by explicitly performing the final - state integration .
this is a non - trivial task and in the next section we describe how to complete it . 0.3 0.3 0.3 0.3 0.3
the phase - space for the @xmath59-particle final - state integration in eq . reads as @xmath68 integrals of this type can be calculated analytically with up to @xmath69 particles in the final state using explicit parametrizations of the phase - space ( for example , see @xcite ) .
however , with @xmath70 particles in the final state such an approach does not look promising .
alternatively , we propose to find such integrals with the method of differential equations ( see , e.g. , @xcite ) , which has been used in a number of recent state - of - the - art computations .
the idea is as follows 1 .
find the integration - by - parts ( ibp @xcite ) rules and reduce all the integrals to the small set of masters .
2 . build a system of differential equations in @xmath2 for the masters .
3 . choose an appropriate basis of new masters , so that the r.h.s of the system of equations vanishes in the limit @xmath71 .
4 . solve a new system of equations as a series in the parameter @xmath72 .
5 . find the integration constants from inclusive integrals .
let us briefly discuss each of the above steps and provide some examples from calculations of the contributions in fig . .
integration - by - parts ( ibp ) @xcite is a powerful tool to reduce a set of integrals of common structure to the small number of master integrals .
nowadays , there are several tools to automatically generate ibp reduction rules , e.g. ` fire ` @xcite , ` litered ` @xcite or ` reduze ` @xcite .
we chose ` litered`out of that list because this tool supports cut propagators , a feature that leads to additional simplifications in final - state integrals , which is essential at higher orders . at nlo ( see fig . ) ` litered`demonstrated good performance and generated all the ibp rules ( in several hours on a standard pc ) and found 9 master integrals , depicted in fig .
[ fig:4 ] . in order to parametrize those integrals
we introduce the following notation @xmath73 with propagators for the contributions from fig .
defined as @xmath74 0.175 ) .
the fat line denotes a propagator raised to second power.,title="fig : " ] 0.175 ) .
the fat line denotes a propagator raised to second power.,title="fig : " ] 0.175 ) .
the fat line denotes a propagator raised to second power.,title="fig : " ] 0.175 ) .
the fat line denotes a propagator raised to second power.,title="fig : " ] 0.175 ) .
the fat line denotes a propagator raised to second power.,title="fig : " ] 0.175 ) .
the fat line denotes a propagator raised to second power.,title="fig : " ] 0.175 ) .
the fat line denotes a propagator raised to second power.,title="fig : " ] 0.175 ) .
the fat line denotes a propagator raised to second power.,title="fig : " ] 0.175 ) .
the fat line denotes a propagator raised to second power.,title="fig : " ] it is worth to mention that this set of masters differs from the one in mellin space calculated for the same contribution , see @xcite .
for the contributions from fig .
` litered`can solve the major part of ibp sectors .
however for a few it takes months of cpu time with no guarantee of completion .
preliminary attempts show that these sectors can be solved with a private version of ` reduze ` , which also features support for cut propagators .
the master integrals from fig . may be integrated explicitly , which at higher orders becomes a challenging task .
a simpler approach is to solve a linear system of differential equations , which is constructed as @xmath75 where matrix @xmath76 is obtained by reducing the l.h.s . of eq . using ibp reduction rules . in the case of masters from fig .
we obtain the following matrix @xmath77 solutions of this system can be found in terms of harmonic polylogarithms ( hpl ) , implemented in the ` harmpol`@xcite package for ` form`or in the ` hpl`@xcite package for ` mathematica ` .
the last step in solving for masters is to fix integration constants which appear in solutions of differential equations .
such constants can be determined conveniently from the inclusive integrals calculated in @xcite .
in the inclusive case , there are only 4 master integrals ( fig . [ fig:5 ] ) , hence ibp rules exist in order to reduce the 9 masters integrated in the @xmath2 variable .
@xmath78 which is basically the first mellin moment .
0.175 0.175 0.185
0.185 such rules can be found with the help of `
litered`and read @xmath79 which are exact in @xmath80 dimensions , i.e. , to all orders in @xmath72 .
in these proceedings we have motivated the necessity for a nnlo qcd computation of the time - like splitting functions from first principles .
we have outlined a strategy how to calculate the required three - loop contributions . based on the physical process
@xmath24 , we have analyzed the transverse fragmentation function @xmath39 in @xmath2-space using dimensional regularization .
the most challenging part of our approach lies in the evaluation of the final - state integrals with a projection to @xmath2-space . for the nnlo case
, this corresponds to four - loop integrals with one massive leg and cut internal propagators .
thanks to an analysis in mellin space and to partial results available in the literature , the number of topologies , and hence the integrals to calculate , can be largely reduced . despite such simplifications
there are , however , still about 80 integrals left to calculate and we employ the integration - by - parts technique and the method of differential equations to obtain the master integrals .
the latter method has become widely used in the last years and we have described its application to our calculations . this work has been supported by the research executive agency ( rea ) with the european union grant pitn - ga-2010 - 264564 ( lhcphenonet ) . the work of o.g . has been partly supported by narodowe centrum nauki with the sonata bis grant dec-2013/10/e / st2/00656 .
thanks for the hospitality of the theory group of the university of hamburg where major part of this research was done .
p. rijken and w. van neerven , _ higher order qcd corrections to the transverse and longitudinal fragmentation functions in electron - positron annihilation _ , _ nucl.phys . _
* b487 * ( 1997 ) 233282 , [ http://arxiv.org/abs/hep-ph/9609377[hep-ph/9609377 ] ] .
a. mitov and s. moch , _ qcd corrections to semi - inclusive hadron production in electron - positron annihilation at two loops _
_ nucl.phys .
_ * b751 * ( 2006 ) 1852 , [ http://arxiv.org/abs/hep-ph/0604160[hep-ph/0604160 ] ] .
a. mitov , s. moch , and a. vogt , _ next - to - next - to - leading order evolution of non - singlet fragmentation functions _ , _ phys.lett . _ * b638 * ( 2006 ) 6167 , [ http://arxiv.org/abs/hep-ph/0604053[hep-ph/0604053 ] ] . a. almasy , s. moch , and a. vogt , _ on the next - to - next - to - leading order evolution of flavour - singlet fragmentation functions _ , _ nucl.phys . _
* b854 * ( 2012 ) 133152 , [ http://arxiv.org/abs/1107.2263 [ arxiv:1107.2263 ] ] . c. duhr , t. gehrmann , and m. jaquier , _
two - loop splitting amplitudes and the single - real contribution to inclusive higgs production at n@xmath84lo _ , _ jhep _ * 1502 * ( 2015 ) 077 , [ http://arxiv.org/abs/1411.3587 [ arxiv:1411.3587 ] ] .
s. badger , e. nigel glover , p. mastrolia , and c. williams , _ one - loop higgs plus four gluon amplitudes : full analytic results _ , _ jhep _ * 1001 * ( 2010 ) 036 , [ http://arxiv.org/abs/0909.4475[arxiv:0909.4475 ] ] .
a. mitov and s. moch , _ qcd corrections to semi - inclusive hadron production in electron - positron annihilation at two loops _ , _
_ * b751 * ( 2006 ) 1852 , [ http://arxiv.org/abs/hep-ph/0604160[hep-ph/0604160 ] ] . | we report on the status of a direct computation of the time - like splitting functions at next - to - next - to - leading order in qcd .
time - like splitting functions govern the collinear kinematics of inclusive hadron production and the evolution of the parton fragmentation distributions .
current knowledge about them at three loops has been inferred by means of crossing symmetry from their related space - like counterparts , which has left certain parts of the off - diagonal quark - gluon splitting function undetermined .
this motivates an independent calculation from first principles .
we review the tools and methods which are applied to attack the problem . |
observations link grbs brief flashes of @xmath0-ray light emitted by sources at cosmological distances with star - forming galaxies and high mass stars @xcite .
these results indicate that grbs are produced by a rare type of supernova where the evolved core of a massive star collapses to a black hole .
normal supernovae , where the stellar core collapses to form a neutron star , are thought to produce nonrelativistic supernova remnant shocks that accelerate cosmic rays with energies @xmath2 ev . the origin of higher energy cosmic rays is controversial .
one possibility is that high - energy ( @xmath3 ev ) cosmic rays are accelerated by the relativistic shocks formed by grb explosions in the galaxy and throughout the universe @xcite , dating back to earlier suggestions of a galactic origin of high - energy cosmic rays @xcite .
if grbs accelerate high - energy cosmic rays , then galactic grbs could be detected as cosmic - ray sources from neutron @xmath4-decay emissions @xcite .
grbs with their jets oriented toward the earth have potentially lethal consequences , and may have contributed to past extinction episodes @xcite .
the sources of the @xmath5 ev cosmic - ray excesses reported from measurements made with the sugar ( sydney university giant air shower recorder ) and agasa ( akeno giant air shower array ) detectors ( @xcite ; see @xcite for review ) have been proposed to result from past grbs in the galaxy @xcite .
to investigate these ideas , the rate of grb events at different fluence levels is estimated , based on recent findings about beaming in grbs @xcite .
a 3d propagation model is used to simulate the sequence of irradiation events that occurs when a grb jet is pointed towards earth .
the results suggest that a grb jet could have produced radiations that contributed to the ordovician extinction event @xcite .
our results are contrary to the claim that the sugar excess could be produced by a grb in the galaxy .
if the auger observatory confirms the sugar cosmic - ray point source , then a model of high - energy cosmic rays from galactic grbs is incomplete .
most of the electromagnetic radiation from a grb is emitted during the prompt and early afterglow phases on timescales of minutes to hours .
we define the bolometric photon fluence @xmath6 with reference to the solar energy fluence @xmath7 ergs @xmath8 received at earth in one second .
significant effects on atmospheric chemistry through formation of nitrous oxide compounds and depletion of the ozone layer is found when @xmath9 @xmath10 @xcite , taking into account the very hard incident radiation spectrum of grbs .
reprocessing of incident grb radiation into biologically effective 200 320 nm uv radiation ( i.e. , with @xmath11 lethality ) on eukaryotes occurs when @xmath12 10@xmath13 @xcite .
achromatic beaming breaks in grb optical / ir afterglow light curves @xcite , if due to jetted grbs , imply typical grb jet opening half - angles @xmath14 .
analyses @xcite show that long - duration grbs have a standard total energy @xmath15 ergs , with @xmath16 , a variance by a factor of @xmath17 , and a low @xmath18 population . here
@xmath18 is the apparent isotropic @xmath0-ray energy release inferred directly from observations .
if the cone of emission from a grb at distance @xmath19 intercepts the line - of - sight to earth , then the radiant fluence is given by @xmath20 .
thus the maximum sampling distance @xmath21 of a grb with apparent isotropic @xmath0-ray energy release @xmath18 to be detected at the fluence level @xmath22 is @xmath23 when @xmath24 , the sampling distance to a typical grb exceeds the @xmath25 pc disk scale height of molecular clouds and ob associations , and we can approximate the distribution of grbs in the galaxy by a uniform disk of radius @xmath26 kpc .
the fluence size distribution of grbs in this approximation is given by @xmath27 where @xmath28 is the rate of grbs in the milky way , and @xmath29 is the probability that the earth lies within the emission cone of a 2-sided jet .
if grbs follow the star formation rate history of the universe , then one grb occurs every @xmath30 years in the milky way @xcite , with @xmath31 1 .
thus @xmath32 estimates ( [ rsamp ] ) and ( [ nsc_1 ] ) show that a grb at a distance @xmath33 kpc with @xmath34 takes place about once every gyr , and more frequently if @xmath35 .
we have developed a numerical model where cosmic rays move in response to a large - scale magnetic field that traces the spiral arm structure of the galaxy , and diffuse through pitch - angle scattering due to magnetic turbulence .
the magnetic field @xmath36 of the galaxy is modeled as a bisymmertric spiral for the galaxy s disk , and a dipole magnetic field for the galaxy s halo @xcite .
the evolution of the particle momentum @xmath37 is found by solving the lorentz force equation @xmath38 , where @xmath39 and @xmath40 are the particle s charge and mass , respectively , @xmath41 is its velocity , and @xmath42 .
magnetic turbulence causes a particle to change its pitch angle by @xmath43 when travelling the mean - free - path @xmath44 .
the energy dependence of @xmath44 is obtained by extrapolating the expression for @xmath44 in a diffusion - model fit @xcite to the measured ionic flux near the knee of the cosmic - ray spectrum @xcite to high energies .
our approach assumes isotropic turbulence that is uniform in the disk and halo of the galaxy ; anisotropic turbulence is more realistic @xcite , but increases the number of free parameters in the model .
the particle s azimuth and cosine angle are randomly chosen between @xmath45 and @xmath46 and between @xmath47 and 1 , respectively , after every @xmath48 ^ 2/c\delta t$ ] steps , where @xmath49 is the time interval of each step in the numerical integration , which is set equal to a small fraction of the local gyroperiod .
the propagation calculations are performed in the test - particle approximation , assuming that the cosmic rays do not affect their surrounding environment .
the energy density of the cosmic - ray shell can , however , greatly exceed the magnetic field energy density of the ism , in which case the protons will sweep up material , causing strong adiabatic losses and rapid deceleration withn a sedov length scale of @xmath50 pc .
this forms a shock and a second phase of cosmic ray acceleration , which will require a hydrodynamic simulation to treat .
the cosmic - ray neutrons decay over the entire jet volume . comparing the ism magnetic - field energy density with the energy density of the neutron - decay protons over
their decay volume shows that the propagation of neutrons with @xmath51 thus our subsequent discussion of @xmath1 ev cosmic - ray neutrons is accurately described in the test particle approximation . the propagation of cosmic - ray protons and neutrons , with and without the effects of diffusive scattering ,
is illustrated in fig . 1 .
cosmic - ray neutrons travel @xmath52 ev ) kpc before decaying .
cosmic - ray protons with energies @xmath53 ev and cosmic - ray neutrons with energies @xmath54 ev escape almost directly into intergalactic space .
protons with energies @xmath55 ev diffusively escape from the galaxy through a combination of larmor motions and pitch - angle scatterings .
2 displays the cosmic - ray halo that surrounds a grb source @xmath56 s after the event .
the grb is modeled by radially - oriented jets with @xmath57 . a conical shell with effective angular extent larger than 0.1 radians forms as a result of directly accelerated protons and neutron - decay protons with @xmath58 ev .
the turbulent wave spectrum that is resonant with these high energy cosmic rays is poorly known . in a simulation with
pitch - angle scattering absent , energy - dependent features and wall - like structures are formed , as shown in the inset to fig . 2 . and
@xmath59 , respectively , with diffusive scattering omitted , and the green and blue curves illustrate the effects on protons with respective energies when diffusive pitch - angle scattering is included .
the red curve shows the path taken by a cosmic - ray neutron with @xmath60 and mean decay length of 3 kpc , including scattering effects on the path of the neutron - decay proton . ] in the ranges @xmath61 @xmath62 , @xmath62
@xmath63 , @xmath63
@xmath64 , and @xmath64
@xmath65 are red , green , turquoise , and purple , respectively .
cosmic - ray neutrons and neutron - decay protons are dark red , yellow , magenta , and dark blue in the respective energy ranges .
( bottom ) side view of the cosmic - ray halo when diffusive scattering is turned off ( grid lines are 2 kpc in separation ) .
, title="fig : " ] 0.2 in in the ranges @xmath61 @xmath62 , @xmath62
@xmath63 , @xmath63 @xmath64 , and @xmath64
@xmath65 are red , green , turquoise , and purple , respectively .
cosmic - ray neutrons and neutron - decay protons are dark red , yellow , magenta , and dark blue in the respective energy ranges .
( bottom ) side view of the cosmic - ray halo when diffusive scattering is turned off ( grid lines are 2 kpc in separation ) .
, title="fig : " ]
about once every several hundred million years , the earth is illuminated by prompt grb photon and neutral radiations of sufficient intensity to have significant effects on the biota through erythema from cascade uv flux @xcite and destruction of the ozone layer catalyzed by the formation of nitric oxide and no@xmath66 odd nitrogen " compounds @xcite .
the destruction of the ozone layer causes the solar uv fluence to greatly exceed the grb cascade uv fluence when the net effects of the grb radiation on the atmosphere are considered @xcite .
such an event might have been responsible for trilobite extinction in the ordovician era through the destruction of plankton @xcite .
if high - energy cosmic - ray production accompanies grbs , then galactic events could also have affected biological evolution due to dna radiation damage by the elevated ground - level muon fluxes induced by @xmath67 ev ( @xmath68 ) cosmic - ray neutrons @xcite .
the prompt energy fluence of neutrons is @xmath69 where an on - axis grb at a distance of @xmath70 kpc accelerates @xmath71 ergs of nonthermal neutrons above @xmath25 tev into a two - sided jet with beaming factor @xmath72% .
the prompt muon number flux due to this type of event is at the level @xmath73 @xcite for grbs within @xmath74 of the zenith .
this value is above the level @xmath75 for 50% mortality of human beings .
the situation could be worse because of various leptonic and hadronic radiation pathways in the grb leading to significant fluxes of @xmath76 mev tev radiation and enhanced cascade uv flux @xcite .
deflection of charged particles by the galactic magnetic field means that cosmic - ray protons arrive after the delay time @xmath77 for @xmath78 , where the characteristic larmor radius @xmath79 kpc is assumed to be much larger than the source distance , and @xmath80 is a characteristic deflection angle . for a grb 1 kpc away , @xmath81 ev protons arrive within angle @xmath80 of the source direction over @xmath82 years , taking @xmath83 .
diffusive scattering spreads out the arrival times and directions compared to this estimate , and implies a chirping " behavior where higher energy particles arrive first . from the preceding expression , the energy dependence of the time delay for a proton with lorentz factor @xmath84
is given by @xmath85 s for @xmath86 , and @xmath87 s for @xmath88 , using the form of @xmath89 from @xcite .
-0.5 in received at earth from an on - axis grb that occurred 1 kpc away .
the fluxes of cosmic - ray neutrons , neutron - decay protons , and protons are shown by the curves that decline , rise , and peak during the initial phase .
the prompt flux actually lasts for some minutes to hours , rather than the 2000 yrs indicated by the numerical simulation , which reflects the approximations of the numerical method .
, title="fig : " ] -1.75 in fig .
3 shows the time - dependence of the flux of cosmic - ray neutrons and protons with lorentz factors @xmath90 received from a grb source located one kpc from the earth .
the received flux of prompt neutrons with this range of lorentz factors would be @xmath91 $ ] @xmath8 s@xmath92 , where @xmath93 is the mean duration of the cosmic - ray emission event .
because of the numerical method , this duration has been artificially increased to @xmath94 s , so that if the actual duration of the event were 600 s , the flux would be @xmath95 greater and the duration @xmath95 shorter than shown in the figure , giving constant total fluence . the flux units shown in fig .
3 apply directly to the cosmic - ray protons and neutron - decay protons , and we see that the cosmic - ray fluence in the prompt and extended phases are roughly equal . the measured energy density of @xmath96 @xmath97 ev cosmic rays is @xmath98 ergs @xmath99 , so fig .
3 shows that the cosmic ray energy density at these energies would be @xmath100 orders of magnitude larger than the currently measured energy density during a period of @xmath101 yrs following such an event . by depositing larger cosmic ray fluence during a much shorter interval , the prompt flux has , however , much greater lethality and effect .
the occurrence of the sequence of extinction events in the ordovician due to a long lasting @xmath102 myr ice age @xcite could happen if the prompt blast induced a long term change in the climate , or if the delayed cosmic rays induce a glaciation @xcite . on -
axis events are considerably more damaging than off - axis events which , though more numerous , release @xmath103 ev cosmic rays that slowly diffuse towards earth over periods of thousands of years and longer .
the hypothesis @xcite that the @xmath54 ev cosmic - ray excesses detected with the agasa and sugar arrays @xcite are cosmic - ray neutrons from a grb is not , however , supported by our simulations .
the relativistic blast waves in a grb accelerate the highest energy cosmic rays over timescales of weeks or less @xcite , and the high - energy neutrons therefore arrive on this same timescale .
cosmic ray protons with energies @xmath104 ev are delayed over a timescale @xmath105 years from a source at the distance of the galactic center . for the sugar excess , which is coincident on the sub - degree ( @xmath106 )
angular scale with a point source , a grb would have to take place within weeks of the observation for cosmic - ray protons to maintain their direction to the source . including the requirement that the grb jet was also pointed towards earth means that an impulsive grb origin is excluded because such an event is highly improbable .
the greater ( @xmath107 ) extent of the agasa excess does not conclusively exclude a grb origin , but here the diffuse excess could simply reflect the greater pathlength for cosmic - ray proton collisions with spiral arm gas along the cygnus arm . because the sugar point source does not admit an impulsive grb solution , only cosmic rays from a persistent source , such as a microquasar , could make such an excess .
the hypothesis @xcite that grbs are sources of @xmath3 ev cosmic rays is therefore incompatible with such a source .
this cosmic - ray origin hypothesis will soon be tested by results from the auger observatory to confirm this source .
if the source is real , then the grb / cosmic - ray model is incomplete .
searches for neutron @xmath4-decay radiation in recent galactic grbs @xcite and around galaxies that host grbs @xcite provide further tests of the hypothesis that high - energy cosmic rays are accelerated by grbs .
the earth resides on the inner edge of a spiral arm , and not in an ob association where high - mass stars , and therefore grbs , are usually found .
the greater likelihood for intense irradiation events by grbs excludes ob associations from the galactic habitable zone @xcite except for planets with very thick ( @xmath108 gm @xmath8 ) atmospheres @xcite needed to limit the radiation effects . 0.1 in * acknowledgements : * this work is supported by the office of naval research and the nasa _ gamma ray large area space telescope _
( glast ) program .
we thank a. atoyan , j. d. kurfess , a. l. melott , k. e. mitman , and s. d. wick for discussions and comments on the manuscript .
we also thank the referees for their reports , including the request to clarify the applicability of the test - particle limit . | the rate of terrestrial irradiation events by galactic gamma - ray bursts ( grbs ) is estimated using recent standard - energy results .
we assume that grbs accelerate high - energy cosmic rays , and present results of three - dimensional simulations of cosmic rays moving in the galactic magnetic field and diffusing through pitch - angle scattering .
an on - axis grb extinction event begins with a powerful prompt @xmath0-ray and neutron pulse , followed by a longer - lived phase from cosmic - ray protons and neutron - decay protons that diffuse towards earth .
our results force a reinterpretation of reported @xmath1 ev cosmic - ray anisotropies and offer a rigorous test of the model where high - energy cosmic rays originate from grbs , which will soon be tested with the auger observatory . |
galactic cosmic rays ( crs ) are mainly constituted by relativistic protons and are believed to be accelerated at supernova remnant ( snr ) shocks via first order fermi mechanism @xcite .
though very popular , this scenario still needs to be conclusively proven by observations .
if crs are indeed accelerated at snrs , these objects must be gamma ray sources .
this is because the crs accelerated at the shock undergo inelastic proton proton interactions with the ambient medium and produce neutral pions which in turn decay into gamma rays @xcite .
several snrs have been detected in gamma rays at both tev ( e.g. * ? ? ?
* ) and gev ( e.g. * ? ? ?
* ) energies , in agreement with such expectations . however , it is often difficult to determine whether the origin of the gamma ray emission is hadronic , and thus related to the acceleration of crs , or due to leptonic mechanisms such as inverse compton scattering . for this reason ,
multi - wavelength studies of snrs have been extensively carried out in an attempt to solve this degeneracy .
though for some individual snrs it has been possible to ascribe the gamma ray emission exclusively and quite confidently to hadronic ( e.g. * ? ?
* ; * ? ? ?
* ) or leptonic ( e.g. * ? ? ?
* ; * ? ? ?
* ) processes , in other cases this ambiguity remains a problem . an alternative way to reveal
the presence of a cr source is by searching for the radiation produced by crs that escape the acceleration site @xcite . at some stage of the dynamical evolution of the snr
, crs are expected to leave the shock region and escape into the interstellar medium .
the details of the escape mechanism are still not very well understood ( see e.g. * ? ? ?
* and references therein ) , but it is generally believed that the ability of a snr in confining particles decreases gradually with the shock speed , with higher energy particles leaving the shock earlier than low energy ones
. once escaped , crs diffuse away from the snr and produce gamma
rays in interactions with the ambient gas . to date , some possible evidence for particle escape from snrs has been pushed forward by the observation of diffuse gamma
ray emission from the vicinity of the shell of the snrs w28 @xcite and w44 @xcite . in both cases ,
the emission is clearly located outside of the shell and it is spatially coincident with the location of massive molecular clouds ( mcs ) .
this would favor a scenario in which the mcs are illuminated by the runaway crs and , being very massive , become prominent gamma
ray sources ( for the case of w28 see e.g. * ? ? ?
* and discussion in section [ sect : w28 ] ) .
besides revealing the presence of a cr source , the gamma
ray emission from runaway crs can also be used to constrain the particles diffusion coefficient in the region surrounding the accelerator .
this is very important for several reasons : first of all , a theoretical determination of the diffusion coefficient is a very complex task ( see e.g. * ? ? ? * ; * ? ? ?
* ) and observational constraints are needed in order to guide and constrain models .
in addition to that , the diffusion of crs is believed to be a non linear process in which crs themselves generate via streaming instability the turbulence they scatter off ( e.g. * ? ? ? * ) .
this is particularly relevant close to cr sources , where the cr density is expected to be very high , and possibly sufficient to suppress significantly the diffusion coefficient through streaming instability @xcite .
thus , an empirical determination of the diffusion coefficient can reveal precious information on the ways in which particles and waves interact in astrophysical plasmas .
most of the studies aimed at the determination of the cr diffusion coefficient from gamma
ray observations rely on the assumption of _ isotropic diffusion _ @xcite .
the common rationale of these approaches can be summarized as follows : if snrs are the sources of crs , they have to convert a fraction @xmath0 of their explosion energy @xmath1 into accelerated particles . if the diffusion of crs proceeds isotropically , after a time @xmath2 from escape crs of a given energy @xmath3 are distributed roughly homogeneously within a distance @xmath4 from the snr . here , @xmath5 is the energy dependent diffusion coefficient of crs .
gamma ray observations of mcs located in the vicinity of w28 or w44 tells us which is the cr density @xmath6 needed to explain the observed emission .
according to what said above , such density has to be of the order of @xmath7 , where the factor @xmath8 contains the information on the shape of the spectral energy distribution of escaping crs .
for aged snrs such as w28 and w44 , the time @xmath2 after the escape can be identified with the snr age @xmath9 , and thus an expression for the diffusion coefficient can be obtained , and reads : @xmath10 .
since the values of all the physical quantities present on the right side of the equation can be inferred from observations , the expression provides a direct estimate of the diffusion coefficient . as an example , we summarize here the results obtained by @xcite in interpreting the gamma ray emission observed from the mcs located close to the snr w28 . they obtained a good fit to the gamma
ray spetrum measured by h.e.s.s . at photon energies @xmath11 gev by assuming a diffusion coefficient for @xmath12 tev crs of the order of @xmath13 @xmath14/s .
the corresponding diffusion length @xmath15 of these particles is of the order of 100 pc .
also the broad band gamma
ray spectrum from gev to tev energies can be fitted by adjusting the energy dependence of the diffusion coefficient and the distances between the snr and the clouds .
the important point here is the fact that the estimated diffusion coefficient is more than one order of magnitude smaller than the one normally adopted to describe the propagation of crs of energy @xmath16 tev in the galactic disk , which is @xmath17@xmath14/s @xcite .
these results are very similar to the ones obtained by other authors by means of similar modeling @xcite and seem to point toward a drop of the diffusion close to the snr w28 .
however , the validity of the assumption of isotropic diffusion of crs needs to be discussed .
in fact , if the intensity of the turbulent field @xmath18 on scales resonant with the larmor radius of particles is significantly smaller than the mean large scale field @xmath19 ( i.e. if @xmath20 ) , then _
cr diffusion becomes anisotropic _ , with particles diffusing preferentially along the magnetic field lines ( e.g. * ? ? ?
* ) . in the limiting
( but still reasonable ) case in which the perpendicular diffusion coefficient can be set equal to zero , the transport of crs across the mean field is mainly due to the wandering of magnetic field lines @xcite .
this is the situation that we investigate in this paper . to give a qualitative idea of the role that anisotropic diffusion can play in these study ,
let us consider an idealized case in which particles that escape a snr diffuse along a magnetic flux tube characterized by a very long coherence length ( i.e. the magnetic flux tube is preserved for a long distance ) . in this case
, after a time @xmath2 particle will diffuse up to a distance @xmath21 along the tube ( here @xmath22 is the _
parallel _ diffusion coefficient ) , while their transverse distribution will be equal to the radius of the snr shock at the time of their escape , @xmath23 , which is of the order of @xmath24 110 pc .
thus , the enhanced cr density in the flux tube will be proportional to @xmath25 instead of @xmath26 as in the isotropic case .
it is easy to see that the estimates of the diffusion coefficient based on the two opposite assumptions of isotropic and one - dimensional diffusion will differ by a factor of @xmath27 , which can be much larger than an order of magnitude !
thus , it is of paramount importance to investigate how the interpretation of gamma
ray observations depends on the assumptions made concerning cr diffusion . in section
[ sect : model ] we develop a model for cr propagation in which crs are strongly magnetized and diffuse uniquely along the magnetic field lines .
the wandering of the field lines is also taken into account , and a diffusion coefficient @xmath28 for the magnetic field lines that depends on the properties of the turbulent field is defined ( see e.g. * ? ? ?
in section [ sect : results ] the model is used to predict the spatial distribution of runaway crs and their spectrum .
finally , we apply the model to the case of the snr w28 in section [ sect : w28 ] . a good fit to data
is obtained , and the estimate of the parallel diffusion coefficient is found to depend on the level of anisotropy of the diffusion . for higher level of anisotropy , i.e. smaller values of @xmath28 , larger values of the particles diffusion coefficient are needed in order to fit data .
a discussion of the results and of future perspectives in this line of research can be found in section [ sect : conclusions ] .
consider a magnetic flux tube whose mean magnetic field @xmath19 is assumed to lie along the @xmath29-axis , perpendicular to the @xmath30 plane .
the wandering of magnetic field lines is due to long wavelength perturbations , i.e. perturbations on scales much larger than the particles gyroradii , with root mean square amplitude @xmath18 .
the condition for the validity of quasi
linear theory is @xmath31 , where @xmath32 and @xmath33 are the field coherence lengths perpendicular and parallel to @xmath19 , respectively @xcite . according to quasi
linear theory , the field lines passing in the vicinity of @xmath34 at @xmath35 are spread over a larger region as they reach a given @xmath29 . the probability distribution describing this spreading of field lines is gaussian and characterized by @xmath36 , where brackets indicate an ensemble average and @xmath28 is a diffusion coefficient for field lines @xcite . for a broad band fourier spectrum of the perturbation
, the coherence lengths can be expressed as @xmath37 , with @xmath38 and @xmath39 the characteristic wave
vectors of the perturbation @xcite . under these circumstances ,
the diffusion coefficient is @xmath40 @xcite .
it is also possible to define the lyapunov length @xmath41 which describes the exponential separation of field lines whose initial separation is smaller than @xmath32 @xcite .
this can be interpreted as the length above which the flux tube is disrupted by field line divergence . for fiducial values of the parameters
the length of the flux tube is of the order of a few hundred parsecs ( see e.g. * ? ? ?
we are interested here in studying the propagation of crs in the presence of magnetic field line wandering .
energetic particles diffuse along and across the magnetic field line as the result of resonant interactions with magnetic field perturbations .
such perturbations are characterized by length scales of the order of the particles larmor radii .
such scales are much smaller than the ones responsible for field line wandering . according to quasi
linear theory , the ratio between the parallel to perpendicular diffusion coefficient is @xmath42 , where @xmath43 is the particle s mean free path along the field line and @xmath44 is its gyroradius . in the interstellar medium it is believed that @xmath45 which implies @xmath46 ( e.g. * ? ? ? * ) .
thus , in the following we will neglect the diffusion of particles perpendicular to the field lines .
in other words , a given particle remains attached to the same field line . under this conditions , in a time interval @xmath47 a particle diffuses along a given field line a distance @xmath48 , but over such a distance @xmath49 along the @xmath29axis the field line is displaced by an amount @xmath50 .
this leads to @xcite : @xmath51 which describes a sub diffusive transport of particles perpendicular to the mean magnetic field @xmath19 .
this behavior of energetic particles due to the combination of the particle diffusion along the field lines and the random walk of the field lines themselves has been often referred to as _ compound diffusion _ , or _
anomalous diffusion_. models of compound diffusion have been developed and used in a great variety of contexts , to study phenomena like the heat transport in tokamak @xcite , the propagation of energetic particles in the solar wind @xcite , the confinement of crs in the galaxy @xcite and their acceleration at astrophysical shocks @xcite . in this paper
, we apply the formalism of compound diffusion to another context , which is the propagation of crs in the vicinity of their sources , i.e. snrs , after they escape the acceleration region .
we will show that in this situation an accurate modeling is needed in order to interpret in a correct way the gamma ray observations of mcs located in the vicinity of snrs . in order to describe the compound diffusion of crs we adopt the mathematical formalism developed by @xcite and we define @xmath52 as the probability to find a field line displaced by an amount @xmath53 after a step of length @xmath29 along the direction of the umperturbed field @xmath19 .
from what said above , it follows that : @xmath54\ ] ] which corresponds to a diffusive behavior of field lines ( flw stands for field line random walk ) .
a similar equation holds for the displacement @xmath55 .
this has to be combined with the probability @xmath56 that a particle moves a distance @xmath29 along the field line in a time @xmath57 . for diffusive transport of particles along the field
we have : @xmath58\ ] ] the probability for a particle to reach the position @xmath59 at the time @xmath2 , when its position at the time @xmath60 was ( @xmath61 ) , is then the product of @xmath62 with @xmath63 : @xmath64 in order to model the escape of crs from a snrs we assume , following @xcite , that particles are injected in the flux tube in the @xmath65-plane at @xmath35 , within a circular region whose radius is equal to the snr shock radius @xmath66 . since crs of different energy
are expected to escape the remnant at different times , the radius of the injection region is an energy dependent quantity .
following @xcite we assume a power law scaling to connect the particle energy of the runaway crs with the time after the supernova explosion : @xmath67 where the implicit assumption has been made that the maximum energy of crs accelerated in a snr @xmath68 is reached at the time @xmath69 which marks the transition between the free expansion and the sedov phases of the snr evolution , and that crs are gradually released in the interstellar medium from that time on .
the sedov phase is characterized by a scaling @xmath70 which gives : @xmath71 which is what is assumed in the following .
other parameterizations of the escape time of crs can be easily implemented . to describe the diffusion of crs along field lines
we adopt a diffusion coefficient which is a power - law in energy : @xmath82 where @xmath83 and @xmath84 are considered here as free parameters . before proceeding in computing the spatial distribution of crs around a snr
we notice that , for @xmath85 , eq . [ eq : frw ] does not provide a good description of the field line wandering .
the reason is that in this regime the lateral displacement of a field line after a step @xmath29 along @xmath19 is of the order of @xmath86 , since @xmath87 represents the angle between the unperturbed ( @xmath19 ) and total ( @xmath88 ) magnetic field @xcite .
an accurate and quantitative analysis of the behavior of a magnetic flux tube in this regime goes beyond the scope of this paper ( see * ? ? ?
* for a more detailed discussion ) .
however , in order to describe this regime in a qualitative way , for @xmath85 we substitute eq .
[ eq : frw ] with : @xmath89}{\pi ( b z)^2}\ ] ] where @xmath90 $ ] is the heaviside function , equal to 1 for @xmath91 and 0 for @xmath92 .
eq . [ eq : conical ] roughly mimics the behavior of a flux tube characterized by a opening angle @xmath93 . in the intermediate region
@xmath94 we use an interpolating function to bridge the behaviors described by eqns .
[ eq : frw ] and [ eq : conical ] .
in this section we compute the spatial distribution of crs expected in the vicinity of a snr at a given time after the explosion . we consider a typical supernova , characterized by the following fiducial values of parameters : an explosion energy of @xmath95erg , a mass of the ejecta equal to @xmath96 , and a density of the circumstellar medium @xmath97@xmath98 .
we further assume that a fraction @xmath99 of the supernova explosion energy is converted into crs , which are injected in the interstellar medium with a power law differential energy spectrum @xmath100 which extends from @xmath101gev to @xmath102pev ( approximately the position of the knee in the cr spectrum ) .
it is known from cr data that @xmath103 should be in the range @xmath104 @xcite .
we adopt here @xmath105 as a representative value . as described in sec .
[ sect : model ] , crs are gradually released from the snr during the sedov phase that goes from @xmath106 yr to @xmath107yr @xcite . for the parallel diffusion coefficient of crs ( eq . [ eq : diffusion ] ) we assume @xmath108@xmath14/s and @xmath109 .
-0.38 cm tev at a time @xmath110kyr after the explosion .
the left panel refers to an isotropic diffusion coefficient of cosmic rays equal to @xmath111@xmath14/s , while the right panel refers to an anisotropic diffusion scenario with @xmath112@xmath14/s , @xmath113pc , and @xmath114 .
the black cross marks the a position at which the cr over - density is equal in the two panels.,title="fig : " ] as a first step , we compare in fig .
[ fig : iso ] the results that are obtained if an isotropic diffusion coefficient is assumed ( as , e.g. , in * ? ? ?
* ; * ? ? ?
* ) , with the ones obtained for the anisotropic diffusion model that we consider in this paper . in both panels of fig .
[ fig : iso ] , the snr is located at the centre of the field and the color code refers to the excess of crs with respect to the average density of crs in the galaxy , which is ( e.g. * ? ? ?
* ) : @xmath115 over - densities are plotted for a particle energy of @xmath116tev and for a time @xmath117kyr after the supernova explosion . here the diffusion coefficient of the magnetic field lines is set equal to @xmath113pc , with @xmath114 ( different values of @xmath28 will be explored in the following ) .
this corresponds to a parallel coherence length of the perturbation of @xmath118pc .
the spatial distribution of crs is strikingly different in the two scenarios : spherically symmetric in the left panel , and strongly elongated in the direction of the magnetic field flux tube in the right panel .
a filamentary diffusion of crs was also found in the numerical simulations by @xcite .
the same parameters have been used to compute the over densities in the two scenarios , with the exception of the cr diffusion coefficient , which in the left panel has been assumed to be isotropic and equal to @xmath119@xmath14/s with @xmath120@xmath14/s .
the choice of two significantly different values for @xmath121 and @xmath83 , with @xmath122 has been made in order to obtain the same level of cr over density in the vicinity of the snr .
as an example , the black cross in fig . [ fig : iso ] shows a position , located @xmath123pc away from the centre of the explosion , where the cr over - density is identical in the two panels . to get comparable values for the cr over density ,
a much smaller ( isotropic ) diffusion coefficient @xmath121 is needed in order to compensate for the larger solid angle over which crs can propagate . as already stressed in the introduction
, this fact must be taken into account when interpreting the gamma
ray observations of molecular clouds illuminated by crs escaping from snrs . this will be discussed in sec .
[ sect : w28 ] , when the model developed here will be applied to fit the gamma
ray observations of the snr w28 . in fig .
[ fig : cr_map ] we show the cr over density around the snr for different values of the particle energy and of the time after the supernova explosion .
the upper , middle , and lower panels refer to a time of 6 , 19 , and 60 kyr after the explosion , respectively .
plots on the first , second , and third column refer to particle energies of @xmath123gev , @xmath124tev , and @xmath125tev , respectively .
the escape of crs is described by eq .
[ eq : escapetime ] , which states that higher energy crs are released first , and lower energy crs escape at later times .
this is the reason why there is no cr excess in the top
left panel of fig .
[ fig : cr_map ] : for the choice of parameters made here , for a snr age of @xmath126kyr particles with an energy of @xmath123gev are still confined within the snr shock . as the age of the snr increases , crs diffuse further away along the flux tube and fill a broader and broader region . as a consequence of that , the cr over density decreases accordingly .
it is evident from these maps that a molecular cloud located in the vicinity of the snr can be illuminated by the escaping crs and become a bright gamma
ray source only if it is located within the flux tube .
a nearby cloud which is not magnetically connected with the snr will not be illuminated by crs , despite its proximity to the snr .
all the plots in fig .
[ fig : cr_map ] refer to a region of size @xmath127pc around the snr . as said in sec .
[ sect : model ] , this roughly represents the expected length of a magnetic flux tube in the galaxy ( e.g. * ? ? ?
* ) . for distances larger than a few hundred parsecs from the snr
, the flux tube loses its identity and it is disrupted due to the exponential divergence of field lines .
thus , the results presented in this paper are accurate and reliable for distances up to a few hundred parsecs and less .
since we assumed here that crs are strongly magnetized , i.e. their remain attached to field lines , their transport across the mean magnetic field is solely governed by the diffusion of field lines , which is described by the diffusion coefficient @xmath28 .
this quantity determines how quickly the field lines diverge as a function of the displacement @xmath29 along the mean field .
this is demonstrated in fig .
[ fig : dm ] , where the cr over density for particles of energy @xmath124tev is shown .
the age of the snr is @xmath128kyr .
three different values for @xmath28 are considered : 0.5 , 1 , and @xmath129pc for the left , middle , and right panel , respectively .
unfortunately , our knowledge of the properties of the interstellar magnetic field is not good enough to allow a determination or an estimate of this parameter . for a snr located in a diffuse interstellar medium ( i.e. with no massive molecular clouds ) the morphology of the resulting gamma
ray emission due to cr proton
proton interaction with the ambient gas would closely follow the spatial distribution of crs .
thus , observing the diffuse gamma ray emission generated by runaway crs around snrs might serve as a tool to explore the structure of the interstellar magnetic field .
the detection of such diffuse emission is within the capabilities of future gamma
ray instruments such as the cherenkov telescope array @xcite . finally , we show in fig . [
fig : spectra ] the spectra of escaping crs at different distances from the snr and at different times after the supernova explosion .
each panel refers to a different epoch : 6 , 19 , and @xmath130kyr for the top , middle , and bottom panel , respectively .
solid curves show the spectra at three different positions on the @xmath29-axis : @xmath131pc , @xmath132pc , and @xmath133pc . for each of these positions we also show the spectra at different distances from the @xmath29-axis : @xmath134pc ( dotted lines ) and @xmath135pc ( dashed lines ) . at high energies , in almost all cases the energy spectra are power laws with slope @xmath136 , where @xmath103 is the slope of the injection spectrum of runaway crs and @xmath84 is the slope of the energy dependent diffusion coefficient ( eq . [ eq : diffusion ] ) .
such a behavior can be inferred from eq .
[ eq : parallel ] .
if one moves to larger values of @xmath29 , the on axis ( i.e. r = 0 ) high energy spectrum preserves the same slope , but its normalization decreases as @xmath137 .
this is due to the fact that the transverse section of the magnetic flux tube is increasing proportionally to @xmath29 , while the cr intensity along a field line is independent of @xmath29 for @xmath138 ( see eqns .
[ eq : frw ] and [ eq : parallel ] ) . a feature common to all the spectra plotted in fig .
[ fig : spectra ] is the presence of a low energy cutoff .
the cutoff is due to the fact that at a given time , only particles of sufficiently large energy had enough time to propagate over the distance @xmath29 .
the cutoff is moving towards lower energies if a longer time is considered , because particles with lower energies have then the time to reach a given position @xmath29 along the axis . finally , some curves ( for example the ones with @xmath139pc and @xmath140pc ) exhibit a quite sharp high energy cutoff .
this cutoff is due to the fact that particles of different energy are injected within different transverse sections of the flux tube .
higher energy particles are released earlier from the snr , when the shock radius is smaller , lower energy ones are injected later , when the shock radius is larger . while diffusing along the field lines ,
crs are displaced in the transverse direction due to field line wandering .
higher energy particles , which have been injected in a smaller region around @xmath35 , need on average a larger transverse displacement in order to reach a given distance @xmath141 from the @xmath29axis .
thus , for small enough @xmath29 , the opening of the magnetic flux tube might not be enough to allow high energy particles to reach @xmath141 , and this explains the presence of the cutoff .
kyr ( upper panel ) , @xmath142kyr ( middle panel ) and @xmath143kyr ( lower panel ) .
solid lines refer to spectra along the @xmath29-axis , oriented as the mean magnetic field , at three different positions ( @xmath144pc , @xmath145pc and @xmath146pc ) . for each distance @xmath29 , three different values of @xmath147 ( the perpendicular distance from the @xmath29-axis )
are also considered : @xmath148pc , @xmath149pc , and @xmath150pc .
the black lines show the cr background.,scaledwidth=47.0% ] in the next section we apply the model developed above to a specific object , i.e. the snr w28 and the molecular clouds located in its proximity .
such clouds have been detected in gamma rays and this has been interpreted by many authors as the result of their being illuminated by crs that escaped the snr .
we will demonstrate that a good agreement can be reached between the predictions of the model and observations and we will discuss the impact of this on the attempts to derive the particle diffusion coefficient close to snrs by means of gamma ray observations .
w28 is an old snr in its radiative phase of evolution , located in a region rich of dense molecular gas with average density @xmath151 . at an estimated distance of @xmath152
the snr shock radius is @xmath153 and its velocity @xmath154 ( e.g. * ? ? ?
* ) . by using the dynamical model by @xcite and assuming that the mass of the supernova ejecta is @xmath155 , it is possible to infer the supernova explosion energy ( @xmath156 ) , initial velocity ( @xmath157 ) , and age ( @xmath158 ) . gamma ray emission has been detected from the region surrounding w28 both at tev @xcite and gev energies @xcite , by hess , fermi , and agile , respectively
the tev emission correlates quite well with the position of three massive molecular clouds , one of which is interacting with the north - eastern part of the shell ( and corresponds to the tev source hess j1801 - 233 ) , and the other two being located to the south of the snr ( tev sources hess j1800 - 240 a and b ) .
the masses of these clouds can be estimated from co measurements and result in @xmath159 , @xmath160 , and @xmath161 , respectively , and their projected distances from the centre of the snr are @xmath24 12 , 20 , and @xmath162pc , respectively @xcite .
the gev emission roughly mimics the tev one , except for the fact that no significant emission is detected at the position of hess j1800 - 240 a. the gamma
ray emission from the clouds in the w28 region has been interpreted by many authors as the result of the interaction of crs that escaped w28 with the dense gas in the cloud @xcite .
all these approaches started from the assumption of isotropic diffusion of crs , and a general consensus was found on the fact that , in order to fit observations , the diffusion coefficient had to be suppressed by a factor of @xmath163 with respect to the average value in the galaxy , which is @xmath164 with @xmath165@xmath14/s and @xmath166 @xcite . in this section
we take a different approach and we apply the model developed in sec . [
sect : model ] to estimate the spectrum of crs and derive the @xmath167-ray emission expected from the clouds in the w28 region .
this approach is radically different from the ones mentioned above because it relies on the more physical assumption that the diffusion of crs is not isotropic , but proceeds mainly along the magnetic field lines .
the results of our modeling are shown in fig .
[ fig : w28 ] , where the gamma ray data from the three molecular clouds are plotted as blue open symbols ( data from fermi ) and black filled dots ( data from hess ) .
the emission from the sources hess j1801 - 233 , and hess j1800 - 240 a and b is plotted in the top , middle , and bottom panel , respectively .
the black dashed lines represent the contribution to the gamma
ray emission from the proton proton interactions of the crs in the galactic background with the inter cloud gas .
the blue long dashed lines represents the contribution to the emission from the runaway crs that escaped from w28 .
the solid red line is the total emission .
the gamma
ray fluxes have been computed following @xcite with an additional multiplicative factor 1.5 to take into account elements heavier than hydrogen in both cosmic rays and ambient gas @xcite .
a good agreement with observations is obtained is a parallel diffusion coefficient @xmath108@xmath14/s with @xmath109 is adopted , together with a diffusion coefficient for field lines @xmath113pc with @xmath114 .
moreover , we assumed that @xmath168 of the total explosion energy has been converted into crs with a spectrum proportional to @xmath169 and extending from @xmath116gev to @xmath170pev . in order to be illuminated by the escaping crs , the three molecular clouds have to be located in the proximity of the axis of the magnetic flux tube ( i.e. the direction of the local mean field ) .
the spectra reported in the figure refers to the positions @xmath171 , @xmath172 , and @xmath173pc and @xmath174 , @xmath175 , and @xmath176pc ( top to bottom panel , respectively ) .
it has to be noted that , due to the number of parameters involved in the model , other sets of parameter values might be found that provide an equally satisfactory fit to data .
this is not surprising , given that several previous modelings of this source provided an equally good fit to data by using a radically different picture ( i.e. isotropic diffusion of crs ) for the transport of particles .
moreover , while a quite small normalization of the ( isotropic ) diffusion coefficient , roughly of the order of @xmath177 @xmath14/s had to be adopted in order to fit data satisfactorily , in the anisotropic case we obtain a good agreement with data for a significantly larger value of the ( parallel ) diffusion coefficient of @xmath178 @xmath14/s .
it might be noticed that this number is close to the standard values inferred for the diffusion of crs in the galaxy .
thus , any attempt to constrain the cr diffusion coefficient from the observations of gamma
rays from the vicinity of snrs needs to take into account that an intrinsic uncertainty exists , and it is related to the unknown nature of the cr transport in the interstellar medium , and in particular to the unknown relative relevance of the transport parallel and perpendicular to the magnetic field lines .
the details of the transport of crs in the galaxy are still little understood . studies of the composition of crs provide us with an estimate of the average confinement time of crs within the galaxy , which can be translated into a spatially averaged diffusion coefficient for crs ( e.g. * ? ? ?
* ; * ? ? ?
whether the cr diffusion coefficient has large spatial variations or it is rather uniform throughout the galaxy is not known , thought a suppression of diffusion close to cr sources might be expected due to cr streaming instability @xcite . to this purpose ,
the detection of gamma ray emission from the vicinity of cr accelerators might be used to constrain the cr diffusion coefficient , and thus assess the importance of such suppression ( e.g. * ? ? ?
* ; * ? ? ?
this is because crs escaping the accelerators would produce gamma rays via proton proton interactions with the ambient medium .
both the morphology of the resulting emission and its spectrum would depend on the functional form ( i.e. energy dependence , level of anisotropy ) of the diffusion coefficient .
an object that has been extensively investigated in this context is the snr w28 .
three massive molecular clouds , with total mass in the @xmath179 range , are located in the vicinity of the snr shell and emit gamma rays .
this has been interpreted as the result of the illumination of the clouds by the crs that escaped the snr .
several models have been proposed to fit these observations , and all of them are based on the assumption that the diffusion of crs proceeds isotropically ( e.g. * ? ? ?
* ; * ? ? ?
* and see sec . [
sect : w28 ] for a complete list of references ) .
there is a general consensus on the fact that the ( isotropic ) diffusion coefficient has to be suppressed by a factor of @xmath180 with respect to the average galactic one in order to explain the observations .
this implies coefficients in the range @xmath181@xmath14/s . in this paper
, the assumption of isotropy of diffusion has been relaxed , and a more physically motivated situation have been investigated , in which crs propagate mainly along the magnetic field lines .
we considered here the limiting scenario in which the diffusion of crs across field lines is very small and thus can be neglected .
in such a situation , the transverse displacement of crs is uniquely due to the wandering of the field lines @xcite .
spectra and morphology of the spatial distribution of crs around snrs have been computed and described .
the main feature is the elongated , filamentary distribution of crs , as opposed to the spherical distribution found in the case of isotropic diffusion . in order to fit the gamma
ray data from the w28 region within this scenario , one has to assume that the molecular clouds in its vicinity are magnetically connected to the snr through a magnetic field flux tube . if this is the case , an accurate fit to data can be obtained . under this assumption ,
particles are bound to the flux tube and thus forced to propagate along a specific direction . for plausible values of the diffusion coefficient of magnetic field lines , in order to obtain the correct cr over density at the location of the molecular clouds a large ( parallel ) diffusion coefficient of the order of @xmath182@xmath14/s has to be adopted .
the fact that a very good agreement has been found with data in the two radically different scenarios characterized by isotropic and anisotropic diffusion tells us that more data needs to be collected from more snrs in order to infer with reasonable confidence the properties of the diffusion of particles escaping their accelerators .
the diffuse emission that these runaway particles would produce in their interaction with the ambient gas is , even in the absence of very massive clouds , within the capabilities of the cherenkov telescope array @xcite .
these observations will provide us with precious informations about the properties of the transport of crs in the galaxy , but also with a direct evidence for the fact that snrs are indeed the accelerators of galactic crs .
we thank f. casse , a. marcowith , f. piazza , v. ptuskin , r. schlickeiser , and l. sironi for helpful discussions .
the work of ln and sg has been supported by anr under the jcjc programme .
gabici , s. , casanova , s. , aharonan , f. a. , rowell , g. , sf2a2010 : proceedings of the annual meeting of the french society for astronomy and astrophysics .
s. boissier , m. heydari - malayeri , r. samadi and d. valls - gabaud , p.313 arxiv:1009.5291 | cosmic rays that escape their acceleration site interact with the ambient medium and produce gamma rays as the result of inelastic proton proton collisions .
the detection of such diffuse emission may reveal the presence of an accelerator of cosmic rays , and also constrain the cosmic ray diffusion coefficient in its vicinity .
preliminary results in this direction have been obtained in the last years from studies of the gamma
ray emission from molecular clouds located in the vicinity of supernova remnants , which are the prime candidate for cosmic ray production .
hints have been found for a significant suppression of the diffusion coefficient with respect to the average one in the galaxy .
however , most of these studies rely on the assumption of isotropic diffusion , which may not be very well justified . here
, we extend this study to the case in which cosmic rays that escape an accelerator diffuse preferentially along the magnetic field lines . as a first approximation
, we further assume that particles are strongly magnetized and that their transport perpendicular to the magnetic field is mainly due to the wandering of the field lines .
the resulting spatial distribution of runaway cosmic rays around the accelerator is , in this case , strongly anisotropic .
an application of the model to the case of the supernova remnant w28 demonstrates how the estimates of the diffusion coefficient from gamma
ray observations strongly depend on the assumptions made on the isotropy ( or anisotropy ) of diffusion . for higher levels of anisotropy of the diffusion ,
larger values of the diffusion coefficient are found to provide a good fit to data .
thus , detailed models for the propagation of cosmic rays are needed in order to interpret in a correct way the gamma ray observations .
[ firstpage ] cosmic rays gamma rays
ism : supernova remnants . |
measurements of the peculiar motions of galaxies in the nearby universe represent one of the most powerful tools available to probe mass fluctuations on scales @xmath9 mpc , complementing current efforts to measure the amplitude of these fluctuations from one - degree scale observations of the cosmic microwave background and on very large scales as obtained from the analysis of cobe data ( grski 1996 ) .
analysis of the peculiar velocity field provides not only a direct test of the gravitational instability picture but it also enables the relationship between galaxies and the underlying mass distribution to be investigated .
moreover , to the extent that this relationship can be adequately modeled , for instance by a linear biasing model , comparison between peculiar velocity and redshift data can be used to determine the parameter @xmath10 , where @xmath11 and @xmath12 are the density and linear biasing parameters , respectively ( dekel 1994 ; strauss & willick 1995 ) . while the importance of cosmic flows to cosmological studies has long been recognized , and a variety of methods for analyzing peculiar velocity data have been developed over the years , the progress of redshift - distance surveys suitable for such studies has been relatively slow .
most of the work in the field relied until recently in the tully - fisher ( tf ) and @xmath13 samples of aaronson ( 1982 ) and lynden - bell ( 1988 , hereafter 7s ) , respectively , complemented by relatively small samples covering particular regions of the sky ( willick 1991 ; courteau 1993 ) .
early analyses led to important results such as the discovery of the great attractor by the 7s and to the rejection of the high - bias cdm model as a viable cosmological model .
however , the sparseness of the samples and the relatively small effective volume they probe also led to results and interpretations that need further confirmation .
these include large derived values of the cosmological density parameter , at variance with other estimates , and the existence of large scale coherent motions , suggestive of excess power at very large scales ( willick 1990 ; mathewson , ford & buchhorn 1992 , hereafter mfb ; courteau 1993 ) .
recently , the observational situation has significantly improved with the completion of large tf surveys in both hemispheres ( mfb ; mathewson & ford 1996 ; giovanelli 1997 a , b ; haynes 1999 a , b ) .
combined , they provide the largest all - sky sample currently available for peculiar velocity studies greatly extending the depth and the number of galaxies of earlier surveys .
there still are , however , points of concern .
first , these surveys were carried out independently with different selection criteria , different observing and reduction techniques and have been assembled in different ways to produce all - sky catalogs suitable for analysis such as the sfi ( da costa 1996 ; giovanelli 1998 ) and mark iii ( willick 1997 ) catalogs .
these catalogs differ both in the data sets used to assemble them and in the transformations employed to map the mfb data in the southern hemisphere onto a common system .
analysis of these catalogs have led to both consistent ( zaroubi 1996 ; freudling 1999 ) and conflicting results ( nusser , davis & willick 1997 ; da costa 1998a ; borgani 1999 ; willick & strauss 1998 ) depending on the type of analysis considered .
second , most redshift - distance surveys of the whole sky currently available , including the recently completed shellflow survey ( courteau 1999 ) , rely primarily on tf distances . while other recent investigations that have used early - type galaxies , have either focused on motions of clusters of galaxies ( efar , wegner 1996 , colless 1999 ; smac , hudson 1999 ; lp10k , willick 1999 ) or have concentrated on pre - chosen areas of the sky which do not render them optimal for all types of analyses ( mller 1998 ; mller , wegner & freudling 1999 ) .
finally , recent attempts to use more accurate distance indicators , such as the sbf method ( tonry , blakeslee , & dressler 1999 ) and sneia ( riess 1999 ) , remain limited in depth and in the number of objects with measured distances to be used for a detailed mapping of the peculiar velocity field ( but see blakeslee 1999 ) . in order to address some of these concerns
we have carried out the so - called enear survey , an extensive spectroscopic and @xmath2band imaging survey of a sample of nearby early - type galaxies ( thus the choice of the name to contrast with the efar project ) brighter than @xmath1 and @xmath0(hereafter enearm ) , extracted from complete redshift surveys , and of galaxies in 28 selected clusters ( enearc ) . currently , the enearm and enearc samples comprise 1359 and 569 galaxies with measured distances , respectively . more importantly , 82% of the enearm galaxies and 58% of the enearc galaxies have distance estimates based on new measurements obtained by our group in both hemispheres , making it the largest homogeneous all - sky survey of its kind conducted by a single group .
the primary goal of this survey has been to extend the 7s sample to fainter magnitudes , thereby probing a volume comparable to that of existing tf surveys such as the sfi .
with such a sample it is possible to : 1 ) investigate the reproducibility of the results obtained from the analysis of peculiar velocity data as measured by different secondary distance indicators ; 2 ) combine it with the tf surveys to produce a homogeneous all - sky catalog of galaxies of different morphological types , sampling both low and high density regions ; and 3 ) provide a denser sampling of the gravitational field .
the enear survey began as an outgrowth of spectroscopic surveys carried out in the southern hemisphere started in 1982 ( ssrs+ssrs2 , da costa 1988 , 1998b ) .
although these surveys were primarily designed for redshift measurements , from the start an attempt was made to obtain higher signal - to - noise spectra for bright early - type galaxies for their eventual use in the measurement of central velocity dispersions and spectral line indices .
the original goal was to provide spectral information for a magnitude - limited sample which could be used as a basis for the statistical analysis of the properties of this population as a function of the local galaxy density and different types of environment .
by 1988 , motivated by the results of the 7s analysis and the start of the sfi tf - survey attention shifted to the peculiar velocity field and @xmath14-band imaging was initiated for a sample selected from a combination of magnitude - limited samples with complete redshift information available at the time .
this sample was later extended to lower galactic latitudes as described below . in the meantime ,
several galaxies with only older spectra were re - observed to ensure the quality and uniformity of the sample . in 1993
the campaign to observe northern galaxies began mostly at the mdm observatory .
as a result we have accumulated a large number of spectra and images from which we have measured redshifts , central velocity dispersions , spectral line indices and several photometric parameters . we have assembled this information together with spectroscopic and photometric parameters available in the literature , to produce a homogeneous catalog of early - type galaxies , hereafter enear catalog .
new data continue to be recorded in the catalog as the observations of early - type galaxies are still in progress .
we hope to make our database accessible on - line in the future .
the large number of galaxies in the enear catalog , most of which with new high - quality spectra and ccd images , its homogeneity and the improved membership assignment of galaxies to groups / clusters for a significant fraction of the galaxies make this catalog a valuable tool for many other applications .
the enear catalog , currently the largest uniform database of its kind , is ideally suited for : the statistical characterization of the properties of present - day early - type galaxies ; stellar populations studies using different spectral line indices ; testing the universality of distance relations such as the and the fundamental plane ( fp ) ; assessing the influence of different types of environment and the dependence of the properties of these galaxies on the local density of galaxies ; and constraining models for the formation and evolution of ellipticals ( bernardi 1998 ) .
we also anticipate that the present sample will be used as a benchmark to comparable high - redshift studies .
the enear survey represents a significant improvement over previous redshift - distance surveys of early - type galaxies .
besides the larger number of galaxies , about a factor of three larger than the 7s sample , most of the galaxies have new measurements of velocity dispersion , in general obtained from higher resolution spectra , and photometric parameters derived from the analysis of @xmath14-band images . to ensure uniformity , a large number of observations
have been obtained of galaxies with measurements available from other authors or from different telescope / instrument configurations enabling all available measurements to be scaled to a uniform system . since the survey has been conducted by a single group ,
all of the data were analyzed using the same procedure , further ensuring their uniformity , and the observations were coordinated so as to reach the same level of completeness over the whole sky .
finally , the homogeneous and well - defined sample suitable for peculiar velocity analysis is a subset of magnitude - limited samples for which complete redshift information is available from surveys such as cfa1+cfa2 ( huchra 1983 ; geller & huchra 1989 ; falco 1999 ) , ssrs+ssrs2 and ors ( santiago 1995 ) , thus allowing an objective assignment of galaxies to groups / clusters .
this paper is the first in a series reporting the results of the enear survey .
its primary goal is to give an overview of the project and provide the basic information which will be used in future papers of this series . in section
[ selection ] , we provide details of the procedure adopted in the selection of a well - defined sample for peculiar velocity analysis and of its general characteristics . we also discuss the selection of the cluster sample , which will be used in subsequent papers to construct template distance relations . in section [ data ] , we briefly describe our new spectroscopic and @xmath14-band imaging data .
we also discuss the procedure adopted for the homogenization of the data either newly obtained with different telescopes / instrument configurations or available in the literature .
such a procedure is of paramount importance in the elimination of inconsistencies between data from different sources in order to minimize systematics .
in addition , we summarize the data currently available in the enear catalog that has been assembled in the course of this project . in section [ redshiftdistance ] ,
we briefly discuss the method employed to estimate galaxy distances , the procedure adopted for pruning the sample of galaxies unsuitable for peculiar velocity studies and describe the criteria employed for assigning galaxies to groups / clusters .
this grouping is crucial for reducing the influence of virial motions and decreasing distance - errors .
we also discuss the completeness of our redshift - distance survey , compare our sample with those of previous peculiar velocity studies and give a preview of the measured peculiar velocity field . finally ,
a brief summary is presented in section [ summary ] .
as pointed out above the primary goal of the enear project has been to extend the volume probed by the 7s sample , which considered early - type galaxies brighter than @xmath15 sampling the peculiar velocity field within an effective volume @xmath3 4000 in radius .
an equally important results is that this complements the sfi tf redshift - distance survey of late spirals ( haynes 1999a , b ) out to @xmath16 7000 . to achieve our goal we assembled all the available complete redshift surveys at the beginning of this project ( in 1988 ) to construct an all - sky sample .
originally the following samples were used : 1 ) the cfa1 sample ( huchra 1983 ) , covering the regions @xmath17 and @xmath18 in the northern galactic cap , and @xmath19 and @xmath20 in the southern hemisphere ; 2 ) the sample of galaxies used in the ssrs ( da costa 1988 ) covering the regions @xmath19 and @xmath21 in the southern galactic cap , and @xmath17 and @xmath22 in the northern galactic cap ; 3 ) the equatorial survey sample of huchra ( 1993 ) filling in the gaps between the first two samples near the equator .
the samples for these redshift surveys were extracted from different catalogs which include : the surface photometry catalogue of the eso - uppsala galaxies ( eso , lauberts & valentjin 1989 ) , the morphological catalog of galaxies ( mcg , vorontsov - velyaminov & arhipova 196 ) , the uppsala general catalog ( ugc , nilson 1973 ) and the catalog of galaxies and clusters of galaxies ( cgcg , zwicky 1961 - 1968 ) . magnitudes for the ssrs galaxies were assigned based on magnitude - diameter relations derived by pellegrini ( 1990 ) who made the first attempt to construct a uniform magnitude - limited sample for the whole sky , with additional checks on the relative differences in the magnitudes of the different catalogs being conducted later ( alonso 1993 , 1994 ) . in the southern hemisphere
the enear magnitudes have been compared to the more recent ssrs2 magnitude system ( da costa 1998b ) .
we find that the magnitudes obtained from the magnitude - diameter relation show a linear relation with a scatter of @xmath23 mag .
more recently , galaxies drawn from the ors ( santiago 1995 ) were added in order to extend the sample towards lower galactic latitudes and cover at least part of the great attractor region .
magnitudes used in the different catalogs were converted into an approximately homogeneous system using statistical corrections between the various systems as described in pellegrini ( 1990 ) and da costa ( 1998b ) . however , as discussed in these papers the individual magnitudes can have errors as large as 0.5 mag .
the morphological types adopted in our final catalog are those from lauberts & valentijn ( 1989 ) , and the types given in the other catalogs have been converted into this system . from the experience accumulated in the morphological classification of the ssrs2 ( da costa 1998b ) we know that morphological misclassifications are frequent and in the course of the observations several objects had to be removed both before , based on the examination of the digitized sky survey ( dss , lasker 1990 ) , and after the observations from the examination of the gathered ccd images , as discussed below . from the all - sky sample
we have drawn 1847 galaxies brighter than @xmath24 , with morphological types @xmath25 and @xmath0 .
these criteria define the original enearm sample considered for peculiar velocity studies , but as shown below it includes a significant number of galaxies which , for different reasons , are not suitable for observations or for estimating distances .
the sample comprises 471 galaxies with @xmath26 , 305 galaxies with @xmath27 , and 1071 with @xmath28 , but these numbers should be viewed only as a rough indication of the morphological mix of the catalog , given the uncertainties in the classification .
the redshift distribution of the sample is shown in figure [ fig : cztot ] , where it is compared to that expected for a @xmath1 magnitude - limited sample and to the redshift distribution of galaxies in the 7s sample .
the predicted distribution was obtained using the luminosity function parameters derived by marzke ( 1998 ) for early - type galaxies , normalized to the area of the sky covered by the enearm sample .
the redshift cutoff was adopted to ensure that the sample depth was comparable to that of the sfi , and as a trade - off between reaching a high level of completeness in the nearby volume and the available observing time .
the enearm sub - sample corresponds to @xmath29 of the total number of early - types in a @xmath1 magnitude - limited sample .
the projected distribution of the enearm sample is shown in figure [ fig : enear ] .
comparison of this figure with that shown , for instance , in santiago ( 1995 ) for an all - sky sample to about the same depth shows that the enearm densely probes the most prominent large - scale structures in the nearby universe including virgo , the great attractor region and its extension to the telescopium - pavo - indus supercluster , and the perseus - pisces complex .
also note the presence of areas devoid of galaxies which correspond to high - extinction regions as shown by santiago ( 1995 ) .
the enearm sample is complemented by a cluster sample , hereafter enearc , which has been used to define template distance relations ( bernardi 1999b ) combining all available cluster data .
the sample consists of groups identified in enearm ( see section [ group ] ) with more than 10 members and at least 5 early - type galaxies to which we have added clusters selected from the literature .
the latter were not identified as groups because either most of the member galaxies are fainter than the enearm limiting magnitude , have redshifts beyond its redshift limit , or are outside the surveyed region at low galactic latitudes .
in addition , to the clusters identified as groups , fainter members have been added according to a membership criteria similar to that described below ( see section [ group ] ) and discussed in more detail by bernardi ( 1999a ) .
a total of 37 groups were identified satisfying our cluster definition , nearly all associated with abell / aco clusters .
the enearc sample currently comprises a total of 28 clusters , 22 of these identified in the volume probed by enearm . for most of these clusters a large number of modern observations are available in the literature .
this provides an adequate sample to derive distance relations . instead of observing new systems preference
was given to the observation of as many galaxies as possible already with available data from different sources , primarily from jrgensen ( 1992 , 1995 a , b ) and 7s .
this was done to enable us : to bring all of the available data , both in clusters and in the field , to a common system ; to substitute old measurements of central velocity dispersions and of photometric parameters , whenever required ; to obtain better resolution and higher @xmath30 spectra for galaxies in clusters for comparison with similar data for galaxies in low density regions ; to enlarge the sample of galaxies with the parameters required for the derivation of the fp ; to enlarge the sample of galaxies in clusters with measured distances to improve the statistics of the template relation .
figure [ fig : czclust ] shows the redshift distribution of the 28 clusters presented in bernardi ( 1999a ) which span a range of redshifts up to @xmath31 with an approximately uniform number per redshift bin . also shown is the redshift distribution of all the cluster members identified using the membership assignment described below .
the spatial distribution of these clusters is shown in figure [ fig : sgcoor ] where three orthogonal projections in supergalactic cartesian coordinates are displayed .
as already stated , the primary objective of the present redshift - distance survey has been to produce an all - sky homogeneous sample for peculiar velocity analysis . with a few exceptions , most recent work in the field
has had to rely on the combination of samples from diverse sources , using different selection criteria , observing strategies and reduction procedures .
this has made the homogenization of the measurements from different sources very difficult and their errors poorly understood . by contrast , even though measurements from different authors are used in compiling our final homogeneous redshift - distance catalog , about 77% and 65% of early - type galaxies with available photometric and spectroscopic data , respectively , are from new observations obtained as part of this project over the whole sky .
this together with the large number of galaxies with multiple observations from different telescope / instrument configurations and in common with other authors enables the internal and external homogenization of the whole data set , the measurement of parameters with an improved accuracy and the possibility of cross - checks between different data sets .
while the imaging observations described here are limited to the enear samples ( enearm+enearc ) presented above , spectra have been obtained for galaxies belonging to the enearm and enearc samples as well as for other early - type galaxies observed as part of several spectroscopic surveys conducted in the southern hemisphere ( ssrs , ssrs2 ) .
measurements of the central velocity dispersion and of spectral line indices for galaxies with spectra deemed suitable for these calculations have also been included in our enear catalog presented below to further enlarge the database of spectroscopic parameters characterizing present - day early - type galaxies for studies other than peculiar velocity analysis . in later papers in this series
we will present these additional measurements together with those obtained from the enear survey proper ( wegner 2000 ; rit 2000 ) .
our spectroscopic observations have been carried out over a long period of time using a variety of telescopes and instruments ( casleo , ctio , eso , lna and mdm ) , detectors and gratings , with the spectral resolution ranging from 2 to 5 .
some of the data originating as far back as the mid-80s were obtained using the intensified photon - counting reticon detector utilized in the southern sky redshift survey ( ssrs , da costa 1991 ) .
the observing and data reduction procedures are similar to those used in the efar survey ( wegner 1999 ) . to date a total of @xmath32 spectra of early - type galaxies taken from the enearm and enearc sample
have been obtained .
however , some of these spectra have not yielded reliable measurements of the spectroscopic parameters of interest ( velocity dispersion ; spectral lines indices ) .
these cases include strong emission - line galaxies with weak absorption features , low surface brightness galaxies yielding low @xmath30 spectra and galaxies too close to bright stars .
others , especially the older reticon data , lead to measurements with large errors compared to our most recent observations and most have been or are in the process of being re - observed . of the total number of available spectra ,
1745 were suitable to measure the central velocity dispersion of 1210 galaxies , including 535 multiple spectra of 327 galaxies .
the number of repeated observations range from two to more than 10 for a few comparison galaxies .
the multiple observations were used to compare spectra obtained with different telescope / instrument configurations and have been used to make our measurements internally consistent .
they have also been used to calibrate our internal error estimates ( wegner 2000 ) .
in addition , we have observed about 200 galaxies with measurements available in the literature to derive statistical corrections and bring those data into a uniform system ( bernardi 1999a ) . by analyzing these spectra
we have measured redshifts , velocity dispersions and line indices ( wegner 2000 ) and figure [ fig : histspect ] shows the distribution of these quantities for all enear spectra deemed suitable for these measurements . also shown
are the same distributions as measured from spectra obtained using a resolution of @xmath162.5 .
over 80% of the spectra were obtained at high - resolution allowing us to make more accurate measurements , especially for galaxies with @xmath33 kms@xmath34 .
the distribution of the differences between our redshifts and those available in the literature is shown in figure [ fig : histcompcz ] .
in general the agreement is excellent leading to an rms of the differences @xmath330 .
note , however , the presence of a few outliers which should be taken into account when computing peculiar velocities .
since our measurements have been obtained from high @xmath30 in computing peculiar velocities we give preferences to these new measurements .
moreover , from our data we find 86 emission - line galaxies , out of which about 60 show no evidence of being misclassified spirals .
some of these have very strong emission lines and weak absorption features which may indicate the presence of residual star - formation .
a detailed description of the observations , data reduction , and derived spectroscopic parameters is given in wegner ( 2000 ) .
finally , we point out that in addition to the line index used by bernardi ( 1998 ) , several other spectral indices are being computed and will be presented in a future paper ( rit 2000 ) .
our spectroscopic parameters have been converted onto a uniform system using the many repeated observations we have in different runs in both hemispheres .
our fiducial system is defined by our high - resolution spectra ( @xmath35 ) .
in addition , we used measurements of 200 galaxies in common with other sources to derive corrections to scale the data in the public domain to our system . from these external comparisons
we find that our errors are typically 8% in @xmath36 and 0.01 mag in , consistent with our internal estimates ( wegner 2000 ) .
since 1988 , photometric observations in @xmath14-band have been made using different telescopes in the northern ( flwo , mdm ) and southern hemispheres ( ctio , eso ) .
about 2200 images have been obtained of which 1896 were taken in photometric conditions . from these images we have measured photometric parameters for 1398 galaxies .
most of the data were taken with large format ccds , allowing for good sky subtraction and , occasionally , for observing more than one galaxy per frame , especially for galaxies in groups / clusters .
some galaxies were later discarded because they were unsuitable for measurements of the photometric parameter of interest ( stellar contamination ; crowded fields ; interacting galaxies or superposed galaxy images ; misclassification ) .
because the imaging observations were conducted over an extended period of time , an attempt was made to obtain repeated observations in every run to provide the necessary data for run - to - run corrections .
such observations have been used to transform all of our measurements into a common internal system .
all images were reduced using standard _ iraf _ tools , and surface photometry was carried out by fitting elliptical and circular aperture to the two - dimensional light distribution of each galaxy .
light profiles were fitted either by a pure de vaucouleurs r@xmath37 law or in combination with an exponential profile to represent the light distribution of bulge+disk systems .
the fits used the algorithm which was developed in the efar survey ( saglia 1997 ) .
this also corrects for the smearing effects due to seeing and provides information about the quality of the fit .
the main photometric parameters derived are @xmath38 , the angular diameter within which the mean surface brightness is equal to @xmath39 mag arcsec@xmath40 , the half - light radius , @xmath41 , the mean surface brightness within this radius , , as well as global quantities such as total magnitudes , @xmath42 , the disk - to - bulge ratio , @xmath43 , ellipticity , position angle and parameters which can be used to characterize the shape of images .
figure [ fig : histphot ] shows the distribution of @xmath44 , @xmath45 , and the disk - to - bulge ratio @xmath43 for all observed galaxies for which these parameters could be measured .
the first three parameters are used in defining the distance relations , while @xmath43 has been used as an indicator of possible misclassifications or inadequate light profile fitting .
details of the observations , data reduction , analysis and tables listing the photometric parameters will be presented by alonso ( 2000 a , b ) .
as mentioned above our derived values of photometric parameters ( @xmath46 , @xmath47 , ) were transformed into an internally consistent system using repeated observations . in total ,
498 observations of 358 galaxies are available for assessing these statistical corrections . by careful examination of all multiple observations available ,
internal conversion relations were derived for each photometric parameter of interest and were used to create a uniform internal system .
in addition , we observed 346 galaxies in common with other sources , including 100 galaxies in the 7s sample for which no images were previously available , to derive corrections to scale the data in the public domain to our system . from the comparison between our measurements and those of other authors we estimate that the typical errors in @xmath44 , @xmath45 , and are 0.017 dex , 0.08 dex , and 0.3 , respectively .
these values are consistent with our internal estimates .
finally , it is also worth mentioning that some @xmath3150 galaxies have also been observed in the @xmath48-band and @xmath3200 in the @xmath49 and @xmath50 near - ir bands .
these data will be presented in forthcoming papers . in the course of this work we have assembled all of the spectroscopic and photometric data gathered by the enear survey , the spectroscopic data from other ongoing observations in the southern hemisphere and most of the available data in the literature into a uniform catalog of early - type galaxies , which we will refer to as the enear catalog .
the main goal of producing this catalog was to minimize the number of galaxies that required observations , and to produce a homogeneous database for other applications . even though extensive , the compilation from the literature is not complete as preference was given to data in the redshift range of interest .
table [ tab : ref ] summarizes the contribution from different sources included in the present catalog listing : in column ( 1 ) the source ; in column ( 2 ) the number of galaxies with measured central velocity dispersion ; in column ( 3 ) number of galaxies with measured line index ; in column ( 4 ) the number of galaxies with measured @xmath38 ; and in column ( 5 ) number of galaxies with available fp parameters . also included are the older compilations of central velocity dispersion measurements carried out by tonry & davis ( 1981 ) and whitmore ( 1985 ) , even though , in general , these data are not used in combination with newer data .
we also give the number of early - type galaxies measured by our group in the southern hemisphere which are not part of the enear survey , as described in this paper , but have been already analyzed by rit ( 1999 ) .
we refer to this additional sample as enear+ . from the literature ,
the bulk of the data still comes from the all - sky sample of 7s galaxies ( faber 1989 and as distributed by d. burstein in electronic form , the so - called mark ii catalog ) .
a revised version of the spectroscopic parameters has recently been kindly provided by d. burstein . with the exception of this sample , most other available data sets are primarily in clusters of galaxies .
it is also important to emphasize that the data as a whole are very inhomogeneous .
the spectroscopic parameters were derived from spectra obtained with very different spectrographs ( photon - counting detectors , using small slits ; ccd detectors , using long - slits ) .
the original spectra also have a wide range of resolutions and spectral coverage .
similarly , the photometric data include old photo - electric photometry and images using small and large format ccds , and were obtained using different filters .
therefore , the measured parameters can not be used as listed in the original sources but must be first transformed into a common scale , utilizing the conversion relations derived from overlapping data as described above . in our final catalogs whenever
data from other authors are used they have been corrected as described in bernardi ( 1999a ) .
this procedure has been helpful in maximizing the use of publicly available data and has enabled us to reach a higher completeness of the enear samples
. currently , in our database there are @xmath51 measurements of velocity dispersion , 3700 values of the @xmath38 parameter and 2304 values of the combination @xmath52 - 0.3 , as used in the fp relation .
these yield 2797 galaxies with measured central velocity dispersion , 1960 galaxies with @xmath38 , 1588 galaxies with @xmath53 , 1847 and 1458 galaxies with the required information to compute their distances using the and fp relations , respectively . from these data ,
distances to 76% of the enearm galaxies and 89% of the enearc sample can be estimated . in our final catalog ,
galaxies with more than one measurement of @xmath38 , redshift and velocity dispersion were combined , discarding measurements from compilations , using an error - weighted mean of the different measurements , eliminating outliers whenever possible .
measurement of the radial component of galaxy s peculiar velocity , @xmath54 , requires knowing its redshift @xmath55 and its redshift independent distance @xmath14 ( in units of ) , which must be determined using a secondary distance indicator .
distances to early - type galaxies may be inferred from relations such as and fp , which relate different observable quantities that depend on the distance of the galaxy ( @xmath38 , @xmath53 ) to a redshift independent quantity such as the central velocity dispersion . for studies of the peculiar velocity field
it is common practice to derive these relations for galaxies in clusters , presumed to be at the same distance . to improve the statistics data from different clusters are combined to derive a template relation ( giovanelli 1997b , bernardi 1999b ) . for the purpose of deriving and fp template relations
we have compiled the enearc sample described above .
it comprises 577 galaxies with @xmath38 , 473 with @xmath53 parameters , and 586 with @xmath36 , yielding a sample of 569 and 431 cluster galaxies to derive a and fp relations , respectively . of these
we have obtained new @xmath2band photometric parameters for 315 galaxies , high - quality spectra for 235 , and combined with other available data we have added 153 new distance measurements , which 93 had no prior measurements . until recently most studies of peculiar velocity using early - type galaxies relied on distances , with the fp being used primarily for more detailed stellar population studies and of possible environment effects .
while it is our intention to compute both distance relations and investigate possible effects on the results of the analysis of the peculiar velocity field , our preliminary analysis of the data is based on the relation derived by bernardi ( 1999b ) , postponing to a future paper the derivation of the fp template relation ( alonso 2000c ) .
this choice was motivated by the larger number of measurements of @xmath38 available , the fact that this quantity is less sensitive to seeing effects , and the parameterization used in fitting the light profiles , as required for determining the half - light radius @xmath41 and the mean surface brightness that enter in the fp relation .
moreover , recent studies have shown that the scatter about this relation is comparable to that obtained using the fp relation ( jrgensen , franx & kjrgaard 1996 ; donofrio 1997 ) . as described in section [ selection ] , the enearm sample is defined from a combination of catalogs , therefore , it is not surprising that some pruning of the sample is required both before and after the observations . in general , galaxies were inspected before the observations , especially after the availability of the dss . at this point
some galaxies were immediately removed because of one of the following reasons : stellar contamination ; crowded fields preventing a good determination of the background ; interacting galaxies ; contamination / superposition by neighboring galaxies of comparable size and/or brightness ; and morphological misclassifications , especially late - type spirals . we have also discarded galaxies during the analysis of ccd images whenever they revealed features not discernible in the dss due to saturation in the inner parts of the images .
a common cause is the presence of relatively bright stars superposed on the central parts of the galaxy image , making the reliable determination of photometric parameters impossible .
a complete description of these cases will be presented in alonso ( 2000a ) .
we have also removed @xmath3 100 galaxies from the sample after analyzing their spectra .
these cases include galaxies with strong emission - lines but very weak absorption features , galaxies with small velocity dispersions ( depending on the spectral resolution used ) and low surface brightness galaxies with low @xmath30 spectra ( @xmath56 ) ( see wegner 2000 ) . in total , roughly 250 galaxies were discarded for these reasons reducing the enearm sample from 1847 to 1607 .
this culled sample is used below as a reference in estimating the completeness of the enear survey .
finally , when compiling the peculiar velocity catalog , galaxies with extremely large peculiar velocities were re - examined and those with old photometric or spectroscopic measurements or with some peculiarity in their images ( prominent bars , shells , spiral arms , dust lane ) that could either affect the derivation of the photometric parameters of interest or lead to a significant departure of the galaxy from the distance relation were removed ( @xmath57 galaxies ) from the sample used for peculiar velocity analysis ( alonso 2000d ) .
in contrast to late - type galaxies , early - type objects tend to reside in virialized clumps and in regions of high density .
therefore , in order to use early - type galaxies to map the peculiar velocity field it is important to assign them to groups and clusters , in order to minimize the impact of virial motions on the analysis of the velocity field .
one advantage of grouping galaxies is that the distances to grouped objects are more accurately determined , as the fractional distance errors are reduced by a factor of @xmath58 , where @xmath59 is the number of early - type galaxies in the group .
since our sample has been drawn from complete redshift surveys , membership assignment of galaxies to groups can be done in a systematic way , using objective friends - of - friends algorithms to search for galaxy groups in the original magnitude - limited catalogs . to this end
we have used the catalogs of groups of maia ( 1989 ) for the ssrs and of geller & huchra ( 1983 ) for the cfa1 .
these groups were defined as associations of galaxies with a density contrast @xmath60 .
these catalogs have been used as they provide the largest sky coverage . in regions covered by cfa2 and ssrs2
we have replaced these groups by those recently compiled by ramella ( 1997 ) for the cfa2 and by ramella ( 1999 ) for the ssrs2 .
these surveys cover a slightly smaller area but extend to fainter magnitudes and consider groups with larger density contrast @xmath61 which , as demonstrated by numerical simulations , are less contaminated by spurious groups ( members ) . finally , groups have also been identified at low galactic latitudes using the ors data kindly provided by b. santiago and m. davis . using the position and parameters of these groups , early - type galaxies
were assigned to a group if their projected separation relative to center of a group was @xmath62 and their redshift satisfied the condition ( @xmath63 . here , @xmath64 is the mean projected separation of the group ( ramella 1989 ) , @xmath65 is the radial velocity of the early - type galaxy , @xmath66 is the mean group velocity and @xmath67 is the velocity dispersion of galaxies in the group .
we adopted these criteria in an attempt to homogenize the assignment of galaxies to groups , given the differences in density contrast used in their original identifications , and to allow for the identification of systems which may have been missed near the edges of the catalogs . adopting the above criteria we find that 59% of the early - type galaxies in enearm are in groups .
we point out that 15% of the galaxies are associated with groups in which the other members are late - type galaxies .
such systems represent about 10% of the total number of groups . in the latter case the peculiar velocity of the galaxy
is computed using the galaxy s distance , while the redshift is that of the group .
groups containing two or more early - type galaxies are treated as single objects with the redshift being given by the mean redshift of the group , including all morphological types , and the logarithm of the distance as the error - weighted mean of the logarithm of the distance of the early - type galaxies in that group .
figure [ fig : multiplicity ] shows the number of groups as a function of the number of early - type galaxies in the group .
we note that in the case of clusters within @xmath0 used to determine the distance relations ( see section [ clusters ] ) we have also used fainter galaxies in the calculation of their redshift and distance .
as discussed above ( see also bernardi 1999a ) , we arbitrarily define clusters to be systems with more than 10 members , of which at least 5 are early - type galaxies .
this cluster sample is used by bernardi ( 1999b ) and alonso ( 2000c ) to define the template distance relation later used to estimate galaxy distances .
finally , we point out that the membership assignment criteria described above was adopted by bernardi ( 1998 ) to investigate the properties of early - type galaxies in different environments .
figure [ fig : enearobj ] shows the projected distribution of the 1141 `` objects '' in the enearm sample .
the upper panel shows the 748 galaxies not assigned to groups , while the lower panel shows 393 groups / clusters .
as expected , the `` isolated '' galaxies are more uniformly distributed across the sky , while the group / cluster sample delineates more clearly the most prominent large - scale structures in the nearby universe .
since our parent sample has complete redshift information , completeness here refers to galaxies for which we have both velocity dispersion and @xmath38 measurements and thus distances .
figure [ fig : complet ] shows the completeness function of the enearm redshift - distance survey as a function of : ( a ) redshift , ( b ) blue magnitude @xmath68 , ( c ) galactic longitude , and ( d ) galactic latitude .
the completeness is computed relative to the sub - sample of 1607 galaxies that remain after excising galaxies , from the original sample of 1847 galaxies , unsuitable for distance estimates , as discussed above . examining the figure we find no strong dependence with the redshift , magnitude , or galactic longitude and latitude except near the southern galactic pole ( @xmath69 ) . in general , the sample completeness is a nearly constant @xmath70% .
this demonstrates that our dataset , even though not complete is remarkably uniform both in depth and in sky coverage .
this is in marked contrast to other catalogs assembled from different sources ( mark iii ) which show large sampling variations in different directions of the sky .
we have not investigated the completeness of the sample as function of morphological types since analysis of our images shows that the morphological classification available from the original catalogs is unreliable and care should be exercised in defining sub - types of the early - type population for any detailed analysis of their properties .
we note that the information presented here is useful for producing realistic mock samples for analysis of the peculiar velocity field .
to underscore the uniformity of the observed sample figure [ fig : enearfdist ] compares the projected distribution , in galactic coordinates , of all galaxies of the enearm sample that have measured distances ( filled circles ) with galaxies for which distances are still unavailable ( open circles ) .
clearly , the sample with measured distances delineates well all major structures in the nearby universe .
we also find that the distribution of missing galaxies does not reveal any particularly under - sampled region .
so far , we have shown all galaxies individually . however , as discussed earlier , early - type galaxies are found predominantly in clusters / groups .
therefore , figure [ fig : eneardistobj ] shows the distribution of independent objects defined by the grouping procedure described in section [ group ] ( and in more detailed in bernardi 1999a ) .
currently , we have distances for 1644 galaxies , 1359 in enearm . the remaining are galaxies in the enearc cluster sample but not in the enearm .
we include these galaxies because they are actually used in the definition of the redshift , distance and error estimate for these clusters .
figure [ fig : eneardistobj ] shows that there are 811 independent objects of which 6 are clusters beyond 7000 .
as briefly mentioned in section [ pruning ] about 200 galaxies show , after closer inspection of cases with extreme values of the peculiar velocity , features in their spectra and/or in their images that can affect the measured parameters that enter the distance relation and as a precaution were removed from the sample .
a detailed description and listing of these cases will be presented by alonso ( 2000d ) . at the present time
the sample being used in our preliminary analysis ( da costa 2000 ; zaroubi 2000 ) consists of 1430 galaxies grouped in 702 objects .
the enearm sample shown in figure [ fig : enearfdist ] is the largest and most homogeneous sample of nearby early - type galaxies currently available for cosmic flow studies . to highlight this ,
figure [ fig : enearmark2 ] compares the projected distributions , in redshift slices , of the enearm and the 7s samples after galaxies are grouped . besides the obvious differences in the total number of galaxies and objects ,
there is a striking difference in the structures probed by the two samples . in particular , in the most distant redshift shell , perseus - pisces is clearly visible in our catalog while is conspicuously absent in the 7s sample .
this accounts for some of the surprises encountered in earlier ( 7s - based ) reconstructions ( dekel 1990 ) .
it is also worth comparing our sample with catalogs built from recently completed tf surveys .
figure [ fig : enearmark2sfi ] compares the projected distribution of individual galaxies having measured peculiar velocities in enearm and the sfi tf - survey of spiral galaxies . from the figure
one can see the complementarity of these samples .
the enearm early - type galaxies delineate the structures more sharply while the spirals are clearly more uniformly distributed .
for this reason , the combination of the two samples is highly desirable , and it is the subject of future work . finally , as a preview for future papers in this series we present in figure [ fig : sgslice ] the homogeneous malmquist bias corrected peculiar velocities of the enearm galaxies within @xmath71 from the supergalactic plane in the cmb rest frame . for comparison
we show the map of the flowfield as probed by the 7s .
the most remarkable features are the suggestion of a back - side infall in the direction of the great attractor region and the flow associated with the perseus - pisces complex nearly absent in the 7s sample .
a detailed analysis of the velocity field will be presented in subsequent papers of this series .
we have presented an overview of the ongoing imaging / spectroscopic survey of nearby early - types being conducted in both hemispheres which allowed the construction of an extensive and homogeneous database consisting of redshifts , central velocity dispersion measurements , spectral line indices and photometric data for roughly 3400 early - type galaxies , with multiple measurements in the nearby universe ( @xmath72 ) . since these galaxies are part of complete redshift surveys additional information is also available regarding the type of environment and measures of the local galaxy density , making the catalog useful for different applications .
we have also discussed in some detail how a magnitude and redshift - limited sample of galaxies with measured distances and redshifts suitable for studies of the local peculiar velocity field has been constructed .
the resulting enearm sample together with the cluster galaxy sample enearc , constitute the basis of the work being carried out by our group to map the peculiar velocity field as probed by early - type galaxies .
we hope that the comparison of the peculiar velocity fields derived from the enear survey and similar tf - based surveys will provide a more definite answer about the true physical nature of the observed motions , thus supporting the assumption that the peculiar motions observed are in fact induced by the gravity field associated to fluctuations of the underlying mass density field
. quantitative analysis of various aspects of the measured velocity field such as the calculation of the dipole , the mass power - spectrum , and determinations of @xmath73 will be presented in subsequent papers of this series . in the future
, we also hope to be able to combine the enear and sfi samples to obtain improved maps of the underlying mass distribution .
finally , we also anticipate the use of our dataset for more detailed stellar population studies .
aaronson , m. , huchra , j. , mould , j. , schechter , p. l. , & tully , r. b. 1982 , apj , 258 , 64 alonso , m. v. , da costa , l. n. , pellegrini , p. s. & kurtz , m. j. , 1993 , aj , 106 , 676 alonso , m. v. , da costa , l. n. , latham , d. w. , pellegrini , p. s. , milone , a. a. e , 1994 , aj , 108 , 1987 alonso , m. v. , bernardi , m. , wegner , g. 2000a , in preparation alonso , m. v. , bernardi , m. , wegner , g.2000b , in preparation alonso , m. v. , bernardi , m. , da costa , l. 2000c , in preparation alonso , m. v. , bernardi , m. , da costa , l.2000d , in preparation bernardi , m. , renzini , a. , da costa , l. n. , wegner , g. , alonso , m. v. , pellegrini , p. s. , rit , c. n. a. , & willmer , c. n. a. 1998 , apj , 508 , l143 bernardi , m. 1999a , in preparation bernardi , m. 1999b , in preparation blakeslee , j. p. , davis , m. , tonry , j. l. , ajhar , e. a. , & dressler , a. 1999 , in cosmic flows : towards and understanding of large - scale structure , eds . s. courteau , m. a. strauss , & j. a. willick , asp conf .
ser . , astro - ph/9910342 borgani , s. , da costa , l. n. , zehavi , i. , giovanelli , r. , haynes , m. p. , freudling , w. , wegner , g. , & salzer , j. j. 1999 , astro - ph/9908155 colless , m. , burstein , d. , davies , r. l. , mcmahan , r. k. j. , saglia , r. p. , & wegner , g. 1999 , mnras , 303 , 813 courteau , s. , faber , s. m. , dressler , a. , & willick , j. a. 1993 , apj , 412 , l51 courteau , s. , willick , j. a. , strauss , m. a. , schlegel , d. , & postman , m. 1999 , in cosmic flows :
towards and understanding of large - scale structure , eds . s. courteau , m. a. strauss , & j. a. willick , asp conf .
astro - ph/9909385 da costa , l. n. , freudling , w. , wegner , g. , giovanelli , r. , haynes , m. p. , & salzer , j. j. 1996 , apj , 468 , l5 da costa , l. n. , nusser , a. , freudling , w. , wegner , g. , giovanelli , r. , haynes , m. p. , salzer , j. j. , & wegner , g. 1998a , mnras , 299 , 425 da costa , l. n. , pellegrini , p. s. , davis , m. , meiksin , a. , sargent , w. l. w. , & tonry , j. l. 1991 , apjs , 75 , 935 da costa , l. n. , pellegrini , p. s. , sargent , w. l. w. , tonry , j. , davis , m. , meiksin , a. , latham , d. w. , menzies , j. w. , & coulson , i. a. 1988 , apj , 327 , 544 da costa , l. n. , willmer , c. n. a. , pellegrini , p. s. , chaves , o. l. , rit , c. , maia , m. a. g. , geller , m. j. , latham , d. w. , kurtz , m. j. , huchra , j. p. , ramella , m. , fairall , a. p. , smith , c. , & lpari , s. 1998b , aj , 116 , 1 dekel , a. 1994 , araa , 32 , 371 dekel , a. , bertschinger , e. , & faber , s. m. 1990 , apj , 364 , 349 donofrio , m. , capaccioli , m. , zaggia , s. r. , & caon , n. 1997 , mnras , 289 , 847 dressler , a. 1987 , apj , 317 , 1 faber , s. m. , wegner , g. , burstein , d. , davies , r. l. , dressler , a. , lynden - bell , d. , & terlevich , r. j. 1989 , apjs , 69 , 763 falco 1999 , pasp , 111 , 438 freudling , w. , zehavi , i. , da costa , l. n. , dekel , a. , eldar , a. , giovanelli , r. , haynes , m. p. , salzer , j. j. , wegner , g. , & zaroubi , s. 1999 , apj , 523 , 1 geller , m. j. , & huchra , j. p. 1983 , apjs , 52 , 61 geller , m. j. , & huchra , j. p. 1989 , in large scale structure and motions in the universe , ed .
m. mezzetti ( dordrecht : kluwer academic publishers ) , p. 3
giovanelli , r. , & haynes , m. p. , freudling , w. , da costa , l. n. , salzer , j. j. , & wegner , g. 1998 , apj , 505 , l91 giovanelli , r. , haynes , m. p. , herter , t. , vogt , n. p. , da costa , l. n. , freudling , w. , salzer , j. j. and wegner , g. 1997a , aj , 113 , 22 giovanelli , r. , haynes , m. p. , herter , t. , vogt , n. p. , da costa , l. n. , freudling , w. , salzer , j. j. and wegner , g. 1997b , aj , 113 , 53 grski , k. m. , banday , a. j. , bennett , c. l. , hinshaw , g. , kogut , a. , smoot , g. f. , & wright , e. l. 1996 , apj , 464 , l11 guzmn , r. , & lucey , j. r. 1993 , mnras , 263 , 47 haynes , m. p. , giovanelli , r. , salzer , j. j. , wegner , g. , freudling , w. , da costa , l. n. , herter , t. , & vogt , n. p. 1999a
, aj , 117 , 1668 haynes , m. p. , giovanelli , r. , chamaraux , p. , da costa , l. n. , freudling , w. , salzer , j. j. , & wegner , g. 1999b , aj , 117 , 2039 huchra , j. p. , davis , m. , latham , d. , & tonry , j. 1983 , apjs , 52 , 89 huchra , j. p. , latham , d. w. , da costa , l. n. , pellegrini , p. s. , & willmer , c. n. a. 1993 , aj , 105 , 1637 hudson , m. j. , lucey , j. r. , smith , r. j. , & steel , j. 1997 , mnras , 291 , 488 hudson , m. j. , smith , r. j. , lucey , j. r. , schlegel , d. j. , & davies , r. l. 1999 , apj , 512 , l79 jrgensen , i. , franx , m. , & kjrgaard , p. 1992
, a&as , 95 , 489 jrgensen , i. , franx , m. , & kjrgaard , p. 1993 , apj , 411 , 34 jrgensen , i. , franx , m. , & kjrgaard , p. 1995a ,
mnras , 273 , 1097 jrgensen , i. , franx , m. , & kjrgaard , p. 1995b ,
mnras , 276 , 1341 jrgensen , i. , franx , m. , & kjrgaard , p. 1996 ,
mnras , 280 , 167 lasker , b. m. , sturch , c. r. , mclean , b. j. , russell , j. l. , jenker , h. , & shara , m. m. 1990 , aj , 99 , 2019 lauberts , a. 1982 , the eso - uppsala survey of the eso(b ) atlas ( mnchen : eso ) lauberts , a. , valentijn , e. a. , 1989 , the surface photometry catalogue of the eso - uppsala galaxies ( garching : eso ) lynden - bell , d. , faber , s. m. , burstein , d. , davies , r. l. , dressler , a. , terlevich , r. , & wegner , g. 1988 , apj , 326 , 19 maia , m. a. g. , da costa , l. n. , & latham , d. w. 1989 , apjs , 69 , 809 marzke , r. o. , da costa , l. n. , pellegrini , p. s. , willmer , c. n. a. , & geller , m. j. 1998 , apj , 503 , 617 mathewson , d. s. , ford , v. l. , & buchhorn , m. 1992 , apjs , 81 , 413 ( mfb ) mathewson , d. s. , & ford , v. l. 1996 , apjs , 107 , 97 mller , k. , wegner , g. & freudling , w. 1999 , a&a suppl , in press , astro - ph/9910392 mller , k. , freudling , w. , watkins , r , & wegner , g. , 1998 , apjl , 507 , 105 nilson , p. 1973
, uppsala general catalog of galaxies , uppsala astron .
ann . , 6 ( ugc ) nusser , a. , davis , m. , & willick , j. a. 1997 , in galaxy scaling relations : origins , evolution and applications , eds
. l. da costa , & a. renzini ( berlin : springer ) , p. 286
pellegrini , p. s. , da costa , l. n. , willmer , c. n. a. , huchra , j. p. , & latham , d. w. 1990 , aj , 99 , 751 ramella , m. , geller , m. j. , & huchra , j. p. 1989 , apj , 344 , 57 ramella , m. , pisani , a. , & geller , m. j. 1997 , aj , 113 , 483 ramella , m. , 1999 , in preparation rit , c. 1999 , ph.d .
thesis , cnpq / observatrio nacional rit , c. , maia , m. a. g. , willmer , c. n. a. 2000 , in preparation riess , a. g. 1999 , in cosmic flows : towards and understanding of large - scale structure , eds .
s. courteau , m. a. strauss , & j. a. willick , asp conf .
ser . , astro - ph/9908237 saglia , r. p. , bertschinger , e. , baggley , g. , burstein , d. , colles , m. , davies , r. l. , mcmahan , r. k. , & wegner , g. 1997 , apjs 109 , 79 santiago , b. x. , strauss , m. a. , lahav , o. , davis , m. , dressler , a. , & huchra , j. p. 1995 , apj , 446 , 457 strauss , m. a. , & willick , j. a. 1995 , physics reports , 261 , 271 tonry , j. l. , blakeslee , j. p. , ajhar , e. a. , & dressler , a. 1999 , astro - ph/9907062 tonry , j. l. , & davis , m. 1981 , apj , 246 , 666 van albada , t. s. , bertin , g. , & stiavelli , m. 1993 , mnras , 265 , 627 vorontsov - velyaminov , b. a. , arhipova , v. p. , & krasnogorskaja , a. a. 1962 - 1974 , morphological catalog of galaxies vols . 1 - 5 ( moscow : moscow state univ.)(mgc ) wegner , g. , colless , m. , baggley , g. , davies , r. l. , bertschinger , e. , burstein , d. , mcmahan , r. k. , & saglia , r. p. 1996 , apjs , 106 , 1 wegner , g. , colless , m. , saglia , r. p. , mcmahan , r. k. , davies , r. l. , burstein , d. & baggley , g. , 1999 , mnras , 305 , 259 wegner , g. , willmer , c. n. a. , bernardi , m. , 2000 , in preparation whitmore , b. c. , mcelroy , d. b. , & tonry , j. l. 1985 , apjs , 59 , 1 willick , j. a. 1990 , apj , 351 , l5 willick , j. a. 1991 , phd thesis , university of california at berkeley willick , j. a. 1999 , apj , 522 , 647 willick , j. a. , courteau , s. , faber , s. m. , burstein , d. , dekel , a. , & strauss , m. a. 1997 , apjs , 109 , 333 willick , j. a. , & strauss , m. a. 1998 , apj , 507 , 64 zaroubi , s. , zehavi , i. , dekel , a. , hoffman , y. & kolatt , t. 1997 , apj , 486 , 21 zwicky , f. 1961 - 1968 , catalog of galaxies and clusters of galaxies , vols . 1 - 6 ( pasadena : california inst . of technology)(cgcg ) lrrrr + source & n@xmath74 & n@xmath75 & n@xmath76 & n@xmath77 + ( 1 ) & ( 2 ) & ( 3 ) & ( 4 ) & ( 5 ) + enear & 1210 & 1036 & 1398 & 1398 + enear+ & 792 & 792 & & + td ( 1981 ) & 229 & & & + wmt ( 1985 ) & 130 &
& & + lc ( 1988 ) & 114 & & 78 & + 7s ( 1989 ) & 543 & 533 & 499 & 179 + d ( 1991 ) & 158 & 151 & 175 &
+ jfk ( 1992 , 1995 ) & 159 & 119 & 200 & 194 + lc ( 1997 ) & 85 & & 84 & + s ( 1997 ) & 88 & 81 & 98 & 98 + | this is the first in a series of papers describing the recently completed all - sky redshift - distance survey of nearby early - type galaxies ( enear ) carried out for peculiar velocity analysis .
the sample is divided into two parts and consists of 1607 elliptical and lenticular galaxies with @xmath0 and with blue magnitudes brighter than @xmath1 ( enearm ) , and of galaxies in clusters ( enearc ) .
galaxy distances based on the and fundamental plane ( fp ) relations are now available for 1359 and 1107 enearm galaxies , respectively , with roughly 80% based on new data gathered by our group .
the and fp template distance relations are derived by combining 569 and 431 galaxies in 28 clusters , respectively , of which about 60% are based on our new measurements . to date
the enear survey has accumulated : 2200 @xmath2band images yielding photometric parameters for 1398 galaxies ; and 2300 spectra yielding 1745 measurements of central velocity dispersions and spectral line indices for 1210 galaxies .
in addition , there are some 1834 spectra of early - type galaxies available in the ssrs+ssrs2 database out of which roughly 800 galaxies yield high - quality measurements of velocity dispersions and spectral line indices , bringing the total number of galaxies with available spectral information to about 2000 . combined with measurements publicly available
, a catalog has been assembled comprising : @xmath3 4500 measurements of central velocity dispersions for about 2800 galaxies ; @xmath3 3700 measurements of photometric parameters for about 2000 galaxies ; distances for about 1900 galaxies .
this extensive database provides information on galaxies with multiple observations from different telescope / instrument configurations and from different authors .
these overlapping data are used to derive relations to transform all available measurements into a common system , thereby ensuring the homogeneity of the database .
the enearm redshift - distance survey extends the earlier work of the 7s and the recent tully - fisher surveys sampling a comparable volume . in subsequent papers of this series
we intend to use the enear sample by itself or in combination with the sfi tully - fisher survey to analyze the properties of the local peculiar velocity field and to test how sensitive the results are to different sampling and to the distance indicators .
we also anticipate that the homogeneous database assembled will be used for a variety of other applications and serve as a benchmark for similar studies at high - redshift .
psfig -1@xmath4 # 1 1@xmath4 2 @xmath5 26 @xmath6 26 @xmath7 2 @xmath8 # 1,aj , # 1 , # 1,apj , # 1 , # 1,apj , # 1 , # 1,apjs , # 1 , # 1,mnras , # 1 , # 1,mnras , # 1 , # 1,pasp , # 1 , # 1,science , # 1 , # 1,nature , # 1 , # 1,_zeitschrift fr astrophysik _
, # 1 , |
we observed ngc 5383 in a broad band ( @xmath14 @xmath15 @xmath1 1240 ) r - filter and a narrow band ( @xmath14 @xmath15 = 25 ) filter at the 1.5 m telescope at palomar on the night of 2 june , 1994 with a 2048 @xmath16 2048 ccd camera in direct imaging mode . in this mode
the camera has a 12@xmath77 field of view with 0@xmath1737 pixels . using the r - band filter we obtained one 300 s and one 600 s exposures which were co - added to obtain a single r - band image .
the image was a single 900 s exposure .
the effective resolution was 1@xmath175 in the r - band image and 2@xmath174 in the image .
for both images we divided by a normalized flat - field frame , subtracted the mean sky brightness , and removed cosmic rays using standard routines in the noao / iraf software package .
we also corrected all images for atmospheric extinction .
we used the standard stars bd+26.2606 and bd+17.4703 for absolute flux calibration using the absolute photometry by oke & gunn ( 1983 ) .
finally , we registered foreground stars in each image with the hubble guide star catalog ( gsc ) and determined astrometric solutions for each image .
the residuals in determining absolute positions are smaller than 0@xmath1703 ; however , systematic errors in the gsc prevent us from achieving an accuracy @xmath18 1@xmath13 .
we used the r - band image to subtract the underlying continuum from the image .
the final continuum - subtracted line image is shown in figure [ hafig ] .
we observed ngc 5383 in the near - infrared j and k@xmath19 bands ( wainscoat & cowie 1992 ) on the nights of 1 , 4 march 1994 at the 1.3 m telescope on kitt peak using the cryogenic optical bench ( cob ) .
the cob used a 256 @xmath16 256 insb detector with 0@xmath1795 pixels ( @xmath14@xmath20 field of view ) .
the total on - source integration time was 24 minutes in the j band and 52 minutes in the k@xmath19 band .
the effective resolution is 2@xmath174 in the k@xmath19 band image and 2@xmath13 in the j band image .
the data were reduced via the procedure described by regan et al .
( 1995 ) . since the k@xmath19 band image traces light primarily from the old stars which are good tracers of the gravitational potential , figure [ kbandfig ] is shown to emphasize the bar potential in ngc 5383 .
we observed the nucleus of ngc 5383 with the f160w filter on camera 2 of the near infrared camera and multi - object spectrometer ( nicmos ) in the hubble space telescope on 13 october 1997 .
the f160w filter is an approximate match to ground - based near infrared h - band filters and has a central wavelength of 1.6 .
our total integration time on source was 704 seconds .
the detector on camera 2 is a 256@xmath16256 hgcdte detector with a plate scale of 0.075 arc seconds pixel yielding a field of view of 192 . during the time period of the observations
the pipeline data reduction for nicmos was still under development leading to output of the pipeline process that was not the best possible calibration of the data .
therefore , we re - calibrated the images using the latest flat fields , darks , and non - linearity files .
we observed co ( j=1 - 0 ) line emission in ngc 5383 with the bima ( berkeley - illinois - maryland association ) array in three different configurations of six dishes from february - june 1994 , and in a compact configuration of nine dishes from july - august 1996 .
the phase and pointing center for these observations was @xmath0(j2000 ) = 13@xmath257@[email protected] , @xmath21(j2000 ) = 41@xmath950@xmath19 46@xmath1700 .
the correlator window was centered at v@xmath22 = 2245 with a total bandwidth of 424 and a resolution of 4.1 . in each case
the line emission was observed in the upper sideband .
the projected baselines ranged from 2.2 to 87.5 k@xmath15 and the single - sideband system temperatures ranged from 500 to 1100 k. we flagged the data with anomalous visibilities resulting from shadowing of an antenna at low elevation and spurious electronic spikes ( birdies and correlator window edges ) .
the complex instrumental gain was calibrated using the quasars 1153 + 495 , 1419 + 543 or 1310 + 323 , observed every 20 - 30 minutes .
the absolute flux calibration was determined from observations of uranus or mars and the structure of the if band pass was determined from observations of the quasars 3c273 or 3c279 .
we imaged the calibrated data set into 10 channels using natural weighting to achieve maximum signal to noise ratio and then cleaned the dirty maps using the hogbm algorithm ( hogbm 1974 ) .
next we employed an iterative phase - only self - calibration process ( thompson , moran , & swenson 1986 ; regan et al . 1995 ) which used the clean components as model inputs for the iterations .
the self - calibration process did not significantly improve the results because our initial atmospheric phase calibration was good .
in other words , our data are limited by thermal noise rather than the dynamic range .
the synthesized beam in the maps is 4@xmath1762 @xmath16 4@xmath1723 at a position angle of 17 degrees .
the final channel maps have a noise level of 50 mjy beam@xmath12 .
we summed over all pixels in the data cube with @xmath23s@xmath24@xmath23 @xmath25 2@xmath26 to form a velocity - integrated total intensity map which is shown in figure [ cofig ] .
we obtained a spectrum of the center of ngc 5383 in the co ( j=10 ) emission line with the single dish , 12 m nrao telescope on 01 december 1998 .
the half - power primary beamwidth of the telescope is 53@xmath13 and the pointing accuracy is @xmath8 5@xmath13 .
we used the spectral line beam position switching observing mode with a 4@xmath20 throw for the off position .
the total on - source integration time was nine minutes with a system temperature of @xmath1350 k. the filter banks were configured in the 2 @xmath16 2 mhz series mode for a total bandwidth of 512 mhz .
an unstable oscillation in one channel prevented us from using data from both channels .
the measured rms noise in a 2 mhz channel is 19.5 mk .
a zeroth order baseline was subtracted from each of the three scans before they were co - added ; the resulting spectrum is shown in figure [ cosdfig ] .
co emission is detected in the central 3.5 kpc diameter region of the ngc 5383 .
we find two prominent co peaks ( labeled ci , cii in figure [ cofig ] ) in the total intensity map , corresponding to the inner termini of the dust lanes .
in addition to these two peaks , we see a third peak ( labeled ciii in figure [ cofig ] ) east - southeast of peak i. peak ciii lies along the northwestern dust lane after it has curved and is heading east across the nuclear region .
there is co emission in an even weaker feature ( labeled iv ) 5@xmath13 north of ciii .
this feature coincides with the most intense star forming region in the nucleus . within the primary beam ( 53@xmath13 fwhm ) of the nrao 12 m telescope , the total co flux in the bima interferometer maps
is @xmath1175 jy .
this is about 60% of the flux ( 9 k ) measured from the nrao spectrum , assuming a conversion factor of 33 jy k@xmath12 for the 12 m telescope .
most of the co detected by the interferometer is concentrated in the central 35@xmath13 of ngc 5383 . the co flux in this 35@xmath13 region ( 165 jy )
can be converted to a total h@xmath27 mass using the standard equation m(h@xmath27 ) = 1.1 @xmath16 10@xmath28 d@xmath29 s@xmath30 ( kenney et al . 1992 ) , where d is in mpc and s@xmath30 is in jy . for the adopted distance of 32.4 mpc ,
the total molecular gas mass is @xmath12 @xmath1610@xmath31 .
we have divided the r and j band images into the k@xmath19 band image to form color maps ( see figure [ dustfig ] ) .
prior to division , the images were registered against each other and the higher resolution images were smoothed to match the lower resolution ( 2@xmath174 ) of the k@xmath19 band image . in both maps
the darker grey scale corresponds to redder colors ( higher extinction ) .
although it is difficult to estimate the dust content in galaxies , it has been shown that optical - infrared colors such as r@xmath32k@xmath19 and infrared - infrared colors such as j@xmath32k@xmath19 can be used to estimate the amount of dust using multiple scattering radiative transfer models ( regan et al .
1995 ) ; however , such models require images at multiple optical and infrared wavelengths to confidently parameterize the optical depth and the ratio of dust scale height relative to the stars . even though the lack of multiple images at different wavelengths prevents us from this quantitative analysis
, we can still gain a qualitative understanding of the dust distribution from the two colors available to us .
in both maps , the main bar dust lanes show up as regions of high extinction . in the r@xmath32k@xmath19 map ,
the northwestern dust lane is relatively straight and narrow with at least two spurs ( indicated by arrows in figure [ dustfig ] ) , which merge into the dust lane , and end in areas with high levels of dust extinction . in the southeastern dust lane the spurs and the dust lane are difficult to distinguish from each other , and the dust lane characteristics are harder to describe .
however , this dust lane is seen more clearly in the j@xmath32k@xmath19 color map , where its appearance and structure is similar to the northwestern lane .
note that the northwestern dust lane , on the other hand , is not clearly visible in the j@xmath32k@xmath19 map .
the differences between the two color maps could be due to a variety of reasons such as different dust lane thicknesses , geometry of the dust / star mixture ( witt 1992 ) , systematic reddening of the near side of a galaxy based on its orientation in the sky ( e.g. elmegreen & block , 1999 ) , or the more remote possibility that the stellar populations are different between the two sides of the bar . at the inner terminus of the dust lanes where the dust lanes begin to curve , two red peaks mark locations of the highest dust extinction .
these dust extinction `` twin peaks '' match the peaks seen in co emission .
interior to these peaks , a s - shaped curve is traced in dust extinction similar to the curve seen in co emission . at the center of the s - shaped curve
we find a peak with the reddest color in the map .
since this central peak coincides with the peak of k@xmath19 band emission it probably reflects excess k@xmath19 band emission from either a change in the underlying stellar population ( excess red giant stars ) , or hot dust ( e.g. , thatte et al .
1997 ; marco & alloin 1998 ) , instead of a large amount of dust extinction .
the s - shaped structure , the dust lanes and the spurs , however , can not be due to k@xmath19 band emission excesses because the k@xmath19 band emission is much smoother ( figure [ kbandfig ] ) ; hence , these must be due to dust extinction .
strong line emission ( figure [ hafig ] ) is detected in the nucleus , the bar ends and the spiral arms . in the nucleus
the peak emission ( indicated by the label hp in figure [ hafig ] ) is located 3@xmath171 northwest of the radio - continuum peak .
we can clearly trace the northwestern dust lane in absorption against the emission as it curves inwards , 3@xmath13 south of hp .
we also see a bright elongated hii region , 5@xmath13 southwest of hp , along the southern edge of the dust lane .
similarly , there is another bright hii region 11@xmath13 east of hp , near the inner terminus of the southeastern dust lane .
hii regions are also observed at both bar ends and along the spiral arms near the bar ends .
since it is unclear whether the observed hii regions belong to the bar end or the spiral arm , we use the term `` bar end / spiral arm hii regions '' to describe any hii regions in this area . the bar end / spiral arm hii regions mostly trail the bar , although there is at least one bright hii region on the leading side of the bar .
the hii regions extend for about 4 kpc on the trailing side of the spiral arms and about half that distance along the leading side .
the study of star formation activity in these regions is interesting because of the changing environment between the bar ends and spiral arms ( kenney & lord 1991 ) .
however , in this paper , we limit the scope of our study and do not study the star formation activity in these regions . instead , we choose to focus on the star formation activity in the bar .
while the nuclear region and the bar ends have strong emission , the bar itself is comparatively weak ; this pattern is consistent with observations of other early - type barred galaxies ( garcia - barreto et al . 1996 ; koopmann & kenney 1996 ; phillips 1993 ) .
most of this limited star formation activity in the bar of ngc 5383 is found in one or two bright star formation sites on each side of the bar , along its leading edges .
these peaks are best seen in the contour image of emission in figure [ hadust ] and their significance is discussed in [ stardust ] .
modeling studies of gas flows in bars constitute an extensive subfield and we refer the reader to excellent reviews in sellwood & wilkinson ( 1993 ) , athanassoula ( 1992a , b ) , teuben ( 1996 ) and references therein .
one can broadly categorize these modeling efforts into two main categories : a ) cloud - based / sticky particle simulations which treat the gas as a set of distinct particles which react to collisions based on some prescription for cloud collisions , and b ) hydrodynamic ( grid - based or smooth particle hydrodynamics ( sph ) ) simulations which treat the gas as an ideal fluid obeying the basic hydrodynamic equations .
both classes of models have advantages and disadvantages and both can provide adequate frameworks in which observations can be interpreted .
both classes of models produce dust lanes but the hydrodynamic models alone can produce the straight dust lanes seen in strongly barred galaxies such as ngc 5383 .
both classes predict gas inflow of approximately the same order of magnitude but the exact route followed by the gas is very different in each class of models . in the hydrodynamic model ,
the gas undergoes a shock at the dust lane and is forced to flow directly inwards in the dust lane ( rvt97 , athanassoula 1992b ) . in the cloud - based models ,
the gas clouds in the dust lane behave like those found in galaxy spiral arms .
the dust lane is formed because the clouds crowd together for a longer time . however , unlike the hydrodynamic models , they eventually cross and leave the dust lane on the leading side of the bar . in this model ,
the clouds can collide , lose angular momentum and spiral down to the center in a few orbits ( combes & gerin 1985 ) .
for strong bars , the hydrodynamic model gas flow is the best match to the observed velocity fields ( e.g. , ngc 1530 rvt97 ) .
it is also particularly successful at producing the high shear and shock seen across the bar dust lanes ( athanassoula 1992 ) ; these can not be produced by the cloud - based / sticky particle models ( regan , sheth & vogel 1999 , hereafter rsv99 ) .
moreover , a recent study of gas kinematics in seven barred spirals also shows that the observed gas kinematics are consistent with the hydrodynamic models and not with cloud - based simulations ( rsv 1999 ) .
so even though the molecular ism is certainly not a diffuse , ideal fluid , it appears that the hydrodynamic models are better at reproducing observations than cloud - based / sticky particle models , especially for describing observations of gas kinematics and morphology in strong bars .
hence we choose to compare the co and dust extinction morphology of ngc 5383 to a hydrodynamic model ; the model chosen is the pst95 standard model . in order to compare the co emission and dust extinction distribution to the hydrodynamic model , we projected the pst95 model ( figure [ modmorph]a ) to match the orientation of ngc 5383 ( figure [ modmorph]b ) and smoothed the model to match the resolutions of the dust extinction and co emission maps ( figures [ modmorph]c , d respectively ) .
the model gas distribution shows an extremely dense nuclear ring with two peaks along the ring near the bar major axis , consistent with pst95 .
however , we emphasize that the location of these peaks varies in the simulations . over time
additional peaks often appear and disappear in the ring , downstream of the dust lane terminus .
the density contrast between the peaks and the dust lane is large ; in the unsmoothed model , the peaks have a column density which is @xmath1 25 to 120 times higher than that in the dust lane , and the average ring column density is @xmath1 25 to 60 times larger . smoothing the models reduces the average density contrast by @xmath150% at the resolution of the dust extinction maps and by @xmath160% at the resolution of the co emission maps .
we also find that the gas density in the dust lane increases inwards along the dust lanes ; from the outer terminus to the inner terminus the density increases by a factor of @xmath14 .
the model also produces offset bar dust lanes and an outer , low density ellipsoidal feature .
when we compare figure [ modmorph ] to figures [ cofig ] & [ dustfig ] , we find significant differences between the model gas morphology and the observations .
the most striking disparity is the density contrast between the dust lane gas and the ring / peak gas .
the contrast in the dust extinction maps between the dust lane and the twin peaks is 1.3 in the j@xmath32k@xmath19 map and 1.5 in the r@xmath32k@xmath19 map . at the same resolution ,
the model column density contrast is an order of magnitude ( @xmath1 @xmath168@xmath3240 ) larger . since co is not detected in the dust lanes , only a lower limit to the co brightness ratio between the peaks and the dust lane can be calculated . in ngc 5383 , this ratio is @xmath120 , similar to that observed in ngc 3351 and ngc 6951 ( kenney et al .
in other galaxies where co is detected in the dust lanes ( e.g. ngc 1530 , see figure 11 or table 3 in downes et al .
1996 , or ngc 7479 , see figure 5 in quillen et al .
1995 ) , the co brightness ratio between the ring / peak and the dust lane gas is only @xmath15@xmath3210 ; at the same resolution , the model density contrast is @xmath115@xmath3270 .
one reason for the high co brightness ratio in ngc 5383 may be that the co emissivity per h@xmath27 molecule is enhanced in nuclear region ( regan et al . 1995 ) or at least at the twin peaks .
since variations in co emissivity may determine the co brightness ratio , dust extinction is perhaps a better tracer of the gas column density because the gas to dust ratio is better constrained than the co to h@xmath27 conversion factor ( sodroski et al .
1995 ; regan 2000 ) .
conversion factor is an upper limit ] therefore comparing the dust extinction maps to the model , we conclude that the ratio of the ring to the dust lane gas surface density is an order of magnitude larger than seen in the observations .
another striking difference between the model and the observations is the central morphology of gas and dust . in the pst95 study , a nuclear ring always formed whenever
the bar had offset dust lanes ( i.e. had an ilr ) , and whenever the bar was thick ( axial ratio @xmath18 5 ) . in ngc 5383 both of these conditions
are met , yet there is no sign of a nuclear ring ; instead , as described in [ comorph ] and [ dustmorph ] , the gas and the dust are distributed in a s - shaped pattern with three peaks . in some galaxies , e.g. ngc 1530 ( regan et al . 1995 ) , higher resolution co maps have resolved the twin peaks structure into a broken ring or partial spiral arms . in ngc 5383 , however , higher resolution uniform weighted maps ( not shown ) do not reveal a ring ; they show the same triple peak structure seen in the lower resolution co map .
also , the dust extinction maps , which are at a higher resolution than the co maps , do not show a ring . in some other `` twin peak '' galaxies , e.g. ngc 6951 , ngc 3351 and m101 ,
a nuclear ring of hii regions has been observed ( kenney et al .
. however , we see no evidence of an ring in our data ; the emission is amorphous and is dominated by the bright peak north of the nucleus ( see figure [ hafig ] ) .
while it is possible that an ring can be obscured by highly variable and patchy dust extinction in the nuclear region ( phillips 1993 ) , the dust distribution in ngc 5383 is not ring - shaped and can not hide an hii ring .
at this point , it worthwhile to mention that the co morphology in the center of bars can have a rich variety of morphologies in addition to `` twin peaks . '' as kenney ( 1996 ) points out , there are spiral arms as in ngc 6951 ( kenney 1996 ) , filled exponential disks as in ngc 3504 ( kenney , carlstrom & young 1993 ) or ngc 4102 , and rings or partial rings as in ngc 4314 ( benedict , smith & kenney 1996 ) .
so although we are only examining a single case , the failure of the hydrodynamic models to produce these morphologies in general is problematic .
lastly , the models differ from the ngc 5383 observations because they produce a low density , ellipsoidal feature which encircles the bar region .
the ellipsoid is present in the models throughout the simulation , and while its density varies , the density is always fairly low .
this feature is not detected in our observations , and we are not aware of any barred galaxy with such a molecular gas feature ; this may be due to the lack of sensitivity of current millimeter wavelength interferometers .
the ellipsoidal feature probably correspond to the inner ring seen in optical images of other galaxies presented in the review by buta & combes ( 1996 ) .
the inner ring is usually a ring or ellipse of star formation which encircles the bar and may be associated with a 4:1 ultraharmonic resonance ( buta & combes 1996 ) .
good examples of this feature are seen in the continuum - subtracted images of ngc 3504 ( kenney et al .
1993 ) and ngc 6782 , ugc 12646 , ic 5240 ( crocker , baugus & buta 1996 ) . before we can discuss the importance and relevance of this feature to gas flows in bars
, we would need to obtain additional observations of molecular gas emission and dust extinction to identify the feature , and we would need to study the models more carefully to establish the role of the gas in this ellipsoidal feature .
while there are many discrepancies between the model gas distribution and the observations , there are also some similarities .
as already shown by a92 , both straight and curved bar dust lanes are reproduced as a function of the bar strength by the hydrodynamic model . in our study , the model ( figure [ modmorph ] ) predicts increasing gas density inwards in the dust lane ; this matches the observations of co emission and dust extinction ( figure [ cofig ] , [ dustfig ] ) .
a similar trend is also seen in the barred spiral ngc 1530 ( see figure 2.10 in regan et al .
1995 , or figure 11 in downes et al 1996 ) . in summary
, the hydrodynamic model reproduces the observed offset dust lanes and the observed increase in gas density inwards along the dust lane .
but there are key differences between the model and the observations .
the model predicts a nuclear ring morphology with an extremely high density contrast between the ring and the dust lanes , whereas observations show a s - shaped morphology with three peaks with a relatively low density contrast between the dust lanes and the peaks .
one reason for the absence of the predicted morphology could be the presence of a nuclear bar which could drive gas inwards from the twin peaks .
we searched for a nuclear bar in a high resolution , high dynamic range nicmos image ( figure [ nicfig]a ) .
the image shows a relatively smooth nucleus with a collection of star forming regions north northwest of the nucleus .
these correspond to the peak hp seen in the lower resolution image ( figure [ hafig ] ) .
the main bar dust lane can also be seen in absorption as it travels inwards towards the nucleus .
there is no evidence of a nuclear bar at the center of ngc 5383 .
since the nucleus in ngc 5383 is quite bright , it is possible that weaker structure in this region may be difficult to discern . to investigate this possibility we applied the technique of unsharp masking where we subtracted a 14 pixel fwhm gaussian smoothed image ( figure [ nicfig]b ) from the original image ( figure [ nicfig]a ) .
pioneered by malin & zealey ( 1979 ) , this technique allows one to distinguish weaker features from brighter ones by reducing the contribution of the bright background ; however , we caution that the unsharp masking technique is not perfect and detailed studies of nuclear structure may benefit from multiple techniques . for our purposes
the unsharp masking technique is adequate ; the resulting image is shown in figure [ nicfig ] .
it shows a well - defined two arm , trailing spiral pattern which starts from the main northwestern dust lane and winds its way down to an unresolved core ; there is no evidence of a nuclear bar .
the observed trailing pattern is not unique to ngc 5383 ; central regions of other galaxies also show similar nuclear trailing spirals ( e.g. ngc 3982 , ngc 3032 , mkn 573 regan & mulchaey 1999 and references therein ) . this trailing spiral arm pattern may be a mechanism for transporting gas further inwards from the location of the inner lindblad resonance ( regan & mulchaey 1999 ) where the gas inflow is known to stall in modeling studies .
models have not predicted this trailing spiral pattern .
analytic solutions predict a leading spiral structure inside the inner inner lindblad resonance ( iilr ) ( combes 1996 ; yuan & kuo 1997 ) .
although a trailing spiral pattern can be excited at the outer inner lindblad resonance ( oilr ) , only a leading spiral is excited by the iilr .
it is unclear whether the oilr excited trailing spiral pattern can propogate inside the iilr .
it is also unclear how the two spiral patterns interact and whether the relative amplitudes of the two patterns can appropriately account for these observations .
these issues are under investigation ( yuan et al . 1999 ) .
the smoothed image used for the unsharp masking ( lower right hand corner in figure [ nicfig ] ) has a resolution and quality comparable to good ground - based images . in this image
, we clearly see a nuclear bar - like feature at the center of ngc 5383 even though there is no bar in the original nicmos image and in the unsharped masked image .
the bar - like feature is created by the smoothing of the light from the bright star forming regions north and south of the nucleus which leads to the roughly north - south elongation , i.e. the fake bar . smoothing the original image with a gaussian with fwhm of 7 pixels
, one can even create a partial circumnuclear ring ( see figure [ nicfig]d ) .
although this is more difficult to see than the bar , an oval encircling the fake nuclear bar is clearly visible in this image . in reality , these structures are simply artifacts of the various smoothing techniques applied to the original image .
this result shows that study of the nuclear structures in a galaxy requires high resolution observations , and that ground - based images can be inadequate and worse , misleading in the identification of these structures .
another important point of note is that even in the h - band , where the effects of dust extinction and star formation are supposed to be minimized , these nicmos images clearly show star forming knots and dust features .
the smoothing experiments show how these contaminants can create false features .
so although near - infrared observations are desirable compared to optical observations ( which are even more sensitive to these contaminants ) , the near - ir data also need to have high resolution to properly investigate nuclear features .
thus , we conclude from this study of the nicmos image of the center of ngc 5383 that a nuclear bar is not responsible for the differences between the observations and the model predictions .
we considered the possibility that the hydrodynamic models could produce offset dust lanes without producing a nuclear ring .
the motivation for this search came from an examination of the bi - modal outcomes ( i.e. offset dust lanes with a ring versus centered dust lanes without a ring ) of the previous studies by a92 and pst95 . since these studies took rather coarse steps in parameter space , we considered the possibility of an intermediate scenario where offset dust lanes could exist without a ring or with some other nuclear gas morphology .
we studied the parameter space by varying three of the four free parameters in the pst95 model ( the four parameters are the bar axial ratio , the central density , the lagrangian radius or equivalently the pattern speed , and the quadrupole moment or equivalently the bar mass @xmath32 see pst95 for details ) in such a way as to encompass the transition region between the two outcomes described above .
though we found some new and interesting results from this experiment ( see appendix a ) , we did not find the intermediate scenario we were searching for ; in every case where offset dust lanes formed , we found that a nuclear ring also formed ( see figure [ modfig]a , b , c ) .
hence , we come to the conclusion that gas hydrodynamic processes alone can not determine the nuclear region gas morphology . initial parameters and the length of time over which the bar has evolved both determine the density of the ring in the model .
since in the pst95 model the inflowing gas simply accretes in the nuclear ring over time , the density contrast between the ring and the dust lanes is destined to increase over time .
the low density contrast in the observations of ngc 5383 might suggest that the bar in this galaxy is fairly young .
however , similar low gas density contrast in several other barred spirals indicates that it is more likely that the gas is somehow being depleted from the ring .
one way of depleting the gas would be to convert the gas into stars .
kenney et al . ( 1993 ) showed that in the nuclear region of ngc 3504 , the toomre q parameter is close to its critical value ; they suggested that the ongoing nuclear star formation could be explained using a simple gravitational instability picture .
elmegreen ( 1994 ) also showed analytically that at typical nuclear ring radii ( r @xmath1 1 kpc ) , star formation can occur via gravitational instability if the ring gas density is roughly 100 @xmath16 disk gas density ( at r @xmath1 10 kpc ) , assuming that the disk gas density is nearly critical . in the models ,
the ring density is a factor of 50 - 200 @xmath16 initial disk density which is sufficiently high for star formation via gravitational collapse . indeed , in the nuclear region in ngc 5383 ( figure [ hafig ] ) , as in numerous other barred spirals ( phillips 1993 ; garcia - barreto et al .
1996 ) , we observe vigorous , circumnuclear , massive star formation .
the nuclear star formation rate in ngc 5383 is at least 7 ( [ starform ] ) .
if continuous star formation is assumed , this rate sets a lower limit on the mass inflow rate .
based on measurements in other barred spirals such as ngc 7479 and ngc 1530 , which are similar to ngc 5383 , it is likely that the mass inflow rate is lower ( quillen et al .
1995 and rvt97 measure inflow rates of 14 ) .
but the mass inflow rate is a difficult quantity to measure and these numbers are highly uncertain .
still it is unlikely that the high nuclear star formation activity can be supported by a proportionally high mass inflow rate over an extended period of time ( for example , an inflow rate of 10 , would deplete the total gas reserve ( 10@xmath33 ) of a typical galactic disk in only a billion years ) .
one can also compare the total gas mass in the center of ngc 5383 to the current star formation rate . using the standard co - h@xmath27 conversion factor ,
the total molecular gas mass in the center of ngc 5383 is 2@xmath1610@xmath31 .
given that the dust and the co intensities differ between the bar and the nuclear region , and given that the dust is probably a better tracer of the molecular gas mass , the total molecular gas mass calculated from the standard conversion factor is probably an upper limit to the total gas mass in the circumnuclear region .
therefore at the observed star formation rate of 7 , the gas reservoir at the center of ngc 5383 can be depleted in less than 3@xmath1610@xmath34 yrs .
so , as has been suggested before , circumnuclear star formation in ngc 5383 is probably an intense and episodic event ( shlosman 1992 ) .
thus the observed low density contrast can be attributed to gas depletion via circumnuclear star formation .
the simulations do not reproduce the observations because , although star formation can occur at the model densities , star formation is not modeled .
we used the luminosity to estimate the star formation rate in various regions in ngc 5383 ( see table [ sfrtab ] ) .
we measure an flux of 2.42 @xmath8 0.09 @xmath16 10@xmath35 in the central 3.5 kpc diameter region and a total flux in the galaxy of 4.37 @xmath8 0.36 @xmath16 10@xmath35 .
we converted the observed flux to a star formation rate using the empirical equation derived by kennicutt ( 1983 ) , @xmath36 in deriving this equation kennicutt assumed an average intrinsic dust extinction of 1.1 magnitudes in a galaxy . since the dust extinction can vary dramatically in star forming regions ( phillips 1993 ) , the exact optical depth is difficult to determine ; hence , we use kennicutt s value to correct the observed fluxes for dust extinction . using an adopted distance of 32.4 mpc
, we calculate the total star formation rate in ngc 5383 to be 12.6 , of which 7 , are formed in the nuclear region . for any single galaxy ,
this type of calculation of the star formation rate is uncertain by a factor of 2 ( kennicutt 1983 ) .
a previous calculation using a different technique and the h@xmath37 ionization line by duval & athanassoula ( 1983 ) resulted in a nuclear star formation rate of 8 ; our results are consistent with their value .
the star formation rate at the northeastern and southwestern bar ends is 1.9 & 3.2 , respectively , smaller than the star formation rate in the nuclear region .
lrrrr h - i & nuclear region & 1074 & 2.42 @xmath8 0.085 @xmath16 10@xmath35 & 7
h - ii & ne bar end region & 1026 & 0.65 @xmath8 0.07 @xmath16 10@xmath35 & 1.9
h - iii & sw bar end region & 1452 & 1.09 @xmath8 0.11 @xmath16 10@xmath35 & 3.2 & the galaxy & 4763 & 4.37 @xmath8 0.36 @xmath16 10@xmath35 & 12.6 the star formation rate of 7 , in the central 3.5 kpc of ngc 5383 is substantially higher than the average star formation rate over the entire disks of sb and sc type galaxies , which varies from 0.1 to 4 , ( kennicutt 1983 ) .
however , such a high star formation rate is not unusual for centers of barred spirals , and such activity is often classified as a starburst ( e.g. , jogee 1999 ; contini et al .
the total mass of h@xmath27 ( 2 @xmath16 10@xmath31 ) calculated earlier is also not unusual and it places ngc 5383 in company of galaxies such as ngc 4536 and ngc 2782 ( jogee 1999 ) which are both gas rich in the nuclear region and hosts to high nuclear star formation activity .
excluding the bar ends and the nuclear regions , early type barred spirals have very low star formation activity in the bar , even though the gas concentration is high in the bar dust lanes ( downes et al .
1996 ) . as already shown in [ hydromorph ] , the density contrast between the dust lanes and the center is small . yet
the star formation activity in the bar is limited compared to the center ; several explanations have been put forth to account for this behavior .
tubbs ( 1982 ) suggested that molecular clouds entering the dust lane could be dispersed due to their high velocities relative to dust lane gas .
athanassoula ( 1992 ) argued that the high shear in the bar dust lanes prevents star formation .
rvt97 pointed to the large divergence in gas streamlines prior to the dust lanes , which they argued could tear apart molecular clouds or prevent their formation . recently
, reynaud & downes ( 1998 ) argued that a combination of the shear and shock in the dust lanes leads to the low star formation rate in the bar .
these hypotheses argue effectively for the inhibition of star formation in the bar dust lanes . yet
some star formation does occur in the bar . here
we investigate precisely where and under what circumstances stars form in the bar .
the star formation sites can be traced by emission , whereas the molecular gas can be traced by either co emission or dust extinction .
the co emission in the bar dust lanes of ngc 5383 is too weak to be detected , so we assume that the dust extinction traces the molecular gas ( as suggested observations of ngc 1530 by downes et al .
1996 and observations of ngc 3627 and ngc 2903 by rsv99 ) .
figure [ hadust ] shows the overlay of the emission on the dust extinction in the bar on ngc 5383 . comparing the relative distribution of the star formation sites to the molecular gas distribution in this figure ,
one immediately notices two important features .
first , the peaks are all located on the leading side of the bar dust lane .
second , these peaks occur preferentially near dust concentrations ( see arrows in figure [ hadust ] ) ; these peaks are found at the ends of dust spurs which are most clearly seen in the r@xmath32k@xmath19 ( figure [ dustfig ] ) image . the peaks and the dust peaks are roughly @xmath1 3@xmath13 , or @xmath1 470 pc apart . in order to place these observations in the context of the hydrodynamic model gas kinematics , we reiterate the predictions of the gas flow in the models
in these models , gas flows in elliptical streamlines until it encounters the main bar dust lane where the hydrodynamic shock redirects the gas inwards ( rvt97 ; rsv99 ) ; _ none _ of the gas encountering the dust lane crosses the dust lane .
hence , the presence of emission on the leading side of the dust lane is surprising .
this is because the recent massive star formation indicated by the emission is thought to occur only in areas of high gas concentration such as giant molecular clouds , or in larger complexes of such clouds called giant molecular associations ( blitz 1994 and references therein ; vogel , kulkarni & scoville 1988 ) ; the area on the leading side of the bar dust lane does not have high gas concentrations .
therefore , we conclude that these young stars must have formed elsewhere . in the bar ,
the only regions suitable for star formation , i.e. regions of high gas concentrations , are the dust lanes and the dust spurs .
the dust lane , however , is thought to be an inhospitable environment for various reasons already outlined above ; the spurs are a plausible alternative .
since there is a one to one correspondence between the peaks and the dust spurs with the peaks located directly across from the spurs , it is likely that the spurs are involved in star formation . in context of the hydrodynamic gas flow , stars forming in the spurs would continue to move along their original elliptical orbits .
these stars would pass ballistically through the dust lane , travel to the leading side and ionize the low density gas there .
this would be consistent with the peaks being located directly across from the dust spurs .
another important reason for the spurs to be conducive to star formation is that , in addition to being regions of high gas density , the spurs are also located in a region of lower shear .
although we do not have a full velocity field map for ngc 5383 which shows the lower shear environment , the lower shear can be clearly seen in the gas , prior to its encounter with the dust lanes , in the maps of other barred galaxies such as ngc 1530 ( rvt97 ) .
the distance between the peaks and the dust concentrations at the end of the spurs puts a strong constraint on the velocity of the newly formed stars .
if the stars form in the spurs , the stars must traverse @xmath1500 pc in 10@xmath38 years and therefore have velocities @xmath1 50 ( in the rotating frame of the bar ) ; indeed , velocities of this magnitude are present in the hydrodynamic model before the gas encounters the dust lane .
so the distance between the peaks and the dust lanes is also consistent with star formation occurring in the spurs .
thus we have addressed where and under what circumstances stars form in the bar , given the predictions of the hydrodynamic model gas flow .
the observations presented here suggest the following scenario : at certain locations along the bar gas becomes highly concentrated and forms spurs .
star formation occurs in the spurs because of the high gas density _ and _ low shear .
the newly formed stars move ballistically through the dust lane , ionizing the low density gas on the leading side of the bar and appearing directly across from the spurs , whereas the gas is redirected inwards down the dust lane . however , one difficulty in applying this scenario to ngc 5383 is the lack of emission in the spurs .
the spurs are simply not dense enough to hide star formation and therefore one might expect to see emission from ongoing star formation in the spurs
. the lack of such emission in ngc 5383 may be because we are observing very low levels star formation activity ( 2 - 3 peaks on each side ) in the bar of this galaxy .
indeed , observations of other galaxies do show emission on both the leading and trailing sides of the bar dust lane ( e.g. , as in ngc 7479 , laine et al .
1999 ) ; however , it is not known whether this emission is associated with dust spurs . in any case , the one - to - one correspondence between the emission and the spurs in ngc 5383 is unlikely to be completely coincidental and therefore it is highly likely that the spurs are involved in the star formation activity .
another interesting aspect of this scenario is that it points out the need for including gas self - gravity in the hydrodynamic model .
the presence of the spurs clearly indicates that gas is becoming highly concentrated along the bar . in the region of the spurs
, the divergent gas streamlines should prevent cloud formation ( as pointed out by rvt97 ) but the presence of the spurs shows the importance of cloud self - gravity and cloud - cloud gravitational interactions .
these effects are not modeled by the hydrodynamic simulations .
so although these models are highly successful at reproducing the velocity fields and the straight dust lane morphology in bars , and they provide an adequate context for interpreting the observations of emission and the spurs in the bar , these models lack the self - consistency necessary to fully explain the star formation process .
the nuclear gas and dust distribution in the prototypical barred spiral ngc 5383 can not be reproduced by the hydrodynamic models . whereas the model always produces a high density contrast circumnuclear ring , the gas and dust
are observed in a low density contrast s - shaped distribution .
we have shown that the discrepancy can not be eliminated by fine tuning the model .
the alternative possibility that a nuclear bar is responsible for the differences is eliminated by using a high resolution nicmos image . applying the unsharp masking technique , we find that the underlying nuclear structure is a trailing spiral pattern . we have also shown how coarser resolution data , such as that found in ground
based images , can lead to false identifications of nuclear bars and rings .
we conclude that the discrepancy between the observed and modeled gas distribution is due to the absence of star formation in the models ; the vigorous 7 circumnuclear star formation , which can deplete the gas and lead to the observed low density contrast , is not modeled by the hydrodynamic simulations . finally , we present an explanation for how stars may form in the bar between the bar ends and the circumnuclear region . in our scenario ,
stars form in dust spurs before the gas encounters the dust lane .
the spurs , unlike the high shear / high density dust lane , are more conducive to star formation because they are regions of low shear and high density .
stars which form in the spurs travel ballistically through the hydrodynamic shock at the dust lane , ionizing the low density gas on the leading side of the dust lane , whereas the gas is redirected down the dust lane .
thus , hii regions can even be found on the leading side of the main bar dust lanes .
we thank the referee , jeff kenney for his careful reading of the original manuscript and his numerous useful suggestions which have resulted in a far superior paper .
we would also like to thank lee armus for obtaining the optical images and eve ostriker for helpful discussions and suggestions which resulted in figure [ ilrs ] .
we thank jim stone for providing the hydrodynamic model code and for his invaluable insights in interpreting the models .
k.s would like to thank m. thornley , l. looney , m. wright , j. morgan and r. forster for their help with bima data acquisition and reduction .
k.s . would also like to acknowledge useful discussions with s. veilleux , y. fernandez , a. koratkar , n. volgenau and j. crawford .
this research was supported by nsf grant ast 9314847 and ast 9613716 .
previous hydrodynamic models of gas flow in a barred potential have resulted in one of two outcomes : either offset dust lanes and a nuclear ring form or centered dust lanes without a ring form . in this section
we describe our experiment intended to find an intermediate case where offset dust lanes could form without the formation of a nuclear ring ? to that end , we have explored the parameter space using the pst95 model near the transition region using finer steps than those used by a92 and pst95 .
lrrr 1 & 5.85 & 4.5e4 & 2.4e4 2 & 5.90 & 4.5e4 & 2.4e4 3 & 5.95 & 4.5e4 & 2.4e4 4 & 6.00 & 4.5e4 & 2.4e4 5 & 6.25 & 4.5e4 & 2.4e4 6 & 6.50 & 4.5e4 & 2.4e4 7 & 6.75 & 4.5e4 & 2.4e4 8 & 7.0 & 4.5e4 & 2.4e4 9 & 7.5 & 4.5e4 & 2.4e4 10 & 6.0 & 1.5e4 & 2.4e4 11 & 6.0 & 2.5e4 & 2.4e4 12 & 6.0 & 3.5e4 & 2.4e4 13 & 6.0 & 4.0e4 & 2.4e4 14 & 6.0 & 4.7e4 & 2.4e4 15 & 6.0 & 4.5e4 & 2.3e4 16 & 6.0 & 4.5e4 & 2.6e4 17 & 6.0 & 4.5e4 & 2.8e4 18 & 6.0 & 4.5e4 & 3.0e4 19 & 6.0 & 4.5e4 & 3.7e4 20 & 6.0 & 4.5e4 & 4.0e4 21 & 6.0 & 4.5e4 & 7.4e4 the pst95 hydrodynamic models are computed with a third - order piecewise , parabolic method ( ppm ) with a cylindrical coordinate system yielding a resolution of 8 pc at a distance of 0.1 kpc .
the gravitational potential used in this model consists of a kuzmin - toomre disk , a bulge and a bar described by a ferrer s ellipsoid .
the model is completely determined by four independent input parameters : the bar axial ratio ( a / b ) , the central density ( @xmath39 , 10@xmath40 kpc@xmath41 ) , the lagrangian radius ( r@xmath42 , kpc , equivalently the pattern speed ) , and the quadrupole moment ( q@xmath43 , 10@xmath40 kpc@xmath44 , equivalently the bar mass ) ( see pst95 for details ) . we chose to set a / b = 4.0 and varied the other three parameters as follows : 1.0e4 @xmath18 @xmath39 @xmath18 7.4e4 , 5.0 @xmath18 r@xmath42 @xmath18 7.5 , 1.5e4 @xmath18 q@xmath43 @xmath18 7.5e4 . since any of these three free parameters can affect the transition between the two outcomes of the model , we varied one parameter at a time while setting the other two parameters to the `` standard '' model values used by pst95 .
as in pst95 , we ran each simulation until the bar had settled into a quasi - static state i.e. until t = 2.0 gyr ( see pst95 for more details on time evolution of the bar ) .
the final azimuthally averaged log density versus radius are plotted in figure [ modfig ] .
we find that the formation of the ring is indeed abrupt in all instances where an ilr exists .
in other words , if offset dust lanes form , a nuclear ring inevitably forms . with this result
, we conclude that the intermediate case , if present , exists in even finer steps of the model parameters ; otherwise we confirm pst95 s assertions about the robustness of the nuclear ring .
although this experiment failed in finding the intermediate scenario which may have explained the observations of ngc 5383 , we discovered other interesting results which add to our current understanding of gas distribution in the hydrodynamic models .
first , when the ring exists , we find that the ring seems to be located at approximately the same distance and has approximately the same thickness growing only slightly thinner with higher central density concentration .
since the commonly accepted explanation for the location of the nuclear ring is that it must exist between the outer ilr ( oilr ) and the inner ilr ( iilr ) , one can conclude that the ring location and thickness should be restricted to a certain range of radii .
we find that in all 21 model runs ( table [ models ] ) which produce a ring , the ring peak is always located between the oilr and the iilr as shown in figure [ ilrs ] .
interestingly , we also find that the location of the ring peak does not coincide with the location of the @xmath45-@xmath46/2 maximum ( figure [ ilrs ] ) as suggested by pst95 except for the standard model . in most cases
the ring peak is close to the maximum of the @xmath45-@xmath46/2 curve , but its location seems to be determined by the location of the oilr for a given model . in other words the location of the ring peak and
the oilr varies in exactly the same way for the different models but by different amounts from model to model . while this relationship is interesting , we note that it may not be physical since the oilr is located at a large distance ( @xmath1 1.5 - 2 kpc ) from the center
; at these distance , the gas is no longer on circular orbits and hence the classical definition of an oilr may not be relevant .
it would be , perhaps , more interesting to compare the location of the ring peak to the x1 and x2 orbits in the various models .
athanassoula , e. 1992 , , 259 , 328 athanassoula , e. 1992 , , 259 , 345 becker , r. h. , white , r. l. , & helfand , d. j. 1995 , , 450 , 559 benedict , f.g . ,
smith , b.j . , & kenney , j.d.p .
1996 , , 111 , 1861 blitz , l. 1993 , protostars and planets iii , eds . e.h .
levy & j.i .
lunine ( tucson : univ . of arizona press ) , 12 burbidge , e.m . ,
burbidge , g.r . , & prendergast , k.h .
1962 , . , 136 , 704 buta , r. & combes , f. 1996 , fundamentals of cosmic physics , 17 , 95 combes , f. , & gerin , m. 1985 , a&a , 150 , 327 combes , f. 1996 , in iau colloquium 157 , barred galaxies , eds .
r. buta , d.a .
crocker , & b.g .
elmegreen , ( san fransisco : asp ) , 286 contini , t. , wozniak , h. , considere , s. , davoust , e. 1997 , , 324 , 41 downes , d. , reynaud , d. , solomon , p. , & radford , s.j.e .
1996 , , 461 , 186 duval , m. f. & athanassoula , e. 1983 , , 121 , 297 elmegreen , b. g. 1994 , , 425 , l73 elmegreen , b. g. , & elmegreen , d. b. 1985 , . , 288 , 438 elmegreen , b.g . & block , d. l. 1999 , , 303 , 133 garcia - barreto , j.a . , & carrillo , r. , venegas , s. & escalante - ramirez , b. 1996 , in iau colloquium 157 , barred galaxies , eds .
r. buta , d.a .
crocker , & b.g .
elmegreen , ( san fransisco : asp ) , 76 heller , c. h. , & shlosman , i. 1994 , , 424 , 84 ho , l.c . , fillipenko , a.v . ,
& sargent , w. l. w. 1997 , , 487 , 591 hogbm , j.a .
1974 , , 15 , 417 huntley , j.m . 1978 , , 225 , 101 jogee , s. 1999 , ph .
d. thesis , yale university kenney , j. d. p. , & lord , s. d. 1991 , , 381 , 118 kenney , j. d. p. , wilson , c. d. , scoville , n. z. , devereux , n. a. , young , j. s. 1992 , , 395 , l79 kenney , j.d.p .
, carlstrom , j.e . , young , j.s .
1993 , , 418 , 687 kenney , j. d. p. 1996 , in iau colloquium 157 , barred galaxies , eds .
r. buta , d. a. crocker , & b. g. elmegreen , ( san fransisco : asp ) , 150 kennicutt , r.c .
1983 , , 272 , 54 koopmann , r.a . , &
kenney , j.d.p .
1996 , in iau colloquium 157 , barred galaxies , eds .
r. buta , d.a .
crocker , & b.g .
elmegreen , ( san fransisco : asp ) , 105 laine , s. , kenney , j.d.p .
, yun , m.s . ,
gottesman , s.t .
1999 , , 511 , 709 malin , d. f. , & zealey , w.j . 1979 , sky & telescope , 57 , 354 marco , o. , & alloin , d. 1998 , , 336 , 823 martin , p. , & roy , j. 1994 , , 424 , 599 martinet , l. , & friedli , d. 1997 , , 323 , 363 mulchaey , j.s . ,
regan , m.w .
, & kundu , a. 1997 , , 110 , 299 norman , c. a. , sellwood , j. a. , & hasan , h. 1996 , , 462 , 114 ohta , k. , sasaki , m. , saito , m. 1986 , pasj , 38 , 677 oke , j. b. , & gunn , j.e .
1983 , , 266 , 713 peterson , c.j . ,
rubin , v.c . ,
ford , w. k. , thonnard , n. 1978 , , 219 , 31 phillips , a.c .
1993 , ph.d .
thesis , university of washington piner , b.g . ,
stone , j.m .
, teuben , p.j .
1995 , , 449 , 508 quillen , a.c . ,
frogel , j.a . , kenney , j.d.p .
, pogge , r.w .
, depoy , d.l .
1995 , , 441 , 549 regan , m.w . , vogel , s.n .
, & teuben , p.j .
1995 , , 449 , 576 regan , m. w. , vogel , s.n . , teuben , p.j .
1997 , , 482 , 135 regan , m.w . , &
mulchaey , j. 1999 , , 117 , 2676 regan , m.w . , sheth , k. , & vogel , s.n .
1999 , , in press regan , m.w .
2000 , , in preparation reynaud , d. , & downes , d. 1998 , , 337 , 671 sandage , a. & bedke , j. 1994 , the carnegie atlas of galaxies , ( publ .
638 ; washington d.c . ,
carnegie institution of washington ) sancisi r. , allen r. j. , sullivan iii , w. t. 1979 , , 78 , 217 sellwood , j. a. , & wilkinson , a. 1993 , rep .
phys , 56 , 173 sellwood , j. a. , & moore , e. m. 1999 , , 510 , 125 sakamoto , k. , okumura , s. k. , ishizuki , s. , scoville , n. z. 1999 , , in press shlosman , i. 1992 , in relationship between active galactic nuclei and starburst galaxies , ed .
a. v. filippenko , ( san fransisco : asp ) , 335 sodroski , t. j. et al .
1995 , , 452 , 262 teuben , p. j. 1996 , in iau colloquium 157 , barred galaxies , eds .
r. buta , d.a .
crocker , & b.g .
elmegreen , ( san fransisco : asp ) , 299 thatte , n. , quirrenbach , a. , genzel , r. , maiolino , r. , & tecza , m. 1997 , , 490 , 238 thompson , a.r . , moran , j.m , & swenson , g.w .
1986 , in interferometry and synthesis in radio astronomy ( john wiley & sons ) tremaine s. , & weinberg , m.d .
1984 , , 209 , 729 tubbs , a.d .
1982 , , 255 , 458 vogel , s.n . ,
kulkarni , s.r .
, & scoville , n.z .
1988 , nature , 402 wainscoat , r.j . , &
cowie , l.l .
1992 , , 103 , 332 weinberg , m.d . 1985 , ,
213 , 451 witt , a.d . ,
thronson , h.a . , & capuano , j.m .
1992 , , 393 , 611 yuan , c. , & kuo , c. 1997 , , 486 , 750 yuan , c. , sheth , k. , regan , m.w . , & vogel , s.n .
1999 , in preparation | we have mapped the barred spiral ngc 5383 using the bima millimeter - wave array for observations of co ( j=10 ) , the palomar 1.5 m for h@xmath0 and optical broadband , and the kitt peak 1.3 m for near - ir broadband .
we compare the observed central gas and dust morphology to the predictions of recent hydrodynamic simulations calculated using the piner , stone , and teuben code . in the nuclear region ,
our observations reveal three peaks lying along a s - shaped gas and dust distribution : two of these are at the inner end of offset bar dust lanes at the presumed location of the inner lindblad resonance , and the other lies closer to the nucleus .
in contrast , the model predicts a circumnuclear ring , not the observed s - shaped distribution ; moreover , the predicted surface density contrast between the central gas accumulation and the bar dust lanes is an order of magnitude larger than observed .
these discrepancies remain for all our simulations which produce offset bar dust lanes , and indicate that the model is missing an essential process or component .
a small nuclear bar might account for the discrepancy , but we rule this out using a hst nicmos ] introduction ------------ barred spirals constitute a large fraction of all disk galaxies ( @xmath1 75% according to a near - ir survey by mulchaey , regan & kundu 1997 ) .
the non - axisymmetry of a bar potential induces gas inflow which can lead to a number of dramatic evolutionary changes in a galaxy .
these include large central concentrations of molecular gas ( kenney et al .
1992 ; 1996 and references therein ; sakamoto 1999 ) , central starburst activity ( heller & shlosman 1994 ; phillips 1993 ; garcia - barretto et al . 1996
; ho et al . 1997 ) , reduction of the overall chemical abundance gradient ( martin & roy 1994 ; martinet & friedli 1997 ) , formation of a new bulge ( norman , sellwood & hasan 1996 ) , transfer of angular momentum to the halo via dynamical friction ( tremaine & weinberg 1984 ; weinberg 1985 ) and the destruction of the bar itself ( norman et al . 1996 ; sellwood & moore 1999 ) .
lrr r.a .
( j2000 ) & 13@xmath257@[email protected] & ( 1 ) dec ( j2000 ) & 41@xmath550@[email protected] & ( 1 ) d@xmath6 & 2@xmath775 & ( 2 ) major axis p.a .
& 85@xmath83@xmath9 & ( 2 ) inclination & 50@xmath9 & ( 2 ) v@xmath10 & 2250 & ( 3 ) v@xmath11 & 2428 & ... adopted distance & 32.4 mpc & ... linear scale & 157 pc arcsec@xmath12 & ... bar length & 110@xmath83@xmath13 & ( 1 ) bar p.a . &
130@xmath84@xmath9 & ( 1 ) though a majority of these effects depend on the gas inflow , there is no consensus on the exact mechanism for the inflow .
there are at least two classes of models : cloud - based / sticky particle simulations which treat the gas as a collection of discrete particles subject to `` sticky '' collisions ( combes & gerin 1985 ) and hydrodynamic simulations which treat gas as an ideal fluid ( athanassoula 1992b and references therein ; piner , stone & teuben 1995 , hereafter pst95 ) .
both models form dust lanes along the leading edge of the bar , have mass inflow , and form circumnuclear rings ( combes & gerin 1985 ; pst95 ) .
the hydrodynamic models have been remarkably successful in explaining the shapes of the main bar dust lanes , the shock signature at the location of the dust lane ( athanassoula 1992 ) , and the overall gas kinematics in the bar ( rvt97 ) .
therefore we use the framework provided by this class of models to ask two important questions : can the nuclear gas and dust morphology in bars be explained by these models ? and can the gas flow predicted by these models provide a basis for explaining the star formation activity in the bar ? to answer these questions , we use a multi - wavelength dataset for the barred spiral ngc 5383 ; the dataset includes maps of millimeter wave co ( j=10 ) emission to trace the molecular gas , emission to trace the ionized component , and broad - band optical and near - ir emission to trace the stellar distribution and dust extinction .
ngc 5383 is a prototypical early hubble type ( sbb ) barred spiral galaxy .
the global characteristics of ngc 5383 are summarized in table [ global ] .
ngc 5383 has been studied with optical images ( burbidge , burbidge & prendergast 1962 ; duval & athanassoula 1983 ; elmegreen & elmegreen 1985 ) , long - slit spectra ( peterson et al .
1978 ; duval & athanassoula , 1983 ) , hi interferometric observations ( sancisi , allen & sullivan 1979 ) , and single - dish co observations ( ohta , sasaki & saito 1986 ) . the standard model for some of the seminal studies of gas flow in barred spirals has also used ngc 5383 as the prototype ( huntley 1978 ; sanders and tubbs 1980 ; tubbs 1982 ; duval & athanassoula 1983 ; a92 ) .
we first present a comparison of nuclear gas and dust morphology in ngc 5383 to the model predictions in [ hydromorph ] .
the gas and dust distribution is a signature of the bar induced inflow and its subsequent evolution .
if the hydrodynamic model can reproduce the observations , then it may be used as an adequate model for studying gas flow in the nuclear region of bars . on the other hand ,
even a failure of the model can refine our understanding by pointing out missing physics in the model s assumptions .
we find that while the model is consistent with the observations in some aspects , there are several striking differences between the two .
possible reasons for these discrepancies are investigated in [ understand ] where we consider presence of a nuclear bar , incomplete parameter space exploration of the model , or lack of star formation processes in the model .
we conclude that the latter is at least partly responsible for the discrepancies . in [ stardust ]
, we compare the sites of star formation with the molecular gas distribution in the bar and investigate where and under what circumstances stars form in the bar .
this is important because bar star formation can affect the net mass inflow .
the analysis is novel because previous studies ( tubbs 1982 ; athanassoula 1992b ; reynaud & downes 1998 ) have only sought to explain the lack of bar star formation .
we present the first explanation of how stars may be forming in the bar , between the bar ends and the circumnuclear region . |
consider an sde of the form @xmath1 where @xmath2 is a @xmath0-dimensional wiener process , @xmath3 is a bounded measurable mapping from @xmath4 to @xmath4 .
according to @xcite there exists a unique strong solution to equation ( [ eq_main ] ) .
it is well known that if @xmath5 is continuously differentiable and its derivative is bounded , then equation ( [ eq_main ] ) generates a flow of diffeomorphisms .
it turns out that this condition can be essentially reduced @xcite , and a flow of diffeomorphisms exists in the case of possible unbounded hlder continuous drift vector @xmath5 .
recently the case of discontinuous drift was studied in @xcite and the weak differentiability of the solution to ( [ eq_main ] ) was proved under rather weak assumptions on the drift .
the authors of @xcite consider a drift vector belonging to @xmath6 for some @xmath7 such that @xmath8 they establish the existence of the gteaux derivative in @xmath9;{\mathds{r}^d})$ ] . in @xcite it is proved that for a bounded measurable drift vector @xmath5 the solution belongs to the space @xmath10 for each @xmath11 and any open and bounded @xmath12 .
the malliavin calculus is used in @xcite .
the aim of our paper is to find a natural representation of the derivative @xmath13 if @xmath5 is discontinuous .
we suppose that for @xmath14 , @xmath15 is a function of bounded variation on @xmath4 , i.e. for each @xmath16 the generalized derivative @xmath17 is a signed measure on @xmath4 .
let @xmath18 be measures from the hahn - jordan decomposition @xmath19 .
denote @xmath20 .
assume that for all @xmath21 @xmath22 is a measure of kato s class , i.e. @xmath23 the condition we impose on the drift is more restrictive than that of @xcite , but it allows us to obtain a representation for the derivative in terms of intrinsic parameters of the initial equation ( see theorem [ theorem_main ] ) .
our methods are different from those used in the papers cited above .
we show that the derivative @xmath24 in @xmath25 is a solution of the following integral equation @xmath26 where @xmath27 is a continuous additive functional of the process @xmath28 , which is equal to @xmath29 if @xmath5 is differentiable , @xmath30 is the @xmath0-dimensional identity matrix .
this representation is a natural generalization of the expression for the derivative in the smooth case . in the one - dimensional case ( see @xcite ) the derivative was represented via the local time of the process .
it is well known that the solution of ( [ eq_main ] ) does not have a local time at a point in the multidimensional situation .
we use continuous additive functionals for the representation of the derivative .
this method can be considered as a generalization of the local time approach to the multidimensional case .
our method can be used in the case of non - constant diffusion and .
the paper is organized as follows . in section [ section_w_functionals ]
we collect some definitions and statements concerning continuous additive functionals .
the main result of the paper is formulated in section [ section_main ] ( see theorem [ theorem_main ] ) . for the proof we approximate equation ( [ eq_main ] ) by equations with smooth coefficients .
the definitions and properties of approximating equations are given in sections [ section_approximation ] , [ section_convergence_derivatives ] .
we prove theorem [ theorem_main ] in section [ section_proof ] .
in this section we collect some facts about continuous additive functionals which will be used in the sequel .
further information can be found in @xcite ; @xcite , ch .
68 ; @xcite , ch .
ii , 6 . let @xmath31 be a homogeneous markov process with a phase space @xmath4 ( see notations in @xcite ) .
assume that @xmath32 has continuous trajectories and the infinite life - time .
denote @xmath33 [ def_w_func ] a random function @xmath34 adapted to the filtration @xmath35 is called a continuous additive functional of the process @xmath36 if it is * non - negative ; * continuous in @xmath37 ; * homogeneous additive , i.e. for all @xmath38 @xmath39 @xmath40 where @xmath41 is the shift operator .
if additionally for each @xmath42 @xmath43 then @xmath44 is called a w - functional .
it follows from definition [ def_w_func ] that a w - functional is non - decreasing as a function of @xmath37 , and for all @xmath45 @xmath46 the function @xmath47 is called the characteristic of a @xmath48-functional @xmath49 [ proposition_uniquely_defined ] a w - functional is defined by its characteristic uniquely up to equivalence .
the following theorem states the connection between the convergence of functionals and the convergence of their characteristics .
[ theorem_convergence_characteristics ] let @xmath50 be w - functionals of the process @xmath36 and @xmath51 be their characteristics .
suppose that for each @xmath52 , a function @xmath53 satisfies the condition @xmath54 then @xmath53 is the characteristic of a w - functional @xmath55 .
moreover , @xmath56 where @xmath57 denotes the convergence in mean square ( for any initial distribution of @xmath58 ) .
[ proposition_uniform_convergence ] if for any @xmath59 the sequence of non - negative additive functionals @xmath60 of the markov process @xmath61 converges in probability to a continuous functional @xmath55 , then the convergence in probability is uniform , i.e. @xmath62}|a_{n , t}-a_t|\to 0 , \
n\to\infty , \
\mbox{in probability}.\ ] ] let @xmath63 be a non - negative bounded measurable function on @xmath4 , let the process @xmath36 has a transition probability density @xmath64 . then @xmath65 is a @xmath48-functional of the process @xmath36 and its characteristic is equal to @xmath66 where @xmath67 let a measure @xmath68 be such that @xmath69 is a function continuous in @xmath70 if we can choose a sequence of non - negative bounded continuous functions @xmath71 such that for each @xmath72 @xmath73}\sup_{x\in{\mathds{r}^d}}\left|\int_{{\mathds{r}^d}}k_t(x , y)h_n(y)dy-\int_{{\mathds{r}^d}}k_t(x , y)\nu(dy)\right|=0,\ ] ] then by theorem [ theorem_convergence_characteristics ] there exists a w - functional corresponding to the measure @xmath68 with its characteristic being equal to @xmath74 formally we will denote this functional by @xmath75 a sufficient condition for the existence of a w - functional corresponding to a given measure is stated in the following theorem . [ theorem_sufficient_condition ] let the condition @xmath76 hold .
then @xmath53 is the characteristic of a w - functional @xmath77 .
moreover , @xmath78 and the sequence of characteristics of functionals @xmath79 converges to @xmath53 in sense of the relation .
let us return to the sde ( [ eq_main ] ) .
let @xmath80 be a solution of equation ( [ eq_main ] ) with bounded measurable @xmath5 .
the transition probability density @xmath81 of the process @xmath80 satisfies the gaussian estimates ( see @xcite ) @xmath82 valid in every domain of the form @xmath83 , y\in{\mathds{r}^d } , z\in{\mathds{r}^d},$ ] where @xmath84 , @xmath85 are positive constants that depend only on @xmath86 and @xmath87 denote by @xmath88 the kernel @xmath89 built on the transition density of the wiener process , i.e. @xmath90 it is easily to see ( @xcite , ch . 8 ,
1 ) that for all @xmath91 @xmath92 where @xmath93 therefore , the kernel @xmath88 has a singularity if @xmath94 ( for @xmath95 ) and the integral @xmath96 is not well defined for all measures .
a measure @xmath68 is a measure of kato s class if @xmath97 it follows from ( [ eq_gaussian_estimates ] ) that a measure @xmath68 satisfies the condition ( [ cond_a ] ) if and only if it belongs to kato s class .
[ cond_a_equiv ] a measure @xmath68 satisfies the condition ( [ cond_a_prime ] ) if and only if @xmath98 the proof is a slight modification of that for the case of @xmath99 given in @xcite , theorem 4.5 ( see also @xcite , exercise 1 on p. 12 ) . here
@xmath100 is a non - negative borel measurable function .
we use the representation ( [ eq_character_2 ] ) in the proof .
[ example_local_time ] let @xmath101 . for each @xmath102 ,
the measure @xmath103 belongs to kato s class and corresponds to the w - functional @xmath104}\left(w_s\right)ds,\ ] ] which is called the local time of a wiener process at the point @xmath105 .
assume that @xmath68 is a measure of kato s class .
this means now that @xmath106)<\infty.$ ] then ( see @xcite , ch .
x , 2 ) the corresponding w - functional can be represented in the form @xmath107 if @xmath108 , then @xmath109 does not belong to kato s class .
this is in consistency with the well - known fact that the local time at a point for a multidimensional wiener process does not exist .
[ example_w_bound_function ] if @xmath110 where @xmath100 is a non - negative bounded function , then @xmath68 is a measure of kato s class and @xmath111 .
[ example_w_manifold ] let @xmath112 be a compact @xmath113-dimensional @xmath114-manifold .
denote by @xmath115 the surface measure .
then for any non - negative bounded function @xmath100 , the measure @xmath116 belongs to kato s class .
[ example_w_hausd_meas ] let @xmath117 assume that a measure @xmath68 is such that @xmath118:\ \ \ \
\nu(b(x,\rho))_\leq k\rho^{d-2+\gamma}.\ ] ] then ( c.f .
@xcite , 2 ) @xmath119:\ \ \ \
\int_{b(x,\rho)}|x - y|^{2-d}\nu(dy)\leq ck\rho^\gamma.\ ] ] this inequality together with remark [ cond_a_equiv ] yields that @xmath68 is a measure of kato s class . in particular ,
the hausdorff measure on the sierpinski carpet in @xmath120 is such a measure ( see @xcite , example 2.2 ) .
we will need the uniform estimates on the moments of a w - functional .
[ prop_gikhman_skorokhod ] for all @xmath121 @xmath122 making use of this proposition one can easily obtain the following modification of khasminskii s lemma ( see @xcite or @xcite , ch.1 lemma 2.1 ) .
[ lemma_exponent_moment ] let the w - function @xmath123 satisfies the condition ( [ cond_a ] ) .
let @xmath55 be the corresponding w - functional .
then for all @xmath124 @xmath125 , there exists a constant @xmath126 depending on @xmath127 and @xmath128 such that for all @xmath39 @xmath129 by the definition of a w - functional , @xmath77 is measurable w.r.t .
the @xmath130-algebra generated by the markov process .
since we have assumed that all the processes are continuous and have the infinite life - times , we may assume that @xmath131 is a measurable function defined on @xmath132 that depends only on behavior of functions on @xmath133 $ ] ( if there is no misunderstanding , sometimes we will consider @xmath77 as a function on @xmath134,{\mathds{r}^d})$ ] ) .
let @xmath135 be a solution of defined on a probability space @xmath136 .
by @xmath137 denote the distribution of the process @xmath28 . in dynkin
s notation @xcite @xmath138 is called a markov family of random functions ( the measures @xmath139 are measures on the space of continuous functions , the measure @xmath140 is a probability on @xmath141 ) .
the composition @xmath142 is an additive functional of @xmath143 corresponding to the measure @xmath144 note that @xmath145 is defined on @xmath136 for any @xmath45 .
if the measure @xmath68 belongs to kato s class , then the corresponding additive functionals of @xmath80 and the wiener process are well defined .
denote the corresponding measurable mappings by @xmath146 and @xmath147 . by the girsanov theorem , for each @xmath45 , the distributions of the processes @xmath28 and @xmath148 are equivalent .
the question naturally arises whether the mappings @xmath146 and @xmath147 are the same .
the answer is positive and it is formulated in the next lemma .
[ lemma_equiv_w_functionals ] let @xmath68 be a measure of kato s class . then for any @xmath45 , @xmath149 for @xmath45 , denote by @xmath150 the process @xmath148 . according to theorem [ theorem_sufficient_condition ] , @xmath151 then by the girsanov theorem , @xmath152 where @xmath153 means the limit in probability .
it remains to show that the characteristics of @xmath154 converge uniformly to @xmath155 as @xmath156 ( see theorem [ theorem_convergence_characteristics ] ) .
this proof is routine and technical , so we postpone it to appendix .
let @xmath5 be a bounded measurable function of bounded variation . denote by @xmath157 the matrix @xmath158 and for @xmath159 , by @xmath160 the signed measure @xmath161 . let @xmath19 be the hahn - jordan decomposition of @xmath160 . further on we suppose that for all @xmath21 the measure @xmath20 belongs to kato s class . by theorem [ theorem_sufficient_condition ] , there exist @xmath48-functionals @xmath162 ( we will denote the corresponding mappings by @xmath163 ) with their characteristics defined according to the formula @xmath164 denote @xmath165 [ remark_hj_dec ] assume that the measure @xmath160 can be represented in the form @xmath166 , where @xmath167 are from kato s class and are not necessarily orthogonal
. then @xmath168 .
recall that the mappings @xmath169 are continuous and monotonous in @xmath37 .
so the function @xmath170 is a continuous function of bounded variation on @xmath171 $ ] almost surely .
the main result on differentiability of a flow generated by equation ( [ eq_main ] ) with respect to the initial conditions is given in the following theorem .
[ theorem_main ] let @xmath172 be such that for all @xmath173 @xmath15 is a function of bounded variation on @xmath4 .
put @xmath174 .
assume that the measures @xmath175 belong to kato s class .
let @xmath176 , be a solution to the integral equation @xmath177 where @xmath30 is the @xmath178-identity matrix , the integral on the right - hand side of ( [ eq_derivative_main ] ) is the lebesgue - stieltjes integral with respect to the continuous function of bounded variation @xmath179 .
then @xmath24 is the derivative of @xmath180 in @xmath181-sense : for all @xmath182 , @xmath45 , @xmath183 , @xmath125 , @xmath184 where @xmath185 is a norm in the space @xmath186 .
moreover , @xmath187 where @xmath188 is the lebesgue measure on @xmath4 . the differentiability was proved in @xcite .
we give a representation for the derivative .
note that the sobolev derivative is defined up to the lebesgue null set .
consider the non - homogeneous sde @xmath189 similarly to the arguments given in section [ section_w_functionals ] a theory of non - homogeneous additive functionals of non - homogeneous markov processes can be constructed .
all the formulations and proofs can be literally rewritten with natural necessary modifications .
unfortunately , there are no corresponding references , therefore we did not carry out the corresponding reasonings . consider examples of functions @xmath5 for which @xmath175 are measures of kato s class . [ remark_cond_for_smooth_functions ]
let for all @xmath190 , @xmath191 be a lipschitz function . by rademacher s theorem @xcite the frecht derivatives
@xmath17 exist almost surely w.r.t .
the lebesgue measure .
it is easy to verify that they are bounded and the frecht derivative coincides with the derivative considered in the generalized sense .
then @xmath22 belongs to kato s class .
let now @xmath192 be a bounded domain in @xmath4 with @xmath114 boundary @xmath193 .
put @xmath194 .
it follows from example [ example_w_manifold ] that for all @xmath195 @xmath22 is a measure of kato s class because ( cf .
@xcite ) @xmath196 where @xmath197 is the outward unit normal vector at the point @xmath198 condition ( [ cond_a_prime ] ) is also satisfied by the measure generated by @xmath5 being a linear combination of the form @xmath199 where @xmath200 , @xmath201 @xmath202 is a bounded domain in @xmath4 with @xmath114 boundary .
further examples of @xmath5 can be obtained as the limits of sequences of the functions of form ( [ eq_func_a ] ) . in one - dimensional case
all the functions of bounded variation generate measures belonginig to kato s class ( see example [ example_local_time ] ) .
see also example [ example_w_hausd_meas ] showing that if @xmath22 are `` hausdorff - type '' measures with a parameter greater than @xmath113 , then @xmath5 satisfies assumptions of the theorem .
the idea of the proof of theorem [ theorem_main ] is to approximate the solution of equation ( [ eq_main ] ) by solutions of sdes with smooth coefficients .
the definition and properties of approximating equations are given in sections [ section_approximation ] , [ section_convergence_derivatives ] .
the proof of the theorem itself is presented in section [ section_proof ] .
for @xmath203 let @xmath204 be a non - negative function such that @xmath205 , and @xmath206 .
put @xmath207 where the function @xmath5 satisfies the assumptions of theorem [ theorem_main ] .
note that @xmath208 and @xmath209 in @xmath210 passing to subsequences we may assume without loss of generality that @xmath211 for almost all @xmath25 w.r.t .
the lebesgue measure .
consider the sde @xmath212 put @xmath213 .
denote by @xmath214 the matrix of derivatives of @xmath215 in @xmath25 , i.e. , @xmath216 then @xmath214 satisfies the equation @xmath217 where @xmath30 is the @xmath0-dimensional identity matrix .
[ lemma_converg_solutions ] _ for each @xmath218 , _ 1 . for all @xmath125 and any compact set @xmath219 , @xmath220 2 . for all @xmath221 @xmath222 where @xmath185 is a norm in the space @xmath4
statement 1 ) follows from the uniform boundedness of the coefficients and the finiteness of the moments of a wiener process ; 2 ) is proved in @xcite , theorem 3.4 . for @xmath21 put @xmath223 by the properties of convolution of a generalized function ( see @xcite , ch .
2 , 7 ) , @xmath224 for each @xmath225 , @xmath21 put @xmath226 and @xmath227 ( c.f .
remark [ remark_hj_dec ] ) .
then , according to theorem [ theorem_sufficient_condition ] , there exist w - functionals @xmath228 of a wiener process on @xmath4 which correspond to the measures @xmath229 and have characteristics of the form @xmath230 the functional @xmath231 is given by the formula @xmath232 ( see example 2 ) . [ lemma_converg_w_functionals ]
_ for each _
@xmath84 , @xmath45 , @xmath233 , @xmath234 , @xmath235 where @xmath236 is the distribution of the process @xmath237 the following simple proposition used for the proof of lemma [ lemma_converg_w_functionals ] is easily checked .
[ proposition_convolution_characteristics ] let @xmath238 be from the kato class , @xmath239 be the characteristics of the corresponding w - functionals of a wiener process , and the representation @xmath240 hold true
. then the relation @xmath241 is fulfilled . to prove the convergence of functionals in mean square
it is sufficient to show that for each @xmath84 , @xmath242 @xmath243 ( see theorem [ theorem_convergence_characteristics ] ) . then the uniform convergence in probability follows from proposition [ proposition_uniform_convergence ] . for each @xmath244 @xmath245
where @xmath246 @xmath247 we have @xmath248 because of the condition ( [ cond_a_prime ] ) , for each @xmath233 , we can choose @xmath249 so small that @xmath250 is less then @xmath251 . to obtain the same estimate for @xmath252 , note that by the associative , distributive and commutative properties of convolution ( see @xcite , ch .
ii , 7 ) , @xmath253 we get @xmath254 consider @xmath255 .
the functions @xmath256 are equicontinuous in @xmath25 for @xmath257 $ ] .
we have @xmath258 then there exists @xmath259 such that for all @xmath260 , @xmath261 .
[ lemma_converg_functionals ] _ for each _
@xmath262 , @xmath233 , @xmath234 , @xmath263 for the proof we make use of the following proposition [ proposition_kulik ] _ let @xmath264 be complete separable metric spaces , @xmath265 be a probability space .
let measurable mappings @xmath266 @xmath267 , @xmath268 , be such that _ 1 .
@xmath269 in probability ; 2 .
@xmath270 in measure @xmath68 , where @xmath68 is a probability measure on x ; 3 .
for all @xmath271 the distribution @xmath272 of @xmath273 is absolutely continuous w.r.t .
the measure @xmath68 ; 4 .
the sequence of densities @xmath274 is uniformly integrable w.r.t .
the measure @xmath68 .
then @xmath275 in probability
. the proof can be found , for example , in @xcite , corollary 9.9.11 or @xcite , lemma 2 .
fix @xmath84 and @xmath45 .
since @xmath276 are measurable functions of a wiener process , we may assume without loss of generality that @xmath277 , { \mathds{r}^d}),$ ] @xmath278 , @xmath279 is the wiener measure , and put @xmath280 @xmath281 , @xmath282 is the distribution of the process @xmath283 ,
@xmath284,{\mathds{r}^d})$ ] , @xmath285)$ ] , @xmath286 , @xmath287 .
then @xmath288 is a sequence of random elements in the space @xmath289 taking values on @xmath290,{\mathds{r}^d})$ ] .
lemma [ lemma_converg_solutions ] entails the convergence @xmath269 in probability @xmath140 uniformly in @xmath83 $ ] .
this implies the first assertion of proposition [ proposition_kulik ] . according to lemma [ lemma_converg_w_functionals ] , @xmath291 as @xmath292 in probability measure @xmath236 uniformly in @xmath83.$ ]
this means that @xmath293 as elements of @xmath290)$ ] in measure @xmath236 .
so the second assertion of proposition [ proposition_kulik ] is justified .
the absolute continuity of the distribution of @xmath294 w.r.t .
the measure @xmath236 follows from girsanov s theorem .
the density is defined by the formula @xmath295 as @xmath296 where @xmath185 is a norm in @xmath4 , we have that for each @xmath297 , @xmath298 ( cf .
@xcite , theorem 6.1 ) .
the uniform integrability of the family @xmath299 follows from the estimate @xmath300 valid for @xmath297 .
thus all the assertions of proposition [ proposition_kulik ] are fulfilled and we have @xmath301 in probability @xmath140 . the lemma is proved .
recall that @xmath302 @xmath303 are the solutions of equations ( [ eq_derivative_main ] ) , ( [ eq_derivative ] ) , respectively . in this section
we show the convergence of the sequence @xmath304 in probability uniformly in @xmath37 .
this together with lemma [ lemma_converg_solutions ] allow us to prove theorem [ theorem_main ] .
[ lemma_converg_derivatives ] 1 . for all @xmath305 , @xmath45 , @xmath182 , @xmath306 2 . for all @xmath307 @xmath308 where @xmath309 for the proof we need the following two propositions
the first one is a version of the gronwall - bellman inequality and can be obtained by a standard argument .
[ prop_gronwall_lemma ] let @xmath310 be a continuous function on @xmath311 , @xmath312 be a non - negative continuous function on @xmath311 , @xmath313 be a non - negative , non - decreasing function , and @xmath314 .
if for all @xmath315 , @xmath316 then @xmath317 [ proposition_moments_a ] for all @xmath125 , @xmath124 @xmath21 there exists a constant @xmath126 such that @xmath318 the statement of the proposition follows from lemma [ lemma_exponent_moment ] and the inequalities ( [ eq_gaussian_estimates ] ) , which allow us to obtain the estimates uniform in @xmath319 . for all @xmath52 , define the variation of @xmath320 on @xmath133 $ ] by @xmath321 and put @xmath322 the variations of @xmath323 are defined similarly . _
the proof of 1)_. we have @xmath324 making use of the gronwall - bellman lemma we get @xmath325 the statement 1 ) follows now from the estimate ( [ eq_estimate_derivative ] ) and proposition [ proposition_moments_a ] . _
the proof of 2)_. we have @xmath326 by proposition [ prop_gronwall_lemma ] , @xmath327 to estimate the right - hand side of ( [ eq_derivative_exponent ] ) we make use of the following proposition .
[ proposition_monot_convergence ] let @xmath328 be a sequence of continuous monotonic functions on @xmath171 $ ] , and @xmath329).$ ] suppose that for each @xmath83,$ ] @xmath330 as @xmath331 then @xmath332}\left|\int_0^t f(s)dg_n(s)-\int_0^t f(s)dg(s)\right|\to 0 , \
n\to\infty.\ ] ] we get @xmath333 consider the first summand in the right - hand side of ( [ eq_hhhh ] ) .
put @xmath334 , @xmath335 and @xmath336 then lemma [ lemma_converg_functionals ] , proposition [ proposition_moments_a ] , and proposition [ proposition_monot_convergence ] provide that @xmath337 in probability .
similarly it is proved that the second summand in the right - hand side of ( [ eq_hhhh ] ) tends to @xmath338 as @xmath339 .
this and statement 1 ) entail statement 2 ) of the lemma .
define approximating equations by ( [ eq_main_n ] ) , where @xmath340 are determined by ( [ eq_a_n ] ) . from lemma [ lemma_converg_solutions ] and
the dominated convergence theorem we get the relation @xmath341}\int_u|\varphi_{n , t}^i(x)-\varphi_t^i(x)|^pdx\to 0 , \ n\to \infty,\ ] ] valid for any bounded domain @xmath342 , @xmath84 , @xmath343 and @xmath344 so for each @xmath173 there exists a subsequence @xmath345 such that @xmath332}\int_u|\varphi_{n_k^i , t}^i(x)-\varphi_{t}^i(x)|^pdx\to 0 \ \mbox{a.s . as }
\ k\to \infty.\ ] ] without loss of generality we can suppose that @xmath346}\int_u|\varphi_{n , t}^i(x)-\varphi_{t}^i(x)|^pdx\to 0 \ \mbox{a.s .
n\to \infty.\ ] ] arguing similarly and taking into account lemma [ lemma_converg_derivatives ] we arrive at the relation @xmath347}\int_u|y_{n , t}^{ij}(x)-y_{t}^{ij}(x)|^pdx\to 0 , \
n\to\infty , \ \mbox{almost surely},\ ] ] which is fulfilled for all @xmath21 @xmath348 since the sobolev space is a banach space , the relations ( [ eq_convergence_solutions ] ) , ( [ eq_convergence_derivatives ] ) mean that @xmath24 is the matrix of derivatives of the solution to ( [ eq_main ] ) .
let us verify .
we have for all @xmath349 @xmath350 it follows from lemmas [ lemma_converg_solutions ] and [ lemma_converg_derivatives ] that @xmath351 the following lemma implies the relation @xmath352 in probability and hence in all @xmath181 .
this completes the proof of the theorem , as and implies .
let @xmath68 be a measure of kato s class .
then for any @xmath125 , @xmath353 , @xmath233 , @xmath354 for @xmath355 , denote by @xmath356 the shift of the measure @xmath68 by the vector @xmath357 , i.e. for each @xmath358 , @xmath359 then @xmath360 note that for fixed @xmath25 and @xmath361 the process @xmath362 can be considered as a markov process starting from @xmath361 , and its distribution is equivalent to the distribution @xmath363 of the wiener process starting from @xmath361 .
indeed , @xmath364 where @xmath365 similarly to the proof of lemma 5 it can be checked that the family of the radon - nikodym densities @xmath366 are uniformly integrable with respect to @xmath367 . by proposition [ lemma_converg_solutions ] and lemma [ proposition_kulik ] to prove ( [ eq_continuity_w_functionals ] ) it suffices to verify that @xmath368 by @xmath369 denote the restriction of the measure @xmath68 to the ball @xmath370 put @xmath371 , @xmath372 .
then @xmath373 it is easy to see that the function @xmath374 is uniformly continuous in @xmath375\times { \mathds{r}^d}.$ ] so by theorem [ theorem_convergence_characteristics ] we have the convergence in probability @xmath376 for any @xmath377 it follows from @xcite , theorem 8.4 that for any @xmath378 and @xmath102 we have the equality @xmath379 a.s . on the set @xmath380}|y+w_s|<r\}.$ ]
this together with entails ( [ eq_convergence_w_functionals_2 ] ) .
note that @xmath381 is a w - functional .
let us estimate its characteristic .
@xmath382 from the estimates ( [ eq_gaussian_estimates ] ) we obtain ( see also the proof of theorem 6.6 . in @xcite )
@xmath383 where @xmath384 @xmath385 .
taking into account ( [ eq_additive ] ) , we get @xmath386 by proposition [ prop_gikhman_skorokhod ] , @xmath387 therefore , the second moment of @xmath388 is bounded uniformly in @xmath63 .
this implies the uniform integrability and , consequently the convergence in @xmath389 holds in ( [ eq_girsanov_theorem ] ) .
then the characteristic of the functional @xmath390 is equal to @xmath391 if we show that @xmath392 then the statement of the lemma follows from proposition [ proposition_uniquely_defined ] .
we have , for each @xmath393 , @xmath394 consider @xmath395 . arguing as in the proof of ( [ eq_similar ] ) we arrive at the inequality @xmath396 making use of ( [ eq_gaussian_estimates ] ) and changing the variables
we get @xmath397 for each @xmath233 , the condition ( [ cond_a_prime ] ) allows us to choose @xmath249 so small that @xmath398 further , @xmath399 the measure @xmath400 converges weakly to the @xmath249-measure at the point @xmath105 .
the function @xmath401 is equicontinuous in @xmath105 for @xmath402 , \
x\in{\mathds{r}^d}$ ] .
so @xmath403 uniformly in @xmath25 and @xmath404 .
besides , from ( [ eq_gaussian_estimates ] ) @xmath405 by the dominated convergence theorem , @xmath406 now the equality ( [ eq_f_tilde ] ) follows from ( [ eq_i_ii ] ) and ( [ eq_iii ] ) .
the lemma is proved .
the authors thank prof . le jan and prof .
kulik for fruitful discussions .
we are appreciate prof .
portenko for his useful remarks to the manuscript .
we also are grateful to the anonymous referee for his thorough reading and valuable comments which helped to improve essentially the exposition . | we consider a @xmath0-dimensional sde with an identity diffusion matrix and a drift vector being a vector function of bounded variation .
we give a representation for the derivative of the solution with respect to the initial data . |
at present , it is unknown whether the neutrinos are massive dirac particles or massive majorana particles .
if neutrinoless double beta decay ( @xmath0 ) is caused by the exchange of three light majorana neutrinos , the observation of @xmath0 is the evidence that the neutrinos are majorana fermions . in the process of neutrinoless double beta decay , @xmath1
the lepton number is not conserved .
neutrinoless double beta decay has not been observed yet .
many experiments of @xmath2 are in progress and planed : candles @xcite , nemo-3 @xcite , solotvino @xcite , cuoricino @xcite , exo-200 @xcite , kamland - zen @xcite , etc .
if neutrinos are majorana fermions , lepton mixing matrix ( mns matrix @xcite ) @xmath3 is represented by six parameters : three lepton mixing angles @xmath4 , cp - violating dirac phase @xmath5 , and two cp - violating majorana phases @xmath6 .
these @xmath7 and @xmath8 are the degrees of freedom of phase which come from the assumption that neutrinos are majorana fermions @xcite . among these six parameters in the lepton mixing matrix , three parameters @xmath9 , @xmath10 , and @xmath11
are measured by experiments @xcite .
the measurement of @xmath11 especially was done recently @xcite , whose effect on the study of @xmath2 had been investigated @xcite .
dirac phase @xmath5 and two majorana phases @xmath7 and @xmath12 are unknown by experiments . if one assumes that neutrinoless double beta decay is caused by the exchange of three light majorana neutrinos , the amplitude of @xmath2 is proportional to the effective majorana mass @xmath13 , @xmath14 where @xmath15 , and @xmath16 are mass of three neutrinos , and @xmath3 is the lepton mixing matrix .
there are two possibilities of neutrino mass spectrum : the normal mass ordering @xmath17 and the inverted mass ordering @xmath18 .
the effective majorana mass @xmath13 depends on the seven parameters : mixing angles @xmath19 , neutrino mass @xmath20 , and majorana phases @xmath21 . as regarding these seven parameters , mass squared differences @xmath22 and @xmath23 , @xmath9 , and @xmath24 are measured by experiments @xcite , whereas the absolute neutrino mass scale and majorana phases @xmath25 ( if majorana particles ) are not measured . while information on majorana phases @xmath25 is obtained by neutrinoless double beta decay experiments @xcite , it is notable that the other experiments measuring the absolute neutrino mass scale are also important to determine the phases @xmath25 as discussed in ref.@xcite . in studying neutrinoless double beta decay ,
people often use the method in which one regards the lightest neutrino mass as a free parameter and analyzes the effective majorana mass @xmath26 for each given value of the lightest neutrino mass .
the effective majorana mass @xmath26 is considered to be a function of two variables @xmath12 and @xmath27 provided that the value of the lightest neutrino mass is given by hand . even if the value of @xmath26 is obtained by @xmath2 experiments
, we can not determine both values of @xmath12 and @xmath27 simultaneously for a given value of the lightest neutrino mass . instead
, we can obtain the constraint between the majorana phases , @xmath12 and @xmath27 .
before @xmath28 was measured by the experiments @xcite , the constraint between @xmath12 and @xmath27 had been examined for given values of @xmath26 and neutrino mass @xmath29 by numerical calculations @xcite . in the numerical calculations ,
the authors gave some values to the parameters , @xmath30 , @xmath31 , and the lightest neutrino mass , respectively .
in this paper , we obtain the analytic equation representing the isocontour of the effective majorana mass @xmath26 on the @xmath32plane for each of the given values of the lightest neutrino mass .
the equation of the isocontour of @xmath26 ( constraint between @xmath12 and @xmath27 ) is derived by the method of perturbation for the normal mass ordering case and the inverted mass ordering case , respectively .
the equation is represented analytically by the six parameters : @xmath30 , @xmath31 , the lightest neutrino mass scale , and of course @xmath26 . because the effective majorana mass is invariant under @xmath33 , @xmath34 the isocontour of @xmath13 around a point @xmath21 in the @xmath35plane is the same as that around the point @xmath36 . if neutrinoless double beta decay is observed , the next challenging task is to restrict the values of the majorana phases @xmath8 and/or @xmath7 . at that stage ,
the analytically expressed constraint between @xmath8 and @xmath7 is useful .
the paper is organized as follows . in section 2 ,
the notations we use are introduced , and the characteristic features of the normal mass ordering and the inverted mass ordering are described , respectively . in section 3 ,
the isocontour of the effective majorana mass @xmath13 in the @xmath32plane ( constraint between @xmath8 and @xmath7 ) is obtained in the case of the normal mass ordering . to describe more concretely ,
the isocontour of @xmath13 around the point of maximum @xmath13 in the @xmath32plane , that around the point of minimum @xmath13 , and that except around the point of maximum or minimum @xmath13 are obtained by the method of perturbation . in section 4 , the same kind of things as section 3 is described in the case of the inverted mass ordering .
section 5 is devoted to conclusions . in appendix ,
the perturbative method used in section 3 is explained .
if we assume that neutrinoless double beta decay ( @xmath2 ) is generated by the exchange of the three majorana neutrinos with light mass , the amplitude of @xmath2 is proportional to the effective majorana mass @xmath37 , @xmath38 where @xmath39 is neutrino mass of @xmath40th mass eigenstate . the unitary matrix @xmath3 is the lepton mixing matrix ( mns matrix ) and parametrized as follows @xcite , @xmath41 where @xmath42 and @xmath43 are sine and cosine of the lepton mixing angle @xmath44 , respectively .
the parameter @xmath5 is the cp - violating dirac phase , and @xmath45 represent the cp - violating majorana phases . among nine parameters , @xmath20 , @xmath46 , @xmath5 , @xmath21 ,
the following five quantities are known by experiments : i.e. , two mass squared differences @xmath47 , @xmath48 , and three mixing angles @xmath46 .
these oscillation parameters are known to be within @xmath49 at @xmath50 @xcite , respectively .
we therefore have the relations @xmath51 and @xmath52 .
the effective majorana mass @xmath53 can be written as @xmath54 where @xmath55 and @xmath56 when obtaining the isocontours of the effective majorana mass @xmath13 in the @xmath32plane , the order of three @xmath57 according to size becomes an important issue , as will be discussed in detail in section 3 and 4 .
the effective majorana mass @xmath13 is invariant under @xmath58 as well as under @xmath59 or @xmath60 .
the @xmath13 depends on seven parameters : mixing angles @xmath19 , neutrino mass @xmath20 , and cp - violating majorana phases @xmath21 .
for these seven parameters , @xmath9 , @xmath24 , @xmath61 , and @xmath62 are measured by experiments @xcite , while the majorana phases @xmath25 ( if majorana particles ) and the absolute neutrino mass scale are not measured .
recently , an upper limit for the sum of the three light neutrino mass has been reported by planck measurements of the cosmic microwave background @xcite , @xmath63 in this paper , we regard the four parameters , @xmath30 , @xmath61 , and @xmath62 as given quantities . the absolute neutrino mass scale is treated as a free parameter .
then , the effective majorana mass @xmath13 can be considered to be a function of two variables @xmath12 and @xmath7 when one sets the size of the absolute neutrino mass scale .
there remains two possibilities with respect to the order of @xmath64 , @xmath65 , and @xmath66 , that is the normal mass ordering @xmath17 and inverted mass ordering @xmath67 .
characteristic features of each mass ordering are discussed in the following . in the normal mass ordering @xmath17
, we take the lightest neutrino mass @xmath64 as the absolute neutrino mass scale , and regard it as a free parameter , @xmath68 the @xmath69 defined in eqs.([be ] ) are represented by use of the lightest neutrino mass @xmath64 , @xmath70 setting the value of the lightest neutrino mass @xmath64 , we can regard the effective majorana mass @xmath13 as the function of @xmath12 and @xmath7 . now let us consider the order of @xmath71 , @xmath72 , and @xmath73 according to size in the normal mass ordering case . for an arbitrary value of the lightest neutrino mass @xmath64 , the following inequality holds , @xmath74 by use of the experimental values @xcite , eq.([bc ] ) , one has the relation @xmath75 at @xmath76 . taking account of this , we assume in this paper that the inequality , @xmath77 is satisfied for an arbitrary value of @xmath64 .
we examine the order of @xmath71 , @xmath72 , and @xmath73 according to size when the value of the lightest neutrino mass @xmath64 is varied .
when taking the @xmath78 limit , we have @xmath79 on the other hand , for @xmath80 , it follows @xmath81 because of @xmath82 and @xmath83 .
therefore , the @xmath71 which is proportional to @xmath64 changes from zero to the maximum of @xmath84 when the value of the lightest neutrino mass @xmath64 is varied . for a concrete illustration
, we show in fig.1 the relation of the lightest neutrino mass @xmath64 and @xmath84 when the values of parameters are set as @xmath85 , @xmath86 , @xmath87 , and @xmath88 .
( solid ) , @xmath89 ( long - dashed ) , and @xmath90 ( dotted ) as a function of the lightest neutrino mass @xmath64 ( in ev ) , respectively , for the normal mass ordering case.,height=264 ] after section 3 , we do not specify the values of @xmath91 , @xmath92 , and @xmath93 .
for the lightest neutrino mass @xmath64 which is regarded as a free parameter , we do not give the value of @xmath64 but specify the size of @xmath71 which is proportional to @xmath64 .
in the inverted mass ordering ( @xmath94 ) , we take the lightest neutrino mass @xmath16 as the absolute neutrino mass scale , and regard it as a free parameter , @xmath95 the @xmath69 defined in eq.([be ] ) are represented by use of the lightest neutrino mass @xmath16 , @xmath96 setting the value of the lightest neutrino mass @xmath16 , we can regard the effective majorana mass @xmath37 as a function of @xmath12 and @xmath7 .
for an arbitrary value of the lightest neutrino mass @xmath16 , we have @xmath97 in the inverted mass ordering because of @xmath98 and @xmath99 .
the later relation , @xmath99 , comes from the following inequality , @xmath100 for a concrete illustration , we show in fig.2 the relation of the lightest neutrino mass @xmath16 and @xmath84 when the values of @xmath91 , @xmath92 , and @xmath93 are the same with those of fig.1 .
( solid ) , @xmath89 ( long - dashed ) , and @xmath90 ( dotted ) as a function of the lightest neutrino mass @xmath16 ( in ev ) , respectively , for the inverted mass ordering case.,height=264 ] as discussed before , we do not specify the values of @xmath91 , @xmath92 , and @xmath93 after section 3 .
one can expect that the @xmath2 experiment will bring the information on the majorane phases @xmath12 and @xmath7 in the future . in this chapter
, we derive the equation representing the isocontour of the effective majorana mass @xmath13 in the @xmath32plane in the normal mass ordering case ( @xmath101 ) .
as discussed in section 2 , the lightest neutrino mass @xmath64 is treated as a free parameter , and when we set the value of @xmath64 , the effective majorana mass @xmath13 is considered as a function of @xmath12 and @xmath7 . if the value of @xmath13 is determined , the constraint on @xmath12 and @xmath7 ( isocontour of @xmath13 in the @xmath32plane ) is obtained ; however , one can not obtain both values of @xmath12 and @xmath7 simultaneously .
we first obtain the isocontour of @xmath13 around the point of maximum @xmath13 in the @xmath32plane , and next those around the point of minimum @xmath13 .
the point of maximum or minimum @xmath13 will be found in the region , @xmath102 in the last subsection , we look for the isocontour except around the point of maximum or minimum @xmath13 .
all these calculations are done on the assumption that the relation @xmath103 holds as discussed in section 2 . for an arbitrary value of the lightest neutrino mass @xmath64 ,
the effective majorana mass @xmath13 takes a maximum value , @xmath105 at the point @xmath106 in the @xmath32plane . in order to obtain the isocontour around @xmath106 , we expand @xmath107 around this point . since the dependence of @xmath107 on @xmath12 and @xmath7 is @xmath108 there appears no odd order term on the two variables @xmath12 and @xmath7 in that expansion .
if one includes the terms up to second order on the two variables @xmath12 and @xmath7 in the expansion and disregards the higher orders , it becomes @xmath109 or @xmath110 where @xmath111 a real symmetric matrix can be diagonalized by a rotational matrix , @xmath112 where @xmath113 if @xmath114 and @xmath115 are satisfied , one has @xmath116 .
since @xmath115 and @xmath114 are satisfied from eq.([cae ] ) , a homogeneous polynomial of second order , eq.([cad ] ) , represents the ellipse in the @xmath32plane . by diagonalization of the real symmetric matrix , eq.([cad ] ) can be rewritten as @xmath117 where @xmath118 and @xmath119 > 0 , \nonumber \\ \tan 2 \theta & = & \frac{2 \vert m_{ee}^{(2 ) } \vert \vert m_{ee}^{(3 ) } \vert } { \vert m_{ee}^{(1 ) } \vert ( \vert m_{ee}^{(2 ) } \vert - \vert m_{ee}^{(3 ) } \vert ) } > 0 , \nonumber \\
\lambda_{\pm } & = & { 1 \over 2 } \left [ \biggl\ { \vert m_{ee}^{(1 ) } \vert \vert m_{ee}^{(3 ) } \vert + 2 \vert m_{ee}^{(2 ) } \vert \vert m_{ee}^{(3 ) } \vert + \vert m_{ee}^{(1 ) } \vert \vert m_{ee}^{(2 ) } \vert \biggr\ } \right .
\nonumber \\ & & ~~ \left .
\pm \sqrt { \vert m_{ee}^{(1 ) } \vert^2 ( \vert m_{ee}^{(3 ) } \vert - \vert m_{ee}^{(2 ) } \vert ) ^2 + 4 ( \vert m_{ee}^{(2 ) } \vert \vert m_{ee}^{(3 ) } \vert ) ^2 } \ , \right ] .
\label{caj}\end{aligned}\ ] ] as @xmath120 and @xmath121 are both positive , we have @xmath122 . in the @xmath32plane ,
the isocontour around the point of maximum @xmath123 becomes the ellipse represented by the following normal equation , @xmath124 where we have neglected the fourth order of two variables @xmath125 and @xmath126 or higher orders . the center of this ellipse is @xmath106 , and the major axis is @xmath127 , the minor axis @xmath128 .
the ratio of the minor axis to the major axis , @xmath129 , depends on the lightest neutrino mass @xmath64 , but not on @xmath13 .
the direction of the @xmath130 axis is produced by a counterclockwise rotation of the @xmath12 axis by the angle @xmath131 ( see fig.3 ) .
axis ( the @xmath132 axis ) and the @xmath12 axis ( the @xmath7 axis).,height=264 ] the effective majorana mass , eq.([ab ] ) , is invariant under @xmath133 , and consequently the ellipse with the center @xmath134 is distributed in the @xmath32plane . when the size of the major axis is a quarter of @xmath135 , @xmath136 ,
the effective majorana mass becomes @xmath137 the ratio of the minor axis to the major axis @xmath129 reflects the form of the ellipse , the major axis reflects the size of the ellipse , and the angle @xmath138 determines the direction of the major axis of the ellipse .
now , we investigate how @xmath129 and @xmath138 change when the value of the lightest neutrino mass @xmath64 is varied .
we also examine the relation between the effective majorana mass @xmath13 and the major axis of the ellipse which is the isocontour of @xmath13 .
the expressions of @xmath139 or @xmath120 are , however , too complicated to study these problems practically . as discussed in section 2 , the @xmath71 changes from zero to the maximum of @xmath84 when the value of the lightest neutrino mass @xmath64 is varied .
when the value of @xmath64 ( or @xmath140 ) is in some regions , @xmath139 or @xmath120 can be expressed simply , and we give three instances in the following .
the first case is that the value of the lightest neutrino mass @xmath64 is large enough to satisfy inequality , @xmath141 owing to this inequality , the terms of order @xmath142 can be neglected , @xmath143 the ratio of the minor axis to the major axis can be written as @xmath144 where we have used @xmath145 . in the region where the condition eq.([cam ] ) is satisfied , both the angle @xmath146 and the ratio @xmath129 increase when the value of the the lightest neutrino mass @xmath64 is decreased .
when the size of the major axis of the ellipse is @xmath147 , the effective majorana mass eq.([cal ] ) becomes @xmath148 in other words , when the effective majorana mass takes the value eq.([cap ] ) , the corresponding isocontour is the ellipse represented by @xmath149 the second case is that the the lightest neutrino mass @xmath64 has smaller value than the first case and satisfies @xmath150 in this case , we have @xmath151 and both the angle @xmath146 and the ratio @xmath129 are larger than the first case , respectively .
when the size of the major axis of the ellipse is @xmath147 , the effective majorana mass becomes @xmath152 the third case is that the lightest neutrino mass @xmath64 has much smaller value than the second case and satisfies @xmath153 in this case , the terms of order @xmath154 can be neglected and one has @xmath155 or @xmath156 and @xmath157 the angle @xmath146 is larger than the second case , while the ratio @xmath129 is much smaller than the second case . in the region where the condition eq.([cau ] ) is satisfied , the angle @xmath146 increases and the ratio @xmath129 decreases when the value of the the lightest neutrino mass @xmath64 is decreased . in the @xmath78 limit , we have @xmath158 when the size of the major axis of the ellipse is @xmath147 , the effective majorana mass becomes @xmath159 the position of the point @xmath21 giving the minimum value of @xmath104 shifts according to the value of the lightest neutrino mass @xmath64 ( or @xmath71 ) , because @xmath160 changes from zero to the maximum of @xmath161 when the value of @xmath64 is varied . in studying the minimum value of @xmath13 , it is convenient to divide @xmath160 into the following three regions , * region ( a ) ; @xmath162 , * region ( b ) ; @xmath163 , * region ( c ) ; @xmath164 .
the minimum value of @xmath13 in each of the three regions becomes @xcite * region ( a ) ; @xmath165 * region ( b ) ; @xmath166 * region ( c ) ; @xmath167 in this subsection , we obtain the isocontour of @xmath13 in the @xmath32plane around the point of the minimum @xmath13 in each of the three regions ( a ) , ( b ) , and ( c ) . moreover , the characteristic features of the isocontour on the boundary between the region ( a ) and ( b ) , or between ( b ) and ( c ) , will be discussed . in the region ( a )
, the effective majorana mass has a minimum value , @xmath169 , @xmath170 at the point @xmath171 .
we expand @xmath107 around this point to obtain the isocontour around the point . in a similar way to the discussion in section3.1
, there appears no odd order term on the two variables @xmath172 and @xmath173 in this expansion around @xmath174 .
up to the second order , the expansion results @xmath175 or @xmath176 where @xmath177 @xmath178 and one has @xmath179 . by diagonalizing the real symmetric matrix as discussed in section 3.1 , we can rewrite eq.([cbac ] ) as @xmath180 where @xmath181 and @xmath182 < 0 , \nonumber \\
\tan 2 \theta & = & - \frac{2 \vert m_{ee}^{(2 ) } \vert \vert m_{ee}^{(3 ) } \vert } { \vert m_{ee}^{(1 ) } \vert ( \vert m_{ee}^{(2 ) } \vert - \vert m_{ee}^{(3 ) } \vert ) } < 0 , \nonumber \\
\lambda_{\pm } & = & { 1 \over 2 } \left [ \biggl\ { \vert m_{ee}^{(1 ) } \vert \vert m_{ee}^{(3 ) } \vert - 2 \vert m_{ee}^{(2 ) } \vert \vert m_{ee}^{(3 ) } \vert + \vert m_{ee}^{(1 ) } \vert \vert m_{ee}^{(2 ) } \vert \biggr\ } \right .
\nonumber \\ & & ~~ \left .
\pm \sqrt { \vert m_{ee}^{(1 ) } \vert^2 ( \vert m_{ee}^{(3 ) } \vert - \vert m_{ee}^{(2 ) } \vert ) ^2 + 4 ( \vert m_{ee}^{(2 ) } \vert \vert m_{ee}^{(3 ) } \vert ) ^2 } \ , \right ] .
\label{cbah}\end{aligned}\ ] ] because of @xmath183 and @xmath184 , one has @xmath185 . in the region ( a ) ,
@xmath186 , the isocontour around the point of the minimum @xmath187 becomes the ellipse represented by the following equation , @xmath188 where the fourth order of two variables @xmath189 and@xmath173 or higher orders have been neglected .
this ellipse has the center at @xmath190 , and the major axis , @xmath191 , the minor axis , @xmath192 .
the direction of the @xmath130 axis is produced by a clockwise rotation of the @xmath12 axis by the angle @xmath193 . from the invariance of @xmath13 , eq.([ab ] ) , the ellipse with the center @xmath194 is distributed in the @xmath32plane . when the size of the major axis is a quarter of @xmath135 , @xmath195 , the effective majorana mass becomes @xmath196 as considered in section 3.1 , we investigate how the ratio of the minor axis to the major axis , @xmath197 , reflecting the form of the ellipse , and the angle @xmath146 determining the direction of the major axis of the ellipse change when the value of the lightest neutrino mass @xmath64 is varied .
although the expressions of @xmath139 or @xmath120 in the region ( a ) are very complicated , these expressions are simplified when the value of @xmath64 ( or @xmath140 ) satisfies the following , @xmath198 neglecting the order @xmath142 , we have @xmath199 in the region where the condition eq.([cbaj ] ) is satisfied , both the absolute value of the angle @xmath200 and the ratio @xmath129 increases when the value of the lightest neutrino mass @xmath64 is decreased .
when the size of the major axis of the ellipse is @xmath147 , the effective majorana mass eq.([cbaix ] ) becomes @xmath201 let us focus on the boundary between the region ( a ) and region ( b ) , @xmath202 on this boundary , the effective majorana mass has a minimum value @xmath203 at @xmath204 .
when the @xmath205 is expanded around the point @xmath174 , there appears no odd order term on the two variables @xmath206 and @xmath173 in the expansion .
the expansion up to the second order takes the same form with eq.([cbab ] ) . because of @xmath207 and @xmath208 on this boundary
, one has @xmath209 and @xmath210 .
this fact shows that , within the approximation up to the second order , the isocontour around the point of of @xmath169 is not an ellipse but a straight line , @xmath211 if one takes into account the higher order terms such as @xmath212 , and @xmath213 , the isocontour would become a closed line .
+ in the region ( b ) , the effective majorana mass has a minimum value @xmath169 , @xmath215 the reason why the minimum value becomes zero @xcite is explained as follows @xcite .
the effective majorana mass @xmath13 is given by the absolute value of the sum of three complex vectors , @xmath71 , @xmath216 , and @xmath217 .
if these three complex vectors form a triangle , @xmath218 , one has @xmath219 @xcite . now , we look for the necessary condition to form a triangle using the assumption @xmath220 made in section 2 . in the case of @xmath221 ,
the necessary condition is @xmath222 , while in the case of @xmath223 , the necessary condition is @xmath224 .
therefore , the necessary condition to form a triangle is @xmath225 , and we can show that the effective majorana mass has a minimum value zero in the region ( b ) .
next , we seek the values of majorana phases @xmath226 and @xmath227 which satisfy @xmath219 @xcite .
since @xmath13 is invariant under @xmath228 , @xmath229 we restrict @xmath226 and @xmath227 within @xmath230 solving the equation , @xmath231 we obtain @xmath232 and then @xmath233 the eq.([cbcd ] ) and the assumption @xmath220 lead to @xmath234 , and then @xmath235 . from eq.([cbcc ] ) , the product of @xmath236 and @xmath237 must be negative , @xmath238 @xcite .
consequently , @xmath237 must be positive and @xmath239 .
we finally find that @xmath226 and @xmath227 satisfy eq.([cbce ] ) and eq.([cbcd ] ) , respectively , and they are within @xmath240 we expand @xmath241 around the point @xmath242 and it is approximated by the terms up to second order , @xmath243 or @xmath244 where @xmath245 @xmath246 and then @xmath114 and @xmath247 . as in section 3.1 , eq.([cbch ] ) can be rewritten as @xmath248 where @xmath249 and @xmath250 , \nonumber \\ \tan 2 \theta & = & \frac { \vert m_{ee}^{(1 ) } \vert^2 - \vert m_{ee}^{(2 ) } \vert^2 - \vert m_{ee}^{(3 ) } \vert^2 } { \vert m_{ee}^{(3 ) } \vert^2 - \vert m_{ee}^{(2 ) } \vert^2 } , \nonumber \\ \lambda_{\pm } & = & { 1 \over 2 } \biggl [ \vert m_{ee}^{(2 ) } \vert^2 + \vert m_{ee}^{(3 ) } \vert^2 \biggr .
\nonumber \\ & & \biggl .
\pm \sqrt { \vert m_{ee}^{(1 ) } \vert^4 + 2 \vert m_{ee}^{(2 ) } \vert^4 + 2 \vert m_{ee}^{(3 ) } \vert^4 - 2 \vert m_{ee}^{(1 ) } \vert^2 ( \vert m_{ee}^{(2 ) } \vert^2 + \vert m_{ee}^{(3 ) } \vert^2 ) } \ , \biggr ] .
\label{cbcm}\end{aligned}\ ] ] the sign of the angle @xmath146 depends on the magnitude of @xmath251 .
when the @xmath251 satisfies @xmath252 , the parameter @xmath253 becomes positive and we have @xmath254 . on the other hand , when the @xmath251 satisfies @xmath255 , the parameter @xmath253 becomes negative and we have @xmath256 . in the region ( b ) , @xmath214 , the isocontour around the point of the minimum @xmath257 becomes the ellipse represented by the following equation , @xmath258 where the third order of two variables @xmath259 and @xmath260 or higher orders have been neglected .
the center of this ellipse is the point @xmath242 . from the invariance of @xmath13 , eq.([ab ] ) and eq.([cbcbz ] ) , the ellipse with the center @xmath261 is distributed in the @xmath32plane .
the expressions of @xmath262 , and @xmath139 are very complicated . however ,
when the value of the lightest neutrino mass @xmath64 ( @xmath140 ) satisfies @xmath263 , three complex vectors , @xmath264 , @xmath265 , and @xmath266 form a right - angle triangle , and the expressions become very simple , @xmath267 and @xmath268 we consider the boundary between the region ( b ) and region ( c ) , @xmath269 on this boundary , the effective majorana mass has a minimum value @xmath203 at @xmath270 .
when the @xmath205 is expanded around the point @xmath271 , there appears no odd order term on the two variables @xmath125 and @xmath173 in the expansion .
the expansion up to the second order takes the same form to eq.([cbcg ] ) . because of @xmath207 and @xmath208 on this boundary
, one has @xmath272 and @xmath273 .
this fact shows that , within the approximation up to the second order , the isocontour around the point of of @xmath169 is not an ellipse but a straight line , @xmath274 if one takes into account the higher order terms such as @xmath275 , and @xmath213 , the isocontour would become a closed line . in the region ( c ) ,
the effective majorana mass has a minimum value @xmath169 , @xmath277 at the point @xmath278 .
we expand @xmath205 around this point and it is approximated by the terms up to second order , @xmath279 or @xmath280 where @xmath281 @xmath282 and then @xmath114 and @xmath247 . as in section 3.1 , eq.([cbec ] ) can be rewritten as @xmath283 where @xmath284 and @xmath285 > 0 , \nonumber \\
\tan 2 \theta & = & \frac{2 \vert m_{ee}^{(2 ) } \vert \vert m_{ee}^{(3 ) } \vert } { \vert m_{ee}^{(1 ) } \vert ( \vert m_{ee}^{(2 ) } \vert + \vert m_{ee}^{(3 ) } \vert ) } > 0 , \nonumber \\
\lambda_{\pm } & = & { 1 \over 2 } \left [ \biggl\ { - \vert m_{ee}^{(1 ) } \vert \vert m_{ee}^{(3 ) } \vert + 2 \vert m_{ee}^{(2 ) } \vert \vert m_{ee}^{(3 ) } \vert + \vert m_{ee}^{(1 ) } \vert \vert m_{ee}^{(2 ) } \vert \biggr\ } \right .
\nonumber \\ & & ~~ \left .
\pm \sqrt { \vert m_{ee}^{(1 ) } \vert^2 ( \vert m_{ee}^{(3 ) } \vert + \vert m_{ee}^{(2 ) } \vert ) ^2 + 4 ( \vert m_{ee}^{(2 ) } \vert \vert m_{ee}^{(3 ) } \vert ) ^2 } \ , \right ] .
\label{cbeh}\end{aligned}\ ] ] because of @xmath286 and @xmath287 , one has @xmath256 . in the region ( c ) , @xmath288 , the isocontour around the point of the minimum
, @xmath289 becomes the ellipse represented by the following equation , @xmath290 where the fourth order of two variables @xmath125 and @xmath173 or higher orders have been neglected .
this ellipse has a center at @xmath270 .
the direction of the @xmath130 axis is produced by a counterclockwise rotation of the @xmath12 axis by the angle @xmath291 .
although the expressions of @xmath139 or @xmath120 are very complicated , these expressions become simple when the value of the lightest neutrino mass @xmath64 ( @xmath140 ) satisfies , @xmath292 owing to this inequality , we can neglect the terms of order @xmath154 , then the angle @xmath138 and the ratio of the minor axis to the major axis @xmath129 become @xmath293 in the region where the condition eq.([cbej ] ) is satisfied , the angle @xmath146 increases and the ratio @xmath129 decreases when the value of the lightest neutrino mass @xmath64 is decreased . when the size of the major axis of the ellipse is @xmath147 , the effective majorana mass becomes @xmath294 in the @xmath78 limit , we have @xmath295 these behavior in the @xmath78 limit resembles those of eq.([caw ] ) in the case of the ellipse representing the isocontour around the point of the maximum @xmath104 . in the preceding subsections
, we have obtained the isocontour around the point of the maximum or minimum @xmath13 in the @xmath32plane .
what are the isocontours not around the point of the maximum or minimum @xmath296 ?
we shall obtain such isocontours in this subsection . in the normal mass ordering case ,
the @xmath251 changes from zero to the maximum of @xmath84 when the value of the lightest neutrino mass @xmath64 is varied .
if the values of @xmath64 ( or @xmath140 ) is in some regions , one can obtain the approximated equations analytically which represent the isocontours not around the point of @xmath297 or @xmath298 . in the following ,
we give such three regions of @xmath251 and derive the approximated equation representing the isocontour in each three cases .
the first case is that the value of the lightest neutrino mass @xmath64 ( @xmath300 ) is large enough to satisfy the following relation , @xmath301 in order to satisfy this relation , @xmath302 should be much larger than @xmath303 , @xmath304 . in this first case
, @xmath90 is much smaller than @xmath251 and @xmath89 . at first , let @xmath90 equal zero and the majorana phase @xmath305 satisfies , @xmath306 , or @xmath307 the other phase @xmath125 can not be determined at all . in practice
the @xmath308 is not zero , and we will obtain the isocontour of @xmath13 by the method of perturbation with @xmath308 , or more formally with @xmath309 .
we rewrite @xmath310 or @xmath311 where @xmath312 and @xmath313 , @xmath314 .
if @xmath315 in the right - handed side of eq.([ccad ] ) is neglected , the phase @xmath316 satisfies @xmath317 or @xmath318 .
\label{ccag}\ ] ] now , we seek the @xmath7 by the perturbation with @xmath319 , and we expand @xmath126 as @xmath320 then , it leads @xmath321 where @xmath322 and so on . owing to the definition of @xmath323
, the @xmath324 satisfies @xmath325 when the order of @xmath326 is neglected in eq([ccai ] ) , one has from eq.([ccad ] ) , @xmath327 the @xmath328 can be determined by @xmath329 and the result is @xmath330 where @xmath331 in the same manner , the coefficient , @xmath332 of the second order of @xmath326 can be obtained , @xmath333 in this expansion , it should be @xmath334 .
this shows that we can not use this expansion near @xmath335 or @xmath336 . when @xmath337 , it becomes @xmath338 , and @xmath339 .
when @xmath340 , it becomes @xmath341 , and @xmath342 . from these
, we understand that the approximation using the expansion in @xmath319 does not work near the maximum value @xmath343 or the minimum value @xmath344 .
since @xmath345 is very small in this case , we only take the first order @xmath346 , and the relation between @xmath7 and @xmath12 becomes @xmath347 the @xmath126 is the sum of the constant @xmath305 and the cosine function of @xmath12 with the period @xmath348 which function is multiplied by the small quantity .
for instance , consider the case that @xmath349 . from eq.([ccab ] ) , the effective majorana mass takes the value @xmath350 and the equation of the isocontour of this @xmath13 becomes @xmath351 where @xmath352 .
when the mass @xmath64 decreases , the factor @xmath353 in the coefficient of @xmath354 increases .
thereby , the contribution of the second term in the right - handed side of eq.([ccar ] ) becomes larger when the value of the lightest neutrino mass @xmath64 decreases .
while the constraint eq.([ccaq ] ) can not determine the value of @xmath12 , it restricts the value of @xmath7 within a rather narrow range . to see it quantitatively , the second term in the right - handed side of eq.([ccaq ] ) is estimated as @xmath355 for instance , again consider the case of @xmath349 and we have @xmath356 which is small because of @xmath357 .
when the value of the lightest neutrino mass @xmath64 decreases , the range restricting the value of @xmath7 enlarges .
the second case is that the lightest neutrino mass @xmath64 has the value satisfying @xmath359 this value of @xmath64 is not so large as that in the first case ( the proceeding sub - subsection ) . in order to satisfy the inequality , @xmath360
, the @xmath64 should be @xmath361 if one puts @xmath362 for the moment , it becomes @xmath363 although @xmath90 is smaller than @xmath160 and @xmath89 , the @xmath345 in this second case is not so small as that in the first case .
when we take into account the second order of @xmath319 in eq.([ccaq ] ) , the @xmath7 is represented as @xmath364 where the coefficient , @xmath332 , is from eq.([ccap ] ) , @xmath365 \nonumber \\ & = & { 1 \over 4 \sin ( 2 \alpha_0 ) } \biggl [ 1 - \frac { \cos ( 2 \alpha_0 ) } { 2 \sin^2 ( 2 \alpha_0 ) } \frac { \vert m_{ee } \vert^2 } { \vert m_{ee}^{(1 ) } \vert \vert m_{ee}^{(2 ) } \vert } \left\ { \cos ( 4 \beta + 2 \phi ) + 1 \right\ } \biggr .
\nonumber \\ & & \hskip2.2 cm + \biggl .
\frac { 1 } { \sin ( 2 \alpha_0 ) } \frac { \vert m_{ee } \vert } { \vert m_{ee}^{(1 ) } \vert } \left\ { \sin ( 4 \beta + \phi -2 \alpha_0 ) - \sin ( \phi + 2 \alpha_0 ) \right\ } \biggr ] .
\label{ccbd}\end{aligned}\ ] ] the coefficient @xmath366 , eq.([ccan ] ) , involves the cosine function of @xmath12 with the period @xmath367 , and the coefficient @xmath368 involves the cosine ( sine ) function of @xmath12 with the period @xmath369 .
the @xmath7 is the sum of the constant @xmath370 , the term involving the cosine function of @xmath12 with the period @xmath367 , and the other term involving the cosine ( sine ) function of @xmath12 with the period @xmath371 .
if needed , the higher orders of @xmath372 can be calculated in the same manner .
the third case is that the value of the lightest neutrino mass @xmath64 (= @xmath374 ) is small enough to satisfy the following relation , @xmath375 in this third case , @xmath251 is much smaller than @xmath376 and @xmath377 . at first , let @xmath251 equal zero , and the difference of two majorana phases @xmath378 satisfies , @xmath379 or @xcite @xmath380 in ref.@xcite , the dependence of the effective majorana mass @xmath26 on @xmath378 , eq.([cccc ] ) , was studied in detail . defining this difference @xmath378 as @xmath381 , we have @xmath382 and @xmath383 in practice , the @xmath251 is not zero , and we will obtain the isocontour of @xmath13 by the method of perturbation with @xmath251 , or more formally with @xmath384 , which is discussed in appendix .
when one takes up to the first order of @xmath385 , the relation between @xmath7 and @xmath12 becomes ( see eq.([op ] ) in appendix ) , @xmath386 where @xmath387 the @xmath7 is the sum of the constant @xmath381 , the @xmath12 , and the cosine function of @xmath12 with the period @xmath348 which function is multiplied by the small quantity . as discussed in appendix , the approximation using the expansion in @xmath385 does not work near the maximum value @xmath343 or the minimum value @xmath344 . in subsection 3.3
, we have considered the three cases and developed the appropriate method of perturbation in each case .
however , the case , @xmath388 , has not been discussed .
unfortunately , we are not able to find a useful method of approximation in this case .
in this chapter , we derive the equation representing the isocontour of the effective majorana mass @xmath13 in the @xmath32plane in the inverted mass ordering case ( @xmath389 ) . as discussed in section 2 , the lightest neutrino mass @xmath16 is treated as a free parameter , and when we set the value of @xmath16 , the effective majorana mass @xmath13 is considered as a function of @xmath125 and @xmath126 . if the value of @xmath13 is determined , the constraint on @xmath125 and @xmath126 is obtained , yet we can not obtain both values of @xmath125 and @xmath126 simultaneously as in the case of the normal mass ordering . in the inverted mass ordering case , the order of three @xmath69 is distinctly known for an arbitrary value of @xmath16 , @xmath390 hence , we can discuss about the isocontour of @xmath13 in short compared to the normal mass ordering case . in the same manner as the normal mass ordering case , we first obtain the isocontour of @xmath13 around the point of maximum @xmath13 in the @xmath32plane , and next those around the point of minimum @xmath13 , restricting the region of @xmath125 and @xmath126 as eq.([cx ] ) . in the final subsection , we obtain the isocontour except around the point of maximum or minimum @xmath13 . for an arbitrary value of the lightest neutrino mass @xmath16 , the effective majorana mass @xmath13 takes a maximum value , @xmath391 at the point @xmath392 in the @xmath32plane .
let us expand @xmath107 around this point . including the terms up to second order on the two variables @xmath125 and @xmath126 in the expansion
, we have the same expressions as those from eq.([cac ] ) to eq.([cak ] ) in section 3 .
if one neglects the fourth order of two variables @xmath125 and @xmath126 or higher orders , the isocontour around the point of maximum @xmath123 becomes the ellipse represented by @xmath393 where @xmath394 and @xmath395 the direction of the @xmath130 axis is produced by a counterclockwise rotation of the @xmath12 axis by the angle @xmath131 .
since @xmath90 is much smaller than @xmath251 and @xmath89 , one can neglect the terms of order @xmath396 in the expressions of eq.([cag ] ) in the inverted mass ordering case .
the @xmath397 and the ratio of the minor axis to the major axis , @xmath398 , are approximated , respectively , @xmath399 where we have used @xmath99 .
we should notice here that , in the inverted ( normal ) mass ordering case , the lightest neutrino mass @xmath400 is regarded as a free parameter .
when the value of the lightest neutrino mass @xmath16 is decreased , both the angle @xmath146 and the ratio @xmath398 decrease monotonously . in the @xmath401 limit
, we have @xmath402 the behavior of the angle @xmath146 and the ratio @xmath129 by decreasing the value of the lightest neutrino mass in this subsection of the inverted mass ordering case is strikingly different from that in the subsection 3.1 of the normal mass ordering case .
when the size of the major axis of the ellipse is a quarter of @xmath135 , @xmath403 , the effective majorana mass becomes @xmath404 owing to the relation eq.([dx ] ) in the inverted mass ordering case , the effective majorana mass takes a minimum value , @xmath405 at the point @xmath406 .
let us expand @xmath107 around this point .
including the terms up to second order in the expansion , we have the same expressions as those from eq.([cbab ] ) to eq.([cbai ] ) in section 3 .
if one disregards the fourth order or higher orders , the isocontour around the point of minimum @xmath169 becomes the ellipse represented by @xmath407 where @xmath408 and @xmath409 this ellipse has the center at @xmath204 , and the direction of the @xmath130 axis is produced by a clockwise rotation of the @xmath12 axis by the angle @xmath193 .
because one can neglect the terms of order @xmath396 in the expressions of eq.([cbah ] ) in the inverted mass ordering case , the angle @xmath146 and the ratio of the minor axis to the major axis , @xmath398 , are approximated , respectively , @xmath410 where we have used @xmath411 .
when the value of the lightest neutrino mass @xmath16 is decreased , both the absolute value of the angle @xmath412 and the ratio @xmath398 decrease monotonously . in the @xmath413 limit , one has @xmath414 the behavior of the angle @xmath146 and the ration @xmath129 by decreasing the value of the lightest neutrino mass in this subsection of the inverted mass ordering case is strikingly different from that in the sub - subsection 3.2.1 of the normal mass ordering case .
when the size of the major axis of the ellipse is @xmath147 , the effective majorana mass becomes @xmath415 in this subsection , we will find the isocontour not around the point of the maximum or minimum @xmath104 .
since the relation @xmath416 holds for an arbitrary value of the lightest neutrino mass @xmath16 in the inverted mass ordering case , we can use the method of perturbation with @xmath417 , which is used in section 3.3.1 .
the relation between @xmath7 and @xmath12 is , from eq.([ccaq ] ) , @xmath418 where @xmath419 the @xmath305 is given by @xcite @xmath420 note that this perturbative calculation can not be used near the point of the maximum @xmath123 or the minimum @xmath169 as discussed in section 3.3.1 .
when the mass @xmath16 decreases , the factor @xmath421 in the coefficient of @xmath354 in eq.([dca ] ) decreases .
hence , the contribution of the second term in the right - handed side of eq.([dca ] ) becomes small when the value of the lightest neutrino mass @xmath16 decreases .
this behavior of the second term in the right - handed side of eq.([dca ] ) by decreasing the value of the lightest neutrino mass in this subsection of the inverted mass ordering case is strikingly different from that in sub - subsection 3.3.1 of the normal mass ordering case . if the @xmath401 limit is taken , one has @xmath422 and @xmath7 approaches the constant not depending on @xmath12 @xcite .
in the @xmath32plane where @xmath12 and @xmath7 are the majorana cp - violating phases , we obtained analytically the equation representing the isocontour of the effective majorana mass @xmath13 by the method of perturbation .
the effective majorana mass @xmath13 is written as @xmath423 where @xmath424 , @xmath425 , and @xmath426 .
the equation representing the isocontour of @xmath13 is expressed by the following six quantities ; @xmath13 , two lepton mixing angles @xmath427 , two neutrino mass squared differences @xmath61 ( responsible for the solar neutrino oscillation ) and @xmath62 ( responsible for the atmospheric neutrino oscillation ) , and the lightest neutrino mass which is regarded as a free parameter .
we studied how the isocontour of @xmath13 changes in the @xmath32plane when the value of the lightest neutrino mass is varied in the case of the normal mass ordering and the inverted mass ordering , respectively . in the normal mass ordering case @xmath17
, we assumed the relation @xmath428 . when the value of the lightest neutrino mass @xmath64 is varied , the @xmath251 changes from zero to the maximum of @xmath429 .
the effective majorana mass @xmath13 has the maximum value at the point @xmath392 , and the isocontour of @xmath13 around this point is the ellipse whose major axis is tilted in the counterclockwise direction of the @xmath12 axis on the @xmath32plane .
when the value of the lightest neutrino mass @xmath64 is varied , we examined how the form of the ellipse or the direction of the major axis changes in the @xmath32plane as discussed in section 3.1 . in finding the minimum value of @xmath13 ,
it is convenient to divide into the following three regions according to the size of @xmath430 , * region ( a ) ; @xmath162 , * region ( b ) ; @xmath163 , * region ( c ) ; @xmath164 . in the region ( a ) , the @xmath13 has the minimum value at the point @xmath406 .
the isocontour of @xmath13 around this point is the ellipse whose major axis is in the clockwise direction of the @xmath12 axis . on the boundary between the region ( a ) and region ( b ) , @xmath431
, the effective majorana mass @xmath13 takes a minimum value at the point @xmath432 . around this point ,
the isocontour of @xmath13 becomes the straight line within the approximation up to the second order . if one takes into account the higher order terms such as @xmath433 , etc .
, the isocontour would become a closed line . in the region ( b ) , the @xmath13 has the minimum value zero at the point @xmath242 where @xmath434 and @xmath234 .
the isocontour around this point is the ellipse . on the boundary between the region ( b ) and region ( c ) , @xmath435
, the @xmath13 has the minimum value at the point @xmath436 . around this point ,
the isocontour of @xmath13 becomes the straight line within the approximation up to the second order . if we take into account the higher order terms such as @xmath437 , etc .
, the isocontour would become a closed line . in the region ( c )
, the @xmath13 takes the minimum value at the point @xmath278 .
the isocontour of @xmath13 around this point is the ellipse whose major axis is in the counterclockwise direction of the @xmath12 axis .
we also found the isocontour not around the point of the maximum or minimum @xmath13 in the @xmath32plane .
the equation representing such isocontour is obtained analytically using the approximation method in the following three cases , ( a)@xmath438(c ) .
+ @xmath439 ( a ) @xmath440 .
+ in this case , the @xmath7 is the sum of the constant term not depending on @xmath12 , and the term of the cosine function of @xmath12 with the period @xmath348 which function is multiplied by the small quantity .
when the value of the lightest neutrino mass @xmath64 is decreased , the contribution of the term involving the cosine function of @xmath12 becomes larger .
while the value of @xmath12 can not be determined at all , the value of @xmath7 can be restricted within a rather narrow range .
+ @xmath439 ( b ) @xmath441 . +
the @xmath7 is the sum of the constant term not depending on @xmath12 , the term proportional to the cosine function of @xmath12 with the period @xmath348 , and the term proportional to the cosine ( sine ) function of @xmath12 with the period @xmath442 .
while the value of @xmath12 can not be determined at all , the value of @xmath7 can be restricted within a certain range . when the value of the lightest neutrino mass @xmath64 is decreased , the range restricting the value of @xmath7 widens .
+ @xmath439 ( c ) @xmath443 . + the @xmath7 is the sum of the constant term not depending on @xmath12 , the @xmath12 , and the term proportional to the cosine function of @xmath12 with the period @xmath348 . when the value of the lightest neutrino mass @xmath64 is decreased , the term involving the cosine function of @xmath12 approaches zero . both the value of @xmath12 and that of @xmath7 can be restricted within certain ranges , respectively .
a thing which has not been discussed is the case , @xmath444 , where the orders of @xmath251 , @xmath72 , and @xmath445 are equal in magnitude .
unfortunately , we can not find a useful method of approximation in this case . in the inverted
mass ordering case @xmath18 , the following relation holds for an arbitrary value of the lightest neutrino mass @xmath16 , @xmath446 the effective majorana mass @xmath13 takes the maximum value at the point @xmath392 , and the isocontour of @xmath13 around this point is the ellipse whose major axis is slightly tilted in the counterclockwise direction of the @xmath12 axis on the @xmath32plane .
when the value of the lightest neutrino mass @xmath16 is decreased , the major axis of the ellipse approaches the @xmath12 axis , and the ratio of the minor axis of the ellipse to the major axis approaches zero .
the @xmath13 has the minimum value at the point @xmath447 , and the isocontour of @xmath13 around this point is the ellipse whose major axis is slightly tilted in the clockwise direction of the @xmath12 axis on the @xmath32plane . when the value of the lightest neutrino mass @xmath16 is decreased , the major axis of the ellipse approaches the @xmath12 axis , and the ratio of the minor axis of the ellipse to the major axis approaches zero .
we also found the isocontour not around the point of the maximum or minimum @xmath13 in the @xmath32plane .
the equation representing that isocontour is such that the @xmath7 is the sum of the constant term not depending on @xmath12 , and the term of the cosine function of @xmath12 with the period @xmath348 which function is multiplied by the small quantity .
when the value of the lightest neutrino mass @xmath16 is decreased , the contribution of the term involving the cosine function of @xmath12 becomes small .
while the value of @xmath12 can not be determined at all , the value of @xmath7 can be restricted within a very narrow range .
so far the constraints between the majorana cp - violating phases @xmath7 and @xmath12 have been investigated mainly by the numerical calculations . at present
, we can not decide whether the mass of the neutrino obeys the normal mass ordering or the inverted mass ordering , and the absolute neutrino mass scale is unknown . under such circumstances remaining numerous possibilities
, we would like to see prospectively the constraints between @xmath7 and @xmath12 ( the isocontour of @xmath13 in the @xmath32plane ) . in this paper
, we obtained the equation of the constraint between @xmath7 and @xmath12 analytically by the method of perturbation , and clarified how the constraint changes in the @xmath32plane when the value of the lightest neutrino mass is varied . since the perturbative calculations are possible up to any order , one can analytically calculate according to the required accuracy in principle .
a positive signal of the neutrinoless double beta decay has not been detected in the experiments , and the upper limit of the effective majorana mass @xmath13 has been reported .
the kamland - zen experiment provides the upper limit @xcite , @xmath448 we expect that the upper limit of @xmath13 will be lowered by the future neutrinoless double beta decay experiments .
when the upper limit of @xmath13 is put at a certain value , one can exclude some parameter regions in the @xmath32plane for a given value of the lightest neutrino mass . as discussed in section 2 , we regard the other four parameters , @xmath30 , @xmath61 , and @xmath62 as given quantities in this paper .
let us assume that the value of the upper limit of @xmath13 becomes so small that the inverted mass ordering is denied @xcite . in the normal mass ordering , we consider here the case that the lightest neutrino mass @xmath64 satisfies the relation , @xmath449 . if the upper limit of @xmath13 becomes @xmath450 then the parameter region inside the ellipse , eq.([caq ] )
, @xmath451 is excluded in the @xmath32plane .
furthermore , if the upper limit of @xmath13 is lowered as @xmath452 then , from eq.([ccab ] ) and eq.([ccar ] ) , the following parameter region , @xmath453 is excluded . in the future , if the neutrinoless double beta decay is detected by experiments and we can confirm the neutrino is majorana fermion , a next challenge is to restrict the values of the majorana phases @xmath7 and @xmath12 .
* appendix *
the effective majorana mass @xmath104 is written as @xmath454 when the value of @xmath104 is given , we will obtain the constraint equation between @xmath7 and @xmath12 under the following condition , @xmath455 the effective mass @xmath104 can be rewritten as @xmath456 where @xmath457 and @xmath458 they satisfy , @xmath459 , @xmath460 , @xmath461 , and @xmath462 if we neglect @xmath463 , we write @xmath464 and @xmath465 or @xmath466 . \label{og}\ ] ] now , we will seek the @xmath467 by the perturbation with @xmath468 , and expand @xmath467 as @xmath469 then , it leads @xmath470 where @xmath471 and so on .
the @xmath472 satisfies @xmath473 when the order of @xmath474 is neglected , one has @xmath475 the @xmath476 can be determined by @xmath477 and the result is @xmath478 where @xmath479 since @xmath480 , we have @xmath481 in this expansion , it should be @xmath482 .
this shows that we can not use this expansion near @xmath483 or @xmath484 . when @xmath485 , it becomes @xmath486 , and @xmath487 . when @xmath488 , it becomes @xmath489 , and @xmath490 . from these
, we understand that the expansion in @xmath385 can not be used near the maximum value @xmath343 or the minimum value @xmath344 .
99 m. doi , t. kotani , and e. takasugi , prog .
83 ( 1985 ) , 1 . for a review , see s. m. bilenky and s. t. petcov , rev .
59 ( 1987 ) , 671 .
w. rodejohann , j. phys .
g 39 ( 2012 ) , 124008 .
s. umehara , t. kishimoto , i. ogawa et al .
c78 ( 2008 ) , 058501 .
j. argyriades et al .
( nemo-3 collaboration ) , nucl .
a847 ( 2010 ) , 168 .
f. a. danevich et al . , phys .
c68 ( 2003 ) , 035501 . c. arnaboldi et al .
95 ( 2005 ) , 142501 .
m. auger et al .
( exo collaboration ) , phys .
( 2012 ) , 032505 .
a. gando et al .
( kamland - zen collaboration ) , phys .
( 2013 ) , 062502 .
z. maki , m. nakagawa , and s. sakata , prog .
( 1962 ) , 870 . s. m. bilenky , j. hosek , and s. t. petcov , phys . lett . 94b ( 1980 ) , 495 ; m. doi et al .
lett . 102b ( 1981 ) , 323 ; j. schechter and j. w. f. valle , phys .
d23 ( 1981 ) , 1666 .
m. c. gonzales - garcia et al .
, jhep 1212 ( 2012 ) , 123 , arxiv : hep - ph/1209.3023 . y. abe et al .
( double chooz collaboration ) , phys .
( 2012 ) , 131801 ; f. p. an et al .
( daya bay collaboration ) , phys .
( 2012 ) , 171803 ; j. k. ahn et al .
( reno collaboration ) , phys .
108 ( 2012 ) , 191802 ; k. abe et al .
( t2k collaboration ) , phys .
( 2014 ) , 061802 .
m. lindner , a. merle , and w. rodejohann , phys .
d73 ( 2006 ) , 053005 .
k. matsuda , n. takeda , and t. fukuyama , phys .
d62 ( 2000 ) , 093001 .
w. rodejohann , nucl .
b597 , ( 2001 ) , 110 .
k. matsuda et al . , phys .
d64 ( 2001 ) , 013001 .
w. rodejohann , arxiv : hep - ph/0203214 .
h. minakata , h. nunokawa , and a. a. quiroga , arxiv : hep - ph/1402.6014 .
h. minakata and o. yasuda , phys . rev .
d56 ( 1997 ) , 1692 .
h. nunokawa , w. j. c. teves , and r. z. funchal , phys . rev .
d66 ( 2002 ) , 093010 .
s. dev and s. kumar , mod .
lett . a22 ( 2007 ) , 1401 .
a. abada and g. bhattacharyya , phys .
d68 ( 2003 ) , 033004 ; s. choubey and w. rodejohann , phys .
d72 ( 2005 ) , 033016 .
p. a. r. ade et al .
( planck collaboration ) , arxiv:1303.5076 .
n. haba and r. takahashi , acta phys .
b45 ( 2014 ) , 61 .
s. pascoli and s. t. petcov , phys .
d77 ( 2008 ) , 113003 . | + we assume that neutrinoless double beta decay is caused by the exchange of three light majorana neutrinos . under this assumption ,
we obtain , by the method of perturbation , the equation representing the isocontour of effective majorana mass which is the function of two cp - violating majorana phases .
the equation representing the isocontour ( constraint equation between two majorana phases ) is expressed analytically by six parameters : two lepton mixing angles , two kinds of neutrino mass squared differences , lightest neutrino mass scale , and the effective majorana mass .
we discuss how the constraint equation between two majorana phases changes when the lightest neutrino mass scale is varied .
tnct-1401 + shinji maedan + _ department of physics , tokyo national college of technology , kunugida - machi , hachioji - shi , tokyo 193 - 0997 , japan _ |
radiative nonleptonic kaon decays represent a source of information on the structure of the weak interactions at low energies .
the flavour - changing neutral current process @xmath11 , induced at one - loop level in the standard model and highly suppressed by the gim mechanism @xcite , is of particular interest .
the @xmath12 processes have been described by the chiral perturbation theory ( chpt ) @xcite ; several models predicting the form factor characterizing the dilepton invariant mass spectrum , and thus the decay rate , have been proposed @xcite .
the first observation of the @xmath13 process was made at the cern ps more than 30 years ago @xcite , followed more recently by bnl e777
@xcite and e865 @xcite measurements .
the most precise of these , e865 , based on a sample of 10300 candidates with 1.2% background , allowed a detailed analysis of the decay form factor and rate , and a test of the next - to - leading order chpt calculation @xcite . in this paper , a new precise measurement of the @xmath14 decay
is reported , based on the full data set collected in 20032004 by the na48/2 experiment at the cern sps .
in addition to spectrum and rate studies , the first observation of the @xmath8 decay allowed setting an upper limit on the charge asymmetry of @xmath13 and @xmath15 decay rates , which can be related to cp violation .
the na48/2 experiment , specifically designed for charge asymmetry measurements @xcite , uses simultaneous @xmath7 and @xmath8 beams produced by 400 gev/@xmath16 primary sps protons impinging on a beryllium target .
charged particles with momentum @xmath17 gev/@xmath16 are selected by an achromatic system of four dipole magnets with zero total deflection ( ` achromat ' ) , which splits the two beams in the vertical plane and then recombines them on a common axis .
the beams pass through momentum defining collimators and a series of four quadrupoles designed to focus the beams at the detector .
finally the two beams are again split in the vertical plane and recombined in a second achromat .
the layout of the beams and detectors is shown schematically in fig .
[ fig : beams ] .
the beams then enter the fiducial decay volume housed in a 114 m long cylindrical vacuum tank with a diameter of 1.92 m , extended to 2.4 m after the first 65 m. both beams follow the same path in the decay volume : their axes coincide within 1 mm , while the transverse size of the beams is about 1 cm . with @xmath18 protons incident on the target per sps spill of @xmath19 s duration , the positive ( negative ) beam flux at the entrance of the decay volume is @xmath20 ( @xmath21 ) particles per pulse ,
of which @xmath22 ( @xmath23 ) are @xmath7 ( @xmath8 ) .
the @xmath24 flux ratio is 1.79 .
the fraction of beam kaons decaying in the decay volume at nominal momentum is @xmath25 .
a detailed description of the na48 detector can be found in @xcite .
the decay volume is followed by a magnetic spectrometer housed in a tank filled with helium at nearly atmospheric pressure , separated from the vacuum tank by a thin ( @xmath26 ) kevlar composite window .
a thin - walled aluminium beam pipe of 16 cm outer diameter traversing the centre of the spectrometer ( and all the following detectors ) allows the undecayed beam particles and the muon halo from decays of beam pions to continue their path in vacuum .
the spectrometer consists of four drift chambers ( dch ) : dch1 , dch2 located upstream , and dch3 , dch4 downstream of a dipole magnet .
the magnet provides a horizontal transverse momentum kick @xmath27 for charged particles .
each dch is composed of eight planes of sense wires . the spatial resolution of each dch is @xmath28 m .
the nominal spectrometer momentum resolution is @xmath29 ( @xmath30 in gev/@xmath16 ) .
the magnetic spectrometer is followed by a plastic scintillator hodoscope ( hod ) used to produce fast trigger signals and to provide precise time measurements of charged particles .
the hodoscope has a regular octagonal shape with a transverse size of about 2.4 m , and consists of a plane of horizontal and a plane of vertical strip - shaped counters .
the hod is followed by a liquid krypton electromagnetic calorimeter ( lkr ) used for photon detection and particle identification .
it is an almost homogeneous ionization chamber with an active volume of 7 m@xmath31 of liquid krypton , segmented transversally into 13248 projective cells , 2@xmath322 @xmath33 each , and with no longitudinal segmentation .
the calorimeter is @xmath34 deep and has an energy resolution @xmath35 ( @xmath36 in gev ) .
spatial resolution for an isolated electromagnetic shower is @xmath37 cm ( @xmath36 in gev ) for the transverse coordinates @xmath38 and @xmath39 .
the lkr is followed by a hadronic calorimeter ( hac ) and a muon detector ( muv ) , both not used in the present analysis .
the @xmath11 rate is measured relative to the more abundant @xmath40 normalisation channel ( where @xmath41 is the so called dalitz decay ) .
the final states of the signal and normalisation channels contain identical sets of charged particles .
thus electron and pion identification efficiencies , potentially representing a significant source of systematic uncertainties , cancel in the first order .
* monte carlo simulation * in order to compute acceptances for signal , normalisation and background channels , a detailed geant3-based @xcite monte carlo ( mc ) simulation is employed , which includes full detector geometry and material description , stray magnetic fields , dch local inefficiencies and misalignment , detailed simulation of the kaon beam line , and time variations of the above throughout the running period .
* event selection * three - track vertices ( compatible with the topology of @xmath11 and @xmath40 decays ) are reconstructed by extrapolation of track segments from the upstream part of the spectrometer back into the decay volume , taking into account the measured earth s magnetic field , stray fields due to magnetization of the vacuum tank , and multiple scattering .
a large part of the selection is common to the signal and normalisation modes .
it requires the presence of a vertex satisfying the following criteria . *
vertex longitudinal position is inside fiducial decay volume : @xmath42 . *
the vertex tracks are required to be consistent in time ( within a 10 ns time window ) and consistent with the trigger time , to be in dch , hod and lkr geometric acceptance , and to have momenta in the range @xmath43 .
track separations are required to exceed 2 cm in the dch1 plane to suppress photon conversions , and to exceed 15 cm in the lkr front plane to minimize particle misidentification due to shower overlaps . *
total charge of the three tracks : @xmath44 . *
particle identification is performed using the ratio @xmath45 of energy deposition of the particle in the lkr calorimeter to its momentum measured by the spectrometer .
the vertex is required to be composed of one @xmath46 candidate ( @xmath47 ) , and a pair of oppositely charged @xmath48 candidates ( @xmath49 ) .
no discrimination of pions against muons is performed . if several vertices satisfy the above conditions , the one with the best vertex fit quality is considered .
the @xmath11 candidates are selected by applying the following criteria to the reconstructed kinematic variables . * @xmath50 momentum within the beam nominal range : @xmath51 . *
@xmath50 transverse momentum with respect to the beam trajectory ( which is precisely measured using the the concurrently acquired @xmath52 sample ) : @xmath53 .
* kinematic suppression of the main background channel @xmath40 ( and other minor backgrounds induced by @xmath54 and @xmath55 decays ) by requiring the @xmath56 mass to be above the @xmath57 mass : @xmath58 , which approximately corresponds to @xmath59 mev/@xmath60 , and leads to an unavoidable loss of @xmath61 of the signal sample .
* @xmath50 invariant mass : @xmath62 .
the signal region , asymmetric with respect to the nominal kaon mass ( fig .
[ fig : mk]a ) , includes a part of the radiative mass tail .
the lower mass limit of @xmath63 corresponds to an upper cutoff @xmath64 for the energy of a single directly undetectable soft radiative photon , and is chosen as a trade - off between making an inclusive analysis , and keeping the background reasonably small .
no restrictions are applied to the @xmath11 sample on the additional energy deposition in the lkr calorimeter , the reason being two - fold : 1 ) to avoid bias due to the radiative @xmath65 decays ( with predominantly soft photons , which makes the lkr response difficult to simulate ) ; 2 ) to decrease sensitivity to accidental activity . for the @xmath40 normalisation mode candidates ,
the presence of a lkr energy deposition cluster ( photon candidate ) satisfying the following principal criteria is required .
* reconstructed cluster energy @xmath66 gev , cluster time consistent with the vertex time , sufficient transverse separations from track impact points at the lkr plane ( @xmath67 cm , @xmath68 cm ) . * @xmath69 invariant mass compatible with a @xmath54 decay : @xmath70 mev/@xmath60 . * the same conditions on reconstructed @xmath71 total and transverse momenta as used for @xmath50 momentum in the @xmath11 selection . * @xmath71 invariant mass : @xmath72 .
( -261,177)*(a ) * ( -35,177)*(b ) * * signal sample * the reconstructed @xmath50 invariant mass spectrum is presented in fig . [ fig : mk]a .
the measured @xmath73 resolution is @xmath74 mev/@xmath60 , in agreement with mc simulation .
the @xmath56 invariant mass resolution computed by mc simulation is @xmath75 mev/@xmath60 .
the number of @xmath0 decay candidates in the signal region is @xmath76 , out of which 4613 ( 2640 ) are @xmath7 ( @xmath8 ) candidates .
the residual background sources after the kinematical suppression of the @xmath54 and @xmath77 decays are the following .
* kaon decays with a @xmath57 in the final state followed by a @xmath41 decay and particle misidentification ( @xmath48 identified as @xmath78 or vice versa ) .
two dominant contributions identified with mc simulations are ( 1 ) @xmath40 with misidentified @xmath48 and @xmath78 ; ( 2 ) @xmath79 with a misidentified @xmath48 from the @xmath54 decay .
the two contributions are of similar size , which means that about 75% of electron candidates in this sample are genuine electrons , and the rest are misidentified pions .
* kaon decays with two @xmath56 pairs in the final state resulting from either @xmath80 decays or external @xmath81 conversions , only a single lepton from each @xmath56 pair constituting the best vertex , without particle misidentification involved .
all electron candidates in this sample are genuine electrons .
backgrounds can be reliably estimated and subtracted using the data sample itself .
for the first of the above background types , the expected kinematic distribution of the lepton number violating `` same lepton sign '' @xmath82 candidates is identical to that of background events ( up to a negligible acceptance correction ) .
for the second background type , the expected sum of distributions of the same lepton sign candidates and the @xmath83 candidates with the unphysical total charge @xmath84 is similarly identical to that of background events .
background contamination is estimated using the sum of the numbers of same lepton sign candidates @xmath85 , and @xmath84 candidates @xmath86 as @xmath87 .
a cross check of the composition of same lepton sign and @xmath84 samples using a tighter electron identification ( based not only on @xmath45 but also on shower properties and track - cluster matching ) confirms the expected fractions of genuine electrons among the electron candidates . * normalisation sample * the reconstructed @xmath71 invariant mass spectrum is presented in fig .
[ fig : mk]b .
the number of @xmath40 candidates in the signal region is @xmath88 , which corresponds to the number of kaon decays in the fiducial decay volume of about @xmath89 .
the only significant background source is the semileptonic @xmath90 decay .
its contribution is not suppressed by particle identification requirements , since no @xmath46/@xmath91 separation is performed .
the background contamination in the signal region is estimated to be 0.15% by mc simulation . *
trigger chain and its efficiency * both @xmath11 and @xmath40 samples are recorded via a two - level trigger chain designed to collect the @xmath52 decays . at the first level ( l1 )
, the hod surface is logically subdivided into 16 non - overlapping square regions ; a coincidence of hits in the two planes of the hod is required to occur in at least two such regions . the second level ( l2 )
is based on a hardware system computing coordinates of hits from dch drift times , and a farm of asynchronous processors performing fast track reconstruction and running a selection algorithm , which basically requires the presence of at least two tracks consistent with a common origin in the decay volume ( closest distance of approach less than 5 cm ) .
l1 triggers not satisfying this condition are examined further and accepted nevertheless if there is a reconstructed track not kinematically compatible with a @xmath92 decay of a @xmath93 having momentum of 60 gev/@xmath16 directed along the beam axis .
the na48/2 analysis strategy for non - rare decay modes involves direct measurement of the trigger efficiencies using control data samples of downscaled low bias triggers collected simultaneously with the main triggers .
however direct measurements are not possible for the @xmath11 events due to very limited sizes of the corresponding control samples .
dedicated simulations of l1 and l2 performance were used instead .
the sources of the trigger inefficiency are the local inefficiencies of the hod and the dchs , which were mapped using special muon runs and the @xmath94 data , and included into the simulations .
the simulated efficiencies and their kinematic dependencies were compared against measurements for the abundant @xmath40 and @xmath94 decays in order to validate the simulations ; simulated values of trigger efficiencies agree with the measurements to a level of a few @xmath95 units .
the simulated values of l1 and l2 inefficiencies for the selected @xmath40 sample are @xmath96 , @xmath97 .
the values of the integral trigger inefficiencies for the @xmath11 sample depend on the a priori unknown form factor ; the corrections are applied differentially in bins of dilepton invariant mass @xmath98 .
indicative values of inefficiencies for a realistic linear form factor with a slope @xmath99 are @xmath100 , @xmath101 .
the @xmath40 sample is affected by a larger trigger inefficiency due to the smaller invariant masses @xmath98 , which means that the lepton trajectories are geometrically closer .
similarly , the inefficiency is enhanced for @xmath11 events at low @xmath98 .
* theoretical input * the decay is supposed to proceed through single virtual photon exchange , resulting in a spectrum of the @xmath3 kinematic variable sensitive to the form factor @xmath2 @xcite : @xmath102 where @xmath103 , @xmath104 , and @xmath105 . on the other hand ,
the spectrum of the angle @xmath106 between @xmath46 and @xmath107 in the @xmath56 rest frame is @xmath108 , @xmath109 , and is not sensitive to @xmath2 .
the expression for two - dimensional decay density can be found for instance in @xcite .
the following parameterizations of the form factor @xmath2 are considered in the present analysis . 1 .
linear : @xmath110 with free normalisation and slope @xmath111 . decay rate and @xmath112 spectrum are sensitive to @xmath113 , not to its sign .
next - to - leading order chpt @xcite : @xmath114 with free parameters @xmath115 and an explicitly calculated pion loop term @xmath116 given in @xcite .
3 . combined framework of chpt and large-@xmath117 qcd @xcite : the form factor is parameterized as @xmath118 with free parameters @xmath119 , see also the appendix .
4 . chpt parameterization @xcite involving meson form factors : @xmath120 .
the resonance masses ( @xmath121 , @xmath122 ) are treated as free parameters in the present analysis . the goal of the analysis is measuring the form factor parameters in the framework of each of the above models , and the computation of the corresponding branching ratios @xmath123 by integration of ( [ theory ] ) normalised to full @xmath93 decay width @xmath124 @xcite . the coulomb correction factor is taken into account following @xcite ( eq .
the effects of radiative corrections to the @xmath11 process are evaluated using the photos @xcite simulation of the @xmath125 decay chain .
the results are cross - checked using a generalized computation for a multi - body meson decay @xcite ( eq .
in addition to affecting the values of the model parameters , radiative corrections are crucial for the extrapolation of the branching ratio from the limited @xmath73 ( equivalently , @xmath126 ) signal region to the full kinematic region ; according to the above models for radiative corrections , about 6% of the total @xmath0 decay rate fall outside the signal region @xmath127 mev .
* fitting procedure * the observed spectra of the data events in @xmath112 ( in the visible region @xmath128 ) and @xmath129 variables are presented in fig .
[ fig : z ] .
the latter spectrum , which contains no information about @xmath2 , is compared to the mc expectation , demonstrating a good agreement .
( -257,184)*(a ) * ( -28,184)*(b ) * the values of @xmath130 in the centre of each @xmath131-bin of @xmath112 , which can be directly compared to the theoretical predictions ( [ theory ] ) , are then computed as @xmath132 here @xmath133 and @xmath134 are the numbers of observed @xmath14 candidates and background events in the @xmath131-th bin , @xmath135 is the number of @xmath40 events ( with background subtracted ) , @xmath136 and @xmath137 are geometrical acceptance and trigger inefficiency in the @xmath131-th bin for the signal sample ( computed by mc simulation , both presented in fig .
[ fig : acc ] ) , @xmath138 and @xmath139 are those for @xmath40 events , @xmath140 is the @xmath112 bin width , and is set to 0.02 .
the external inputs are the kaon lifetime @xmath141 s , and branching ratios of the normalisation decay modes @xmath142 , @xmath143 @xcite .
( -257,178)*(a ) * ( -30,178)*(b ) * while the simulated values of geometric acceptances and trigger efficiencies are subject to relatively large systematic uncertainties ( discussed in the following section ) , their ratios @xmath144 and @xmath145 appearing in ( [ dgdz ] ) are affected by these uncertainties to a lesser extent due to cancellations of most systematic effects .
the computed values of @xmath130 vs @xmath112 and the results of the fits to the four considered models are presented in fig .
[ fig : fit]a .
the corresponding squared form factors @xmath146 normalized by a condition @xmath147 are presented in fig .
[ fig : fit]b .
( -260,188)*(a ) * ( -45,188)*(b ) * branching ratios @xmath148 in the full kinematic range corresponding to each model are computed using the measured value of the model parameters , their statistical uncertainties , and correlation matrices . in addition , a model - independent branching ratio @xmath149 in the visible kinematic region @xmath128 is computed by integration of @xmath130 .
@xmath149 differs from each of the model - dependent brs computed in the visible @xmath112 range by less than @xmath150 .
the differences between the model - dependent brs come from the region @xmath151 , as can be seen in fig .
[ fig : fit]a . *
systematic uncertainties * * particle identification .
* imperfect mc description of electron and pion identification inefficiencies @xmath152 and @xmath153 can bias the result only due to the momentum dependence of @xmath152 and @xmath153 , due to identical charged particle composition , but differing @xmath46 and @xmath154 momentum spectra of the signal and normalisation modes .
inefficiencies were measured with the data to be in the ranges @xmath155 and @xmath156 , depending in momentum .
systematic uncertainties due to these momentum dependencies not perfectly described by mc were conservatively estimated assuming momentum - independent mc @xmath152 and @xmath153 . * beam line description . * despite the careful simulation of the beamline including time variations of its parameters , the residual discrepancies of data and mc beam geometries and spectra may bias the results . to evaluate the related systematic uncertainties , variations of the results with respect to variations of cuts on track momenta , lkr
cluster energies , total and transverse momenta of the @xmath157 final states , track distances from beam axis in dch planes , and distances between particle impact points at the lkr calorimeter surface were studied .
stability of the results in bins of decay vertex longitudinal coordinate was studied as well .
the maximum observed variations of the fit parameters are taken conservatively as the systematic uncertainties .
* background subtraction . * for background subtraction in the @xmath50 sample
the spectra of same lepton sign and @xmath84 events were used , as described above .
this method has a limited statistical precision ( with an average of @xmath158 background events in a bin of @xmath112 ) .
the uncertainties of the measured parameters due to background subtraction were conservatively taken to be equal to the corrections themselves . * trigger efficiency . *
as discussed earlier , the corrections for trigger inefficiencies were evaluated by simulations . in particular , l1 and l2 corrections for the measured br have similar magnitudes of a few @xmath159 .
no uncertainty was assigned to the l1 correction , due to relative simplicity of the trigger condition , and the consequent robustness of the simulation . on the other hand , the uncertainty of the l2 efficiency correction
was conservatively taken to be equal to the correction itself . * radiative corrections . *
uncertainties due to the radiative corrections were evaluated by variation of the lower @xmath160 invariant mass cut .
the results show reasonable stability with respect to the variation , as can be expected given a good mc description of the @xmath50 radiative mass tail visible in fig .
[ fig : mk]a .
estimation of systematic uncertainties as differences of fit results obtained with radiative corrections calculated using the photos simulation and the formalism given in @xcite leads to similar results as those obtained with the first method .
* fitting method .
* uncertainties due to the fitting procedure were evaluated by comparing the fit results obtained with a bin width @xmath161 with those obtained with the standard bin width of @xmath162 . * external input . * substantial uncertainties arise from the external input , as @xmath163 is experimentally known with a limited relative precision of 2.7% @xcite .
the only parameter not affected by this external uncertainty is the slope @xmath4 of the linear form factor describing the shape of the spectrum .
.corrections and systematic uncertainties ( excluding external sources ) . [ cols= " > ,
> , > , > , < , > , < , > , > " , ] the branching ratio in the full kinematic range , which is computed as the average between the two extremes corresponding to the models ( 1 ) and ( 4 ) , and includes an uncertainty due to extrapolation into the inaccessible region @xmath151 , is @xmath164 the uncertainty of the above result is correlated with those of the earlier measurements ( those due to external input and model - dependence uncertainties ) .
a comparison to the most precise of these measurements by the bnl e865 collaboration @xcite using the same external input , and taking into account correlation of external uncertainties between the two analyses , shows a @xmath165 level of agreement . in conclusion ,
the obtained br is in agreement with the previous measurements @xcite .
measurements of the decay parameters and brs were performed separately for @xmath7 and @xmath8 decays . in particular , the branching fractions in the full kinematic range and their statistical uncertainties were measured to be @xmath166 this allows making a first measurement of the direct cp violating asymmetry of @xmath7 and @xmath8 decay rates in the full kinematic range .
considering only the uncorrelated systematic uncertainties ( those due to background subtraction ) between @xmath7 and @xmath8 samples , @xmath167 the other systematic uncertainties are correlated between the @xmath7 and @xmath8 samples , and thus cancel in @xmath168 .
a conservative limit for the charge asymmetry of @xmath169 at 90% cl can be deduced from the above value .
this result is compatible with cp conservation , however the achieved precision is far from the standard model expectation @xmath170 @xcite and even the susy upper limit @xmath171
@xcite for the cp violating asymmetry .
from a sample of 7253 @xmath11 decay candidates with 1.0% background contamination , the branching fraction in the full kinematic range has been measured to be @xmath172 , in agreement and competitive with previous measurements .
the shape of the form factor which characterizes the decay has been evaluated in the framework of four models , giving consistent results with previous measurements .
the first simultaneous observation of both charge kaon decays into @xmath50 allowed to establish an upper limit for the cp violating asymmetry of @xmath7 and @xmath8 decay rates of @xmath173 at 90% cl .
we gratefully acknowledge the staff of the cern sps accelerator and the beamline for the excellent performance of the beam .
we also thank the technical staff of the participating laboratories , universities and affiliated computing centres for their efforts in operation of the experiment and data processing .
we are grateful to samuel friot , gino isidori and victor pervushin for valuable discussions .
the paper @xcite that formulates one of the considered models does not contain the full definition for the form factor @xmath174 : the function @xmath175 appearing in the eq .
( 3.19 ) for @xmath174 is defined only below the @xmath176 threshold .
the expressions for @xmath175 both below and above the threshold , kindly provided by the authors of ref .
@xcite , are presented below .
let us define @xmath177 .
then for @xmath178 ( below the threshold ) , @xmath179 and for @xmath180 ( above the threshold ) , @xmath181 99 s.l .
glashow , j. iliopoulos and l. maiani , phys .
d2 * ( 1970 ) 1285 .
g. ecker , a. pich and e. de rafael , nucl .
b291 * ( 1987 ) 692 .
g. dambrosio _ et al .
_ , jhep * 9808 * ( 1998 ) 4 . s. friot , d. greynat and e. de rafael , phys . lett . * b595 * ( 2004 ) 301 . a.z .
et al . _ ,
phys . part .
* 5 * ( 2008 ) 76 [ hep - ph/0611175 ] .
_ , phys . lett . *
b56 * ( 1975 ) 201 . c. alliegro
_ et al . _ ,
* 68 * ( 1992 ) 278 .
* 83 * ( 1999 ) 4482 .
et al . _ , eur .
j. * c52 * ( 2007 ) 875 .
et al . _ , .
geant detector description and simulation tool , cern program library long writeup w5013 ( 1994 ) .
e. barberio and z. was , comp .
comm . * 79 * ( 1994 ) 291 .
g. isidori , eur .
j. * c53 * ( 2008 ) 567 . c. amsler
_ et al . _ ( pdg ) , phys . lett .
* b667 * ( 2008 ) 1 .
et al . _ ,
* 84 * ( 2000 ) 2580 .
p. lichard , phys .
* d60 * ( 1999 ) 053007 .
a. messina , phys .
lett . * b538 * ( 2002 ) 130 . g. dambrosio and dao - neng gao , jhep * 0207 * ( 2002 ) 068 . | a sample of 7253 @xmath0 decay candidates with 1.0% background contamination has been collected by the na48/2 experiment at the cern sps , which allowed a precise measurement of the decay properties . the branching ratio in the full kinematic range was measured to be @xmath1 , where the uncertainty includes also the model dependence .
the shape of the form factor @xmath2 , where @xmath3 , was parameterized according to several models , and , in particular , the slope @xmath4 of the linear form factor @xmath5 was determined to be @xmath6 .
a possible cp violating asymmetry of @xmath7 and @xmath8 decay widths was investigated , and a conservative upper limit of @xmath9 at 90% cl was established .
european organization for nuclear research + cern - ph - ep-2009 - 005 + 2 march 2009 the na48/2 collaboration + j.r .
batley , a.j . culling , g. kalmus , c. lazzeroni@xmath10 , d.j .
munday , + m.w .
slater@xmath10 , s.a .
wotton + _ cavendish laboratory , university of cambridge , cambridge , cb3 0he , uk@xmath10 _
+ r. arcidiacono@xmath10 , g. bocquet , n. cabibbo@xmath10 , a. ceccucci , d. cundy@xmath10 , v. falaleev , m. fidecaro , l. gatignon , a. gonidec , w. kubischta , a. norton@xmath10 , a. maier , + m. patel , a. peters + _ cern , ch-1211 genve 23 , switzerland _ + s. balev@xmath10 , p.l .
frabetti , e. goudzovski@xmath10 , p. hristov@xmath10 , v. kekelidze , v. kozhuharov@xmath10 , l. litov , d. madigozhin , e. marinova@xmath10 , n. molokanova , i. polenkevich , + yu . potrebenikov , s. stoynev@xmath10 , a. zinchenko + _ joint institute for nuclear research , 141980 dubna , moscow region , russia _ + e. monnier@xmath10 , e. swallow , r. winston + _ the enrico fermi institute , the university of chicago , chicago , il 60126 , usa _ + p. rubin@xmath10 , a. walker + _ department of physics and astronomy , university of edinburgh , jcmb king s buildings , mayfield road , edinburgh , eh9 3jz , uk _ + w. baldini , a. cotta ramusino , p. dalpiaz , c. damiani , m. fiorini@xmath10 , a. gianoli , m. martini , f. petrucci , m. savri , m. scarpa , h. wahl + _ dipartimento di fisica delluniversit e sezione dellinfn di ferrara , i-44100 ferrara , italy _ + a. bizzeti@xmath10 , m. calvetti , e. celeghini , e. iacopini , m. lenti , f. martelli@xmath10 , g. ruggiero@xmath10 , m. veltri@xmath10 + _ dipartimento di fisica delluniversit e sezione dellinfn di firenze , i-50125 firenze , italy _ + m. behler , k. eppard , k. kleinknecht , p. marouelli , l. masetti@xmath10 , u. moosbrugger , c. morales morales , b. renk , m. wache , r. wanke , a. winhart + _ institut fr physik , universitt mainz , d-55099 mainz , germany@xmath10 _ + d. coward@xmath10 , a. dabrowski , t. fonseca martin@xmath10 , m. shieh , m. szleper , + m. velasco , m.d .
wood@xmath10 + _ department of physics and astronomy , northwestern university , evanston , il 60208 , usa _
+ g. anzivino , p. cenci , e. imbergamo , a. nappi , m. pepe , m.c .
petrucci , + m. piccini , m. raggi@xmath10 , m. valdata - nappi + _ dipartimento di fisica delluniversit e sezione dellinfn di perugia , i-06100 perugia , italy _
+ c. cerri , r. fantechi + _
sezione dellinfn di pisa , i-56100 pisa , italy _ + g. collazuol , l. dilella , g. lamanna , i. mannelli , a. michetti + _ scuola normale superiore e sezione dellinfn di pisa , i-56100 pisa , italy _ + f.
costantini , n. doble , l. fiorini@xmath10 , s. giudici , g. pierazzini , m. sozzi , s. venditti + _ dipartimento di fisica delluniversit e sezione dellinfn di pisa , i-56100 pisa italy _ + b. bloch - devaux , c. cheshkov@xmath10 , j.b .
chze , m. de beer , j. derr , g. marel , e. mazzucato , b. peyaud , b. vallage + _ dsm / irfu cea saclay , f-91191 gif - sur - yvette , france _ + m.
holder , m. ziolkowski + _ fachbereich physik , universitt siegen , d-57068 siegen , germany@xmath10 _
+ s. bifani@xmath10 , c. biino , n. cartiglia , m. clemencic@xmath10 , s. goy lopez@xmath10 , f. marchetto + _ dipartimento di fisica sperimentale delluniversit e sezione dellinfn di torino , + i-10125 torino , italy _ + h. dibon , m. jeitler , m. markytan , i. mikulec , g. neuhofer , l. widhalm + _ sterreichische akademie der wissenschaften , institut fr hochenergiephysik , + a-10560 wien , austria@xmath10 _ + _ accepted for publication in physics letters b. _ |
the program computes a fibonacci number .
a number @xmath0 is represented by @xmath0 1-s .
* sample * : 5 is represented as 1 1 1 1 1 3 is represented as 1 1 1 : number @xmath0 * sample * ( n = 7 ) : 1 1 1 1 1 1 1 : fibonacci-@xmath0 * sample * ( fibonacci-7 ) : 1 1 1 1 1 1 1 1 1 1 1 1 1
here are alphabets of states and symbols . initial state : @xmath1 ; halting state : @xmath2 ; internal states : @xmath3 , @xmath4 , @xmath5 , @xmath6 , @xmath7 , @xmath8 , @xmath9 , @xmath10 , @xmath11 , @xmath12 , @xmath13 , @xmath14 , @xmath15 , @xmath16 , @xmath17 , @xmath18 , @xmath19 , @xmath20 , @xmath21 , @xmath22 , @xmath23 , @xmath24 , @xmath25 , @xmath26 , @xmath27 , @xmath28 , @xmath29 , @xmath30 , @xmath31 , @xmath32 , @xmath33 , @xmath34 , @xmath35 , @xmath36 , @xmath37 , @xmath38 , @xmath39 , @xmath40 , @xmath41 , @xmath42 , @xmath43 , @xmath44 , @xmath45 , @xmath46 , @xmath47 , @xmath48 , @xmath49 , @xmath50 .
empty symbols alphabet : @xmath51 ; input alphabet : @xmath52 ; internal alphabet : @xmath53 , @xmath54 .
the table contains 100 rules . [ cols="^,^,^,^,^,^",options="header " , ]
c++ simulator of a turing machine has been used to compute several fibonacci numbers .
the simulator can be downloaded at | a turing machine that computes fibonacci numbers is described . |
since the discovery of the nuclear modification in the nucleon structure function @xmath4 by the european muon collaboration ( emc ) @xcite , various models have been proposed to investigate the nuclear effect .
several years ago , the new muon collaboration ( nmc ) @xcite accurately measured the structure functions @xmath5 at very small @xmath0 .
these data as well as the emc data @xcite provide an insight for investigating the physics details in the shadowing region(@xmath6 ) .
the need for testing the models of the emc effect becomes more and more important for further understanding of the nature of the nuclear effect .
bickerstaff et al .
@xcite found that although most of the theoretical models can provide good explanations for the emc effect , they can not describe the nuclear drell - yan process well .
the nmc @xcite analyzed the inelastic @xmath7 production data and obtained the gluon distribution ratios @xmath8 which indicate a modification of the gluon distribution in the nuclear environment .
kumano @xcite has made progress towards the unified description of the emc effect and the nuclear gluon distribution in the @xmath1-rescaling model with the recombination effect .
although the @xmath1-rescaling model is a useful and effective model in explaining deep - inelastic data in the valence dominated medium @xmath0 region , it is not clear whether the momentum distributions of sea quarks and gluons follow the same rescaling as that for valence quarks in the nuclear environment .
according to the qcd theory , sea quarks have different dynamic properties while interacting with gluons .
there is a constraint condition on the number conservation for valence quarks but not for sea quarks .
so it is reasonable to believe that the modification of sea quark distributions should be different from that of valence quarks . for further discussion , we assume that the transverse size of partons in a nucleon can be described by an average parton transverse momentum which is related to the longitudinal distributions of partons .
the relationship between the parton transverse momentum and longitudinal momentum , which is obtained by analysing the hadron tensor in the deep inelastic scattering process , indicates that the transverse size of partons in a nucleon depends sensitively on the effective nucleon mass .
the change of this transverse size gives rise to an additional parton evolution in the nuclear environment refer to the normal qcd parton evolution thus resulting in new parton distributions .
the probabilities of the additional evolution are directly related to the ratio of the average parton transverse momentum in the bound nucleon to that in the free nucleon . according to these probabilities ,
we give the self - consistent integral equations for parton distribution modifications due to the additional evolution in the nuclear environment .
considering the fact that the nucleon diameter is about 1.8@xmath9 , in an infinite momentum frame , one can estimate that the longitudinal size of a nucleon in a lorentz contracted nucleus is @xmath10 and the longitudinal localization size of a parton with the momentum @xmath11 is @xmath12 .
if a parton has a small @xmath0 and consequently its dimension ( @xmath13 ) exceeds the nucleon longitudinal size ( @xmath14 ) , it leaks out of the nucleon and recombines with partons of other nucleons .
we suppose that the partons participating in recombination have undergone the normal qcd evolution and the additional evolution in the nuclear environment before the fusion . after considering the additional evolution in the nuclear environment with the parton recombination effects
, we find that the new model can give a well explaination to the nuclear effects of the structure function , nuclear drell - yan process and inelastic @xmath7 production .
finally , a quantitative comparison between the ape model and the @xmath1-rescaling model is made .
partons in a free nucleon have their own distributions with a specified scale .
it is generally accepted that the transverse size of partons in a nucleon is affected by the nuclear environment .
we introduce an average transverse momentum to describe the change of the transverse size of partons . in order to obtain the average transverse momentum of partons ,
let us repeat the @xmath15 variable analysis @xcite of the hadron tensor @xmath16 in the deep inelastic scattering process shown in fig.1 .
@xmath16 can be written as @xmath17|p \rangle \nonumber \\ & = & ( -g_{\mu\nu}+\frac{q_{\mu}q_{\nu}}{q^2})w_1 + \frac{1}{m_n^2}(p_{\mu}-\frac{p\cdot q}{q^2}q_{\mu } ) ( p_{\nu}- \frac{p \cdot q}{q^2}q_{\nu})w_2\end{aligned}\ ] ] where @xmath18 is the mass of a nucleon .
if we take the frame in which @xmath19 and neglect the quark mass , then @xmath20 where @xmath21\ ] ] is the quark tensor in the lab . frame and @xmath22 is a lorentz - invariant wave function . to extract @xmath23 and @xmath24 from @xmath16 , we introduce two contractions @xmath25 @xmath26 and
then @xmath23 and @xmath24 can be expressed as @xmath27 @xmath28 if we take the frame in which @xmath29 and @xmath30 , then @xmath31 by means of eq.([tensor ] ) and the contractions defined by eq.([b ] ) and eq.([c ] ) , one can find the relation @xmath32 taking @xmath23 and @xmath24 in eq.([b ] ) and eq.([c ] ) as those in refs.@xcite and @xcite and neglecting @xmath33 and @xmath34 terms , one easily obtains @xmath35 furthermore , we assume that there is a similar relationship between the gluon transverse momentum and longitudinal momentum distributions , i.e. , @xmath36 eq.([kq ] ) and eq.([kg ] ) with the bound nucleon effective mass @xmath37 @xmath38 have been written in a unified form for partons in both a free nucleon(n ) and a bound nucleon(a ) in a nucleus with the mass number @xmath39 .
the effective mass of the bound nucleon is less than that of the free one , i.e. , in terms of eq.([kq ] ) and eq.([kg ] ) , the average transverse momentum of partons in the bound nucleon is less than that in the free nucleon , which mainly causes the parton evolution in the nuclear environment . by using the same technique in deriving the famous altarelli - parisi evolution @xcite
, we obtain the evolution probabilities for the evolution from a parton with momentum fraction @xmath40 to another parton with the momentum fraction @xmath0 ; @xmath41 ln \frac{{\langle k_{\perp}^2(x , q^2 ) \rangle}^{q_j^n } } { { \langle k_{\perp}^2(x , q^2 ) \rangle}^{q_j^a}}\ ] ] @xmath42ln \frac{{\langle k_{\perp}^2(x , q^2)\rangle}^{q_j^n } } { { \langle k_{\perp}^2(x , q^2)\rangle}^{q_j^a}}\ ] ] @xmath43 ln \frac{{\langle k_{\perp}^2(x , q^2)\rangle}^{g^n } } { { \langle k_{\perp}^2(x , q^2)\rangle}^{g^a}}\ ] ] @xmath44 @xmath45 ln \frac{{\langle k_{\perp}^2(x , q^2)\rangle}^{g^n } } { { \langle k_{\perp}^2(x , q^2)\rangle}^{g^a}}\ ] ] where the strong interaction coupling constant @xmath46 is @xmath47\ ] ] with @xmath48 and @xmath49,@xmath50 . @xmath51 and @xmath52 are color factors for @xmath53
. therefore , the quark distributions modified by the additional parton evolution(ape ) in the nuclear environment can be written as follows : @xmath54 @xmath55 @xmath56 @xmath57 @xmath58 where the input parton distributions @xmath59 and @xmath60 are taken from refs .
@xcite- @xcite .
in fact , these equations are the self - consistent integral equations since the evolution probabilities @xmath61 contain the parton transverse momenta which are related to the parton longitudinal momentum distributions @xmath62 and @xmath63 by eq.([kq ] ) and eq.([kg ] ) . in the calculation of the parton distributions in the bound nucleon
, the numerical solution of these equations is obtained by the iteration method .
although the leakage in the nuclear environment can occur for all partons , the most important contribution arises from partons with the largest spatial uncertainty , i.e. , those with the small @xmath0 .
we assume that the leak - out partons are sea quarks and gluons for which the momentum cutoff function @xcite is taken as @xmath64 namely , the distributions of leak - out(lk ) partons are @xmath65 where the input parton distributions @xmath66 , @xmath67 , @xmath68 , @xmath69 , @xmath70 , @xmath71 and @xmath72 together with @xmath73 and @xmath74 are those having undergone the normal qcd parton evolution and the additional parton evolution due to the nuclear environment .
the momentum cutoff function is shown as a function of @xmath0 in fig.2 with various @xmath75 .
as discussed in section 1 , the partons with @xmath76 can leak out of a nucleon .
we then find in fig.2 that an appropriate choice of @xmath75 is 2.0@xmath9 .
according to the original idea of the parton recombination by close , qiu and roberts @xcite , a leak - out parton can fuse with a parton from another nucleon .
in general , the modification of a parton distribution @xmath77 , due to the process of producing the parton @xmath78 with the momentum @xmath79 by fusion of partons @xmath80 and @xmath81 , is given by @xcite , @xmath82 @xmath83 where @xmath84 is given by refs .
@xcite and @xcite , @xmath85 the nuclear radius is @xmath86 with @xmath87 @xmath9 @xcite . the function @xmath88 defined by ref .
@xcite is introduced to keep the momentum conservation .
the parton fusion function @xmath89 is a probability for producing a parton @xmath78 with momentum @xmath79 by a fusion of partons @xmath80 and @xmath81 with momentum @xmath90 and @xmath91 , respectively .
it is related to a splitting function @xmath92 in the altarelli - parisi equations @xcite by @xmath93c_{{p_1}{p_2}\rightarrow p_3}.\ ] ] where @xmath94 is the ratio of color factors in the processe @xmath95 .
using the fusion functions , one can obtain the modifications of parton distributions due to the recombination(rc)@xcite . with our notations , the explicit expressions for the modifications @xmath96 and @xmath97 are @xmath98 @xmath99[1 + { [ \frac{x - x^\prime}{x}]}^2 ] \nonumber \\ & & - \frac{k}{6 } \int_0 ^ 1 \frac{dx^\prime}{x^\prime } xx^{\prime}[q^{ape}(x)g^{lk}(x^\prime)+q^{lk}(x)g^{ape}(x^\prime ) ] \frac{x}{x+x^\prime}[1 + { [ \frac{x}{x+x^\prime}]}^2 ] \nonumber \\ & & - \frac{4k}{9}x \int_0 ^ 1dx^\prime x x^\prime [ q^{ape}(x){\bar{q}}^{lk}(x^\prime)+q^{lk}(x){\bar{q}}^{ape}(x^\prime ) ] \frac{x^2+{x^\prime}^2}{{(x+x^\prime)}^4}\end{aligned}\ ] ] and @xmath100 @xmath101 \frac{1}{x^2}[\frac{x^\prime}{x - x^\prime}+ \frac{x - x^\prime}{x^\prime}+ \frac{x^\prime ( x - x^\prime)}{x^2 } ] \nonumber \\ & & - \frac{3k}{4}x \int_0 ^ 1 dx^\prime x x^\prime [ g^{lk}(x)g^{ape}(x^\prime)+g^{ape}(x)g^{lk}(x^\prime ) ] \nonumber \\ & & \times \frac{1}{(x+x^\prime)^2 } [ \frac{x}{x^\prime}+ \frac{x^\prime}{x}+ \frac{xx^\prime}{(x+x^\prime)^2 } ] \nonumber \\ & & + \frac{4k}{9}x\int_0^xd{x}^\prime \sum \limits_{i}[x^\prime q_i^{ape}(x^\prime)(x - x^\prime ) { \bar{q}}_i^{lk}(x - x^\prime ) \nonumber \\ & & + x^\prime \bar{q}_i^{ape}(x^\prime)(x - x^\prime ) q_i^{lk}(x - x^\prime)]\frac{1}{x^4}[{x^\prime}^2+(x - x^\prime)^2 ] \nonumber \\ & & - \frac{k}{6 } \int_0 ^ 1dx^\prime x g^{ape}(x ) \sum \limits_{i}[x^\prime q_i^{lk}(x^\prime ) + x^\prime { \bar{q}}_i^{lk}(x^\prime ) ] \frac{1}{x+x^\prime}[1 + [ \frac{x^\prime}{x+x^\prime}]^2 ] \nonumber \\ & & - \frac{k}{6 } \int_0 ^ 1dx^\prime x g^{lk}(x ) \sum \limits_{i}[x^\prime q_i^{ape}(x^\prime ) + x^\prime { \bar{q}}_i^{ape}(x^\prime ) ] \nonumber \\ & & \times
\frac{1}{x+x^\prime}[1 + { [ \frac{x^\prime}{x+x^\prime}]}^2]\end{aligned}\ ] ] the explicit @xmath1 dependence in the parton distributions is not shown in the above two equations in order to simplify the notation .
therefore , we write , for the nuclear gluon distribution , @xmath102 similarly , for the nuclear quark distribution , @xmath103 where @xmath104 is the modification due to the gluon shadowing(gs ) and is derived from the evolution equations . a crude estimate of @xmath105 is given by @xcite @xmath106}{\partial x } \label { sdq}\ ] ] using this relation
, we expect that the sea - quark distribution is also affected by the gluon modification as follows : @xmath107}{\partial x}\ ] ] where the step function @xmath108 is introduced because the relation is valid only at very small @xmath0 ( @xmath0 less than @xmath109 with @xmath109=0.1 ) .
we have considered the additional parton evolution and parton recombination in sections 2 and 4 , respectively , and gotten the nuclear parton distributions at @xmath110 . to compare the results with the experimental data or display the @xmath1-dependence of the model , these distributions
are evolved to those at lager @xmath1 by using the ordinary altarelli - parisi equations @xcite . for the valence quark(@xmath111 ) , sea quark(@xmath112 ) , and gluon(g ) , the evolution equations are @xmath113\ ] ] @xmath114+g^{a}(y , t)p_{qg}[\frac{x}{y}]]\ ] ] @xmath115+g^{a}(y , t)p_{gg}[\frac{x}{y}]]\ ] ] where @xmath116 is the splitting function , and @xmath117 is defined by @xmath118\ ] ] we used the solution of these integral equations in the leading order and the evolution subroutine of ref .
@xcite in our numerical calculation .
we now proceed to compute the modification in the structure function . in order to compare with the experimental data ,
let us define the ratio of the average nuclear structure function to the deuteron structure function as @xcite @xmath120 with the average nuclear structure function : @xmath121\ ] ] where @xmath122 and @xmath123 are the free neutron and proton structure functions , respectively .
@xmath124 is the nuclear structure function .
the second term compensates for the neutron excess . by considering the fermi motion of nucleons in the nucleus
, @xmath124 can be written as @xcite @xmath125 with @xmath126 , @xmath127 is the separation energy of a nucleon in the single - particle state @xmath128 , and @xmath129 is the single - particle wave function of the nucleon in the momentum space , which satisfies the light - cone normalization : @xmath130 in the following calculation , @xmath127 and @xmath129 are taken from ref .
@xcite . by using the nuclear parton distributions
after considering the additional parton evolution , recombination and the ordinary altarelli - parisi evolution to higher @xmath1 , the bound nucleon structure function @xmath131 can be expressed as @xmath132 @xmath133 + 2[s^a(x , q^2)+ \bar{s}^a(x , q^2)]\}\end{aligned}\ ] ] we now calculate @xmath134 in eq.([rad ] ) .
the deuteron structure function @xmath135 is taken from ref.@xcite .
similar to ref.@xcite , we take @xmath136}\ ] ] where the constant 0.48 is determined by @xmath137 @xcite . the theoretical result in the range
@xmath138 can not be well described since the gluonic modification due to eq .
( [ sdq ] ) is valid at very small @xmath0 .
however , using the kmrs - b0@xcite , our calculation results indicate that the modified parton distribution can be well parametrized by the analytical form of the corresponding input distribution .
the parameters in the analytical form of the modified parton distributions can be obtained by fitting the numerical results .
so we obtain a smooth curve of the ratio @xmath139 which is compared with the experimental data of the emc effect @xcite in fig.3 .
the result is in qualitative agreement with the experimental data . to show the distinct effect of the additional parton evolution , the following results are no longer presented by fitting numerical results and taking into account the fermi - motion correction . in order to investigate the @xmath1-dependence of the ratio @xmath139 ,
the results obtained by using the grv inputs@xcite and the kmrs - b0 inputs@xcite are plotted in fig.4(a ) and fig.4(b ) , respectively .
one can easily find that the result obtained by using the kmrs - b0 inputs is less dependent on @xmath1 . in fig.5 ,
the results with various input parton distributions of different parameterizations are presented , and they show that , in the small @xmath0 region , the recombination results after undergoing the additional parton evolution are very sensitive to the input sea - quark and gluon distributions . using the duke - owens(2 ) inputs of ref .
@xcite with hard gluon distributions , theoretical results underestimate the shadowing , which is not difficult to be understood since the parton parameterizations of ref .
@xcite did not fit to the experimental data in the small @xmath0 region .
this is the reason why we take kmrs - b0 @xcite or grv @xcite parameterizations as input parton distributions in most of the following calculations .
several years ago , the e772 collaboration @xcite at fermilab published the data of high - mass dilepton production measured in the nuclear drell - yan ( dy ) process .
these data aroused special attention in clarifying the different explanations for the nuclear effect on the parton distributions .
their results show that the ratio of dy dimuon yield per nucleon on a nuclear target to that on a free nucleon is slightly less than unity if the momentum fraction @xmath0 carried by a target quark is less than 0.1 .
the ratios over the range @xmath140 , however , do not reveal distinct nuclear dependence . in order to show the prediction of nuclear drell - yan ratio in our present model ,
let us begin with the definition of the nuclear drell - yan ratio : @xmath141 where the differential cross section of the nuclear drell - yan process is : @xmath142 with @xmath143\ ] ] in a more specified case for the proton - nucleus reaction : @xmath144xs^a(x , q^2)\ ] ] @xmath145 @xmath146\ ] ] where the total sea quark is @xmath147 .
by using the nuclear parton distributions in our model , we calculated the nuclear drell - yan ratios for @xmath148 . the integral range for @xmath149 in eq.([dyr ] ) is determined according to the kinematic region of the experiment in ref.@xcite , i.e. @xmath150 , and @xmath151 . to avoid the uncertainty derived by the crude estimate in eq .
( [ sdq ] ) for taking into account the gluonic modification , the results , except those in the range @xmath152 , are compared with the e772 @xcite data in fig.6 .
fig.6 indicates that the ape model as well as the @xmath1-rescaling model can explain the drell - yan ratio reasonably well .
however , in the small @xmath0 region(@xmath153 ) , the result obtained by the ape model seems to be better than that of the @xmath1-rescaling model . in 1992 , the nmc @xcite analyzed the inelastic @xmath7 production result with @xmath154 by the color singlet(cs ) model @xcite and obtained the gluon distribution ratios @xmath8 .
the result sheds light on gluon distributions in nuclei .
the results of @xmath8 calculated in our model together with the experimental data from nmc are presented in fig.7 , which are qualitatively similar to those given by the @xmath1-rescaling model in ref.@xcite . due to
the gluon shadowing , the ratios @xmath8 are less than 1 at @xmath155 .
it is noteworthy that experimental errors are very large in comparison with typical theoretical modifications . in order to test our model ,
we should wait for more accurate measurements of @xmath156 , for example , a proposed experiment at rhic @xcite in the small @xmath0 region for investigating details of the gluon shadowing .
it is worthwhile to compare the results of the present model with those of the @xmath1-rescaling model .
although these two models seem to have the same physics origin , they have different descriptions of the nuclear effect on sea quarks and gluons . in the ape model , the evolution of valence quarks , sea quarks and gluons
are determined by different interaction vertex factors , and their effects on their own momentum distributions are also different .
however , the @xmath1-rescaling model , ignoring the difference in additional parton evolution modes for different kinds of partons , describes the influence of the nuclear environment on the distribution functions of valence and sea quarks by using the same @xmath1-rescaling mechanism . in the medium @xmath0 region where the valence quarks
dominate , the @xmath1-rescaling model can explain the emc effect reasonably well , indicating that the nuclear effect on the valence quark momentum distribution can indeed be well described by the @xmath1-rescaling mechanism .
in the small @xmath0 region where sea quarks and gluons dominate , it is not clear whether the @xmath1-rescaling picture could be used also for sea quarks and gluons . to avoid this uncertainty
, we attempt to find a unified description of the nuclear effect on sea quarks and gluons as well as valence quarks .
this is the purpose of our investigation . to make a quantitative comparison between the ape model and the @xmath1-rescaling model ,
some results of these two models are shown in figs.8 - 9 in different @xmath0 scales .
firstly , figs.8 - 9 indicate that both models give reasonable explanations of the emc effect in the medium @xmath0 region .
however , the expectations of the ratio @xmath157 for two different input distributions ( grv and duke - owens ) are almost the same in the ape model but distinctly different in the @xmath1-rescaling model in which the ratio is slightly sensitive to the choice of free parton distributions . in the lager @xmath0 region , there are distinct modifications of the recombination effect in the ratio of @xmath157 which are similar to , but different from the contributions of the nuclear fermi motion ( cf .
fig.3 ) . for the small @xmath0 region
, we are always seeking for some mechanisms which account for the nuclear shadowing effect .
unfortunately , neither the ape model nor the @xmath1-rescaling model can give suitable explanation of the nuclear shadowing effect .
it has been emphasized by ref.@xcite that the shadowing due to the recombinations in the nuclear environment is mainly produced as a result of the modification in the gluon distribution . considering this recombination effect , both the @xmath1-rescaling model ( see ref .
@xcite ) and the ape model ( see fig.4 ) can fairly explain the emc and nmc shadowing data by using the kmrs - b0 input distributions .
but , by using harder input gluon distributions such as those of the duke - owens(2)@xcite , the theoretical results ( see figs.8 - 9 ) of these two models underestimate the shadowing . it seems , however , that by considering the recombination effect , the prediction of nuclear shadowing in the ape model is closer to the experimental data than that in the @xmath1-rescaling model with these harder input gluon distributions .
this is possibly attributed to the better descriptions of the nuclear effect on sea quarks and gluons in our model . as for the prediction of the nuclear drell - yan ratio
, the result in fig.6 indicates that , with the same input parton distributions of grv , the ape model has almost the same prediction of the nuclear drell - yan ratio for @xmath148 as the @xmath1-rescaling model in the range @xmath158 although there is a slight difference in the nuclear shadowing region ( @xmath159 ) .
the results of the nuclear drell - yan ratios for other nuclei such as c , fe , w are not presented since the results are qualitatively similar .
the results of these nuclear drell - yan ratios do not show a distinct nuclear dependence , which is consistent with the experimental fact .
in summary , we have put forward the ape model which attempts to explain the nuclear effect on the parton distributions by the change of the transverse momentum of partons in the nuclear environment instead of the @xmath1-rescaling model . by means of the ape model with the parton recombination effect ,
we investigated the nuclear effect including the nuclear structure function @xmath160 in the whole @xmath0 region , nuclear drell - yan process and nuclear gluon distribution .
the results of the nuclear structure function , in the small @xmath0 region ( @xmath161 ) , are of a little @xmath1-dependence but are very sensitive to the input sea - quark and gluon distributions . by using the proper input parton distributions
, we obtained reasonably good agreements with the experimental data of the emc effect , nuclear drell - yan ratio and nuclear gluon distributions .
in addition , the comparison between the ape model and the @xmath1-rescaling model shows , on one hand , that the results are qualitatively similar since they have the same physics origin ; and on the other hand , with better description of the nuclear effect on sea quarks and gluons , the value of @xmath162 in the ape model is less sensitive to the choice of the free parton distributions than that of the @xmath1-rescaling model .
our present investigation should be considered as the first step for seeking for more subtle mechanism with the same physics origin as the @xmath1-rescaling model but with less uncertainties .
further , we are looking forward to a unified description of the nuclear effect and the nuclear shadowing within the framework of the ape model .
this work was supported by national natural science foundation of china and natural science foundation of jiangsu province of china .
one of the authors ( j.j.y ) would like to thank dr .
s. kumano for his correspondence about the a - p equation subroutine , and dr .
vogt for sending their computer programs for calculating the parton distributions in ref .
he also expresses thanks to prof .
h. q. shen and dr .
s. a. aruna for helpful discussions .
150 emc , j. j. aubert et al .
, 123b(1983)275 .
nmc , p. amaudruz et al .
, z. phys .
c51(1991)387 .
emc , m. arneodo et al .
b333(1990)1 ; phys .
211b(1988)493 .
bickerstaff , m.c .
birse and g.a .
miller , phys .
d33(1986)3228 .
nm collab . , p. amaudruz et al . ,
b371(1992)553 .
s. kumano , phys .
c48(1993)2016 ; phys .
298b(1993)171 .
f. e. close , r. g. roberts , and g. g. ross , phys
129b(1983)346 ; nucl .
b296(1988)582 ; f. e. close , r. l. jaffe , r. g. roberts and g. g. ross , phys .
d31(1985)1004 .
m.glck and e. reya , nucl .
. h. georgi and h. d. politzer , phys .
. 36(1976)1281 .
( e:37(1976)68 ) ; phys .
d14(1976)1829 .
r. barbieri , j. ellis , m. k. gaillard and g. g. ross , phys .
64b(1976)171 ; nucl .
b117(1976)50 .
g. altarelli and g. parisi , nucl .
b126(1977)298 .
j. kwiecinski , a. d. martin , w. j. stirling , and r. g. roberts , phys .
d42(1990)3645 .
m. glck , e. reya and a. vogt , z.phys .
c67(1995)433 .
d. w. duke and j. f. owens , phys .
d30(1984)49 .
f. e. close , j. qiu , and r. g. roberts , phys .
d40(1989)2820 .
a. h. mueller and j. qiu , nucl .
b268(1986)427 ; j. qiu , ibid , b291(1987)746 ; e. l. berger and j. qiu , phys .
206b(1988)141 .
r. c. barrett and d. f. jackson , nuclear sizes and structure ( clarendon , oxford , 1977 ) . s. kumano and j. t. londergan , comput .
phys . commun .
69(1992)373 .
g. l. li , k. f. liu and g. e. brown , phys .
213b(1988)531 .
g. l. li , z. j. cao and c. s. zhong , nucl .
a509(1990)757 .
l. l. frankfurt and m. i. strikman , phys .
76b(1978)333 .
a.w hendry , d. b. lichtenberg and e. predazzi , nuovo cimento 92a(1986)427 .
g. e. brown and m. rho , phys .
237b(1990)3 .
d. m. alde et al . , phys .
64(1990)2479 .
e. l. berger and d. jones , phys .
d23(1981)1521 ; a. d. martin , c. k. ng and w. j. stirling , phys .
. 191b(1987)200 .
p. braun - munzinger and g. david , in : proc .
workshop on gross properties of nuclei and nuclear excitation ( hirschegg , austria , january 1992 ) ed .
fig.1 . deep inelastic scattering in the free parton model .
fig.4(a - b ) .
comparisons of the ratio @xmath157 with emc-90 @xcite and nmc @xcite data by using ( a ) the grv @xcite input and ( b ) the kmrs - b0 @xcite input . @xmath164 and
@xmath165 for the solid and dashed curves , respectively .
comparisons of the ratio @xmath157 with emc-90 @xcite and nmc @xcite data by using different input distributions .
the solid , dashed and dotted curves are for the kmrs - b0 @xcite , the grv @xcite and duke - owens(2 ) @xcite , respectively .
@xmath164 .
the nuclear drell - yan ratios @xmath166 predicted by the present model(solid curve ) and the @xmath1-rescaling model(dashed curve ) are plotted versus x by using the grv input distributions at @xmath1=20 @xmath167 .
the experimental data are taken from the e772 collaboration @xcite .
fig.8(a - b ) . the ratio @xmath168 given by the ape model
is plotted .
( a ) in the linear x scale ; ( b ) in the logarithmic x scale .
results by using the grv inputs(solid curve ) and the duke - owens(2 ) inputs(dashed curve ) at @xmath1=4 @xmath167 .
fig.9(a - b ) . the ratio @xmath168 given by the @xmath1-rescaling model
is plotted .
( a ) in the linear x scale ; ( b ) in the logarithmic x scale .
results by using the grv inputs(solid curve ) and the duke - owens(2 ) inputs(dashed curve ) at @xmath1=4 @xmath167 . | the relationship between the parton transverse momentum and longitudinal momentum , which is obtained by analysing the hadron tensor in the deep inelastic scattering ( dis ) process , indicates that the transverse size of partons in a nucleon is sensitively affected by the effective nucleon mass .
the change in this transverse size gives rise to an additional parton evolution ( ape ) in the nuclear environment refer to the normal qcd parton evolution thus resulting in new parton distributions .
the self - consistent integral equations for these new parton distributions are given .
having undergone the normal qcd evolution and the additional evolution in the nuclear environment , the partons with the small @xmath0 participate in recombination and contribute to the nuclear shadowing effect .
the reasonably good descriptions of the nuclear effect of structure functions , nuclear drell - yan ratios and nuclear gluon distributions are presented .
a quantitative comparison between the ape model and the @xmath1-rescaling model is made . the nuclear effect induced by additional parton evolution and recombination in nuclear environment jian - jun yang@xmath2 , guang - lie li@xmath3 |
this research was supported at nyu by nsf grant dmr-1006575 , usaro grant no .
w911nf0710643 as well as the partner university fund ( puf ) of the embassy of france . | we present a two - current - pulse temporal correlation experiment to study the intrinsic subnanosecond nonequilibrium magnetic dynamics of a nanomagnet during and following a pulse excitation .
this method is applied to a model spin - transfer system , a spin valve nanopillar with perpendicular magnetic anisotropy .
two - pulses separated by a short delay ( @xmath0 ps ) are shown to lead to the same switching probability as a single pulse with a duration that depends on the delay .
this demonstrates a remarkable symmetry between magnetic excitation and relaxation and provides a direct measurement of the magnetic relaxation time .
the results are consistent with a simple finite temperature fokker - planck macrospin model of the dynamics , suggesting more coherent magnetization dynamics in this short time nonequilibrium limit than near equilibrium . the control of magnetization on short timescales has become an area of intense research and many methods are used to excite a magnetic system , including spin - currents @xcite , magnetic fields @xcite and optical pulses @xcite .
it also has recently become possible to drive individual nanometer scale magnetic elements far from equilibrium using spin - currents and probe their dynamical response . however , most electronic transport studies focus on the excitation during the drive pulses @xcite .
less is known about the relaxation processes following a pulse excitation .
yet , this is very important for fundamental and practical reasons .
first , the time scales of the relaxation far from equilibrium are not known and may differ from those determined in experiments that probe low amplitude magnetic excitations , such as ferromagnetic resonance .
second , these time scales determine the speed at which nanomagnets can be written and read in magnetic memories , including in spin - transfer torque magnetic random access memories ( stt - mram ) .
it is now well known that spin - polarized currents and spin - transfer torques ( stts ) can reverse the direction of the magnetization on subnanosecond timescales @xcite .
in fact , many experimental methods have been developed to study spin - transfer switching in both spin valves and magnetic tunnel junctions ( mtjs ) .
while mtjs are promising for applications because of their large magnetoresistive ( mr ) readout signals ( @xmath1% ) , all - metallic spin valves permit one to apply much larger currents and thus can be driven farther away from equilibrium .
direct single - shot time - resolved electrical measurements have been carried out in mtjs @xcite , but have not been reported in spin valve nanopillars , because of their low impedance and small magnetoresistance ( mr ) ( @xmath2% ) , which results in their switching signals being to small for nanosecond measurements .
further , time - resolved stt measurements are insensitive to the magnetization dynamics after the excitation , as the excitation current is the source of the electrical readout signal . therefore ,
to time - resolve the relaxation dynamics , a different experimental method is needed .
in this paper we present an all - electrical two - pulse correlation method that yields quantitative information on the form and timescales of magnetic excitation and relaxation , with time resolution .
this method is applied to a model spin - transfer system , a spin valve nanopillar with perpendicular magnetic anisotropy .
two - pulses separated by a short delay ( @xmath0 ps ) are shown to lead to the same switching probability as a single pulse with a duration that depends on the delay demonstrating a remarkable symmetry between magnetic excitation and relaxation and providing a direct measurement of the magnetic relaxation time .
the observed symmetry is shown to be consistent with a finite temperature macrospin model .
the system we studied has two stable magnetic states @xmath3 and @xmath4 , which are nearly degenerate in equilibrium .
we are interested in how the system relaxes to these states after being excited by a spin - polarized current .
the basic idea of our method is to compare the switching probability of two pulses separated by a delay much longer than the magnetic relaxation time to that of a composite pulse , i.e. two pulses separated by a short delay . with a composite pulse ,
the first pulse alters the magnetic state on a time scale that changes the dynamics excited by the second pulse . by choosing the amplitudes and polarities of the first and second pulse we can study different aspects of the relaxation processes . for example , to study the relaxation back to the initial state ( i.e. state @xmath3 ) after a pulse , we can apply two pulses with identical polarities that both would be of the appropriate polarity to switch the sample from state @xmath3 to @xmath4 .
if the delay is much longer than the relaxation time @xmath5 , the two pulses can be considered to be independent events and the total probability of switching from state @xmath3 to state @xmath4 is : @xmath6 here @xmath7 , @xmath8 and @xmath9 are the switching probabilities from state @xmath3 to @xmath4 for the first , the second and for both pulses combined and we have assumed no reverse switching during the measurement , i.e. @xmath10 . if @xmath11 is within the relaxation time of the system , the total switching probability will increase compared to eq . : @xmath12 using these equations we can resolve the relaxation dynamics of the sample by measuring and comparing the switching probabilities as a function of @xmath11 @xcite . our sample is a spin - valve nanopillar consisting of a nico free layer and exchange coupled nico and copt multilayers as the reference layer , both of which have perpendicular magnetization anisotropy ( pma ) .
( for the full layer stack see @xcite . )
such pma samples have a simple uniaxial anisotropy energy landscape that permits a direct comparison to theoretical models . the sample presented in this work is @xmath13{nm}\times\unit[100]{nm}$ ] in size , fabricated by e - beam and optical lithography @xcite .
we have studied more that @xmath14 samples of different shapes and sizes and found similar results .
the equilibrium states are antiparallel ( @xmath15 ap ) and parallel ( @xmath16 p ) magnetization alignment of the free and reference layers and there is @xmath17 mr , well within the accuracy of our method .
the room temperature coercive field is @xmath13{mt}$ ] , large enough so that no reverse switching occurs @xcite .
the zero temperature switching current for antiparallel ( ap ) to parallel ( p ) switching is @xmath18{ma}$ ] .
we use an arbitrary waveform generator to apply two pulses separated by a time delay @xmath11 , as illustrated by the inset of fig .
[ fig:1](a ) .
the state of the spin valve is determined by a resistance measurement using a lock - in amplifier both before and after the injection of the double - current pulse .
we note that the lock - in current is small enough ( ) not to affect the spin valves and not to induce reverse switching @xcite . all experiments described here were conducted at room temperature .
initially the sample was brought into the ap state by the application of an easy - axis magnetic field of appropriate direction and magnitude . a short current pulse of positive polarity @xmath19{ma}$ ] and @xmath20{ns}$ ]
was then applied .
a longer pulse of positive polarity would eventually switch the sample into the p state ( as @xmath21 ) .
however , the duration has been chosen such that the switching probability is nearly zero ( @xmath22 ) , as shown by the dashed curve ( i.e. the single pulse curve ) in fig .
[ fig:1](a ) at . after a variable delay from to
, a second pulse with the same amplitude ( @xmath23 ) and a variable duration @xmath24 , from to , was applied .
the total switching probability distribution @xmath25 was then measured by repeating the process @xmath26 to @xmath27 times for each pulse set . as seen in fig .
[ fig:1](a ) , for delays longer than the switching probability distribution is the same as that of the second pulse alone @xmath28 , which is consistent with eq . for @xmath29 .
this indicates that these delays are longer than the magnetization relaxation time .
the magnetization state excited by a current pulse decays to equilibrium in under , within our experimental accuracy .
however , for delays less than , the switching probability @xmath25 is larger than that of just a single pulse . remarkably , even with a composite pulse , the switching probability versus the second pulse duration is seen to follow the same functional form as that for a single pulse
this is shown in fig .
[ fig:1](b ) , where the probability data , shifted by an amount @xmath30 a quantity that depends on the delay between the pulses is seen to closely overlap .
this clearly demonstrates that the switching probability of a composite pulse ( @xmath31 ) is equal to that of a single pulse with an increased duration ( @xmath32 ) .
since this is satisfied for all @xmath24 , a natural conclusion is that the ensemble averaged state of the sample after the first pulse and the delay ( @xmath33 ) is the same as that after a single pulse with a duration of @xmath34 , as illustrated schematically by the pulse shapes in the inset of fig .
[ fig:1](b ) . from the data shown in fig .
[ fig:1 ] we determine @xmath34 as a function of the delay @xmath11 .
this is shown as crosses in fig .
[ fig:2 ] .
the black solid curve is an exponential fit to the data where @xmath35 is the lifetime of the excitation : @xmath36 with the parameters @xmath37{ns}$ ] and @xmath38{ns}$ ] . for no delay , @xmath39 , clearly @xmath40 and thus @xmath41 , which is indeed fulfilled within the time resolution of the experiment ( @xmath42 ps ) .
eq .
links the duration of the excitation , the delay and the corresponding equivalent pulse duration , which is related to the magnetization relaxation rate . versus delay @xmath11 .
the blue crosses are determined from the experimental data in fig .
[ fig:1 ] .
the black solid curve is an exponential fit to eq . .
the green dash - dotted line is a calculation from the llg equation for @xmath43 .
the red dashed curve is a calculation based on a fokker - planck analysis ( eq . ) . ]
the data in fig .
[ fig:2 ] is unexpected at first , as a zero temperature macrospin landau - lifshitz - gilbert ( llg ) model predicts a linear relationship between @xmath34 and @xmath44 , as shown by the green dash - dotted line in fig .
[ fig:2 ] .
this linear relationship follows from the llg equation .
for a uniaxial nanomagnet in zero applied magnetic field , the polar angle of the magnetization @xmath45 , i.e. the angle between the magnetization and the easy axis , satisfies @xcite : @xmath46 with : @xmath47 @xmath48 @xmath49 is the relaxation time due to damping and depends on the material parameters , the anisotropy field @xmath50 , the gilbert damping parameter @xmath51 and the gyromagnetic ratio @xmath52 .
@xmath49 is typically hundreds of picoseconds .
@xmath53 is the zero temperature critical current as we mentioned earlier , representing the threshold current for switching .
it depends on the magnetization density @xmath54 , the nanomagnet s volume @xmath55 and the spin polarization of the current @xmath56 . from eq .
both excitation ( @xmath57 ) and relaxation ( @xmath58 ) follow the same functional form for small angles @xmath45 and are only different in their respective time constants .
it is straightforward to show that the duration of a single pulse ( @xmath34 ) , which brings the sample to the same state as the first pulse ( @xmath59 ) followed by the delay ( @xmath11 ) satisfies : @xmath60 hence the initial slope of @xmath34 versus @xmath11 , shown as the green dashed dotted curve in fig .
[ fig:2 ] , depends on the current overdrive , @xmath61 , and provides an independent method to determine the zero temperature critical current .
however , for long delays this equation is unphysical , as it predicts a negative @xmath34 .
a physical picture is that during the first pulse the magnetization evolves from a thermal distribution of initial angles near the north pole ( @xmath62 , where @xmath63 is the energy barrier and @xmath64 is the thermal energy ) to a larger angle @xmath65 . during the delay , and
provided @xmath66 , the polar angle decays back toward the north pole .
however , in this zero temperature model the magnetization decays back to @xmath67 , i.e. to an angle that can be less than the initial angle , @xmath68 .
the difference between the llg model prediction and the experimental results can be explained by thermal fluctuations both during the excitation and relaxation processes . as is well known , thermal fluctuations lead to a spread of the initial angles of magnetization about the north pole .
a spin - current pulse works to rotate the magnetization away from the easy axis direction and , in essence , amplifies these magnetization fluctuations .
once the magnetization deviates from the easy axis direction , the magnetization polar angle grows exponentially .
however , thermal noise also slows down the relaxation from a simple exponential decay process predicted by the llg equations and leads to relaxation to a thermal distribution .
therefore , thermal fluctuations not only change the timescales of the excitation and relaxation processes , they also alter their functional forms . to model the influence of thermal fluctuations
, we simulated the double pulse measurement using the fokker - planck equation @xcite , which has the following form for a uniaxial macrospin system : @xmath69 \nonumber \\ & + & \frac{1}{2\xi\tau_d}\frac{\partial ^ 2 \mathcal{p}}{\partial \theta ^ 2 } \label{eq : fp}\end{aligned}\ ] ] where @xmath70 is the probability density of the magnetization at the angle @xmath45 and the time @xmath71 .
@xmath72 is the dimensionless energy barrier , @xmath73 . for a uniaxial nanomagnet @xmath74 .
we start with a boltzmann distribution centered around the easy axis in the ap state : @xmath75 where @xmath76 is a normalization constant so that @xmath77 .
for the simulation we use a fixed duration for the first pulse of , the same as in our experiments and vary the duration of the second pulse for different delays . by fitting the measured single pulse switching probability distribution ( i.e. the red dashed curve of fig . [ fig:1](a ) ) to eq . .
we obtain @xmath78 and @xmath79{ps}$ ] .
we then can track the ensemble averaged magnetization projection , @xmath80 as a function of time @xmath71 , as shown in fig .
[ fig : fp](a ) .
the shaded region depicts the angular range containing @xmath81 to @xmath82 of the states . during the first current pulse ( @xmath83{ns } \leq t \leq \unit[0.9]{ns}$ ] )
, the switching process begins .
the averaged magnetization projection decreases while the occupation probability distribution spreads out , where the majority of the distribution is still centered around @xmath84 . during the delay ( @xmath85{ns } \leq t \leq \unit[1.9]{ns}$ ] ) , the averaged magnetization projection increases while the occupation probability distribution narrows toward the initial state .
then during the second pulse ( @xmath86{ns}$ ] ) , the averaged magnetization projection decreases again as the total distribution expands during the beginning part of the switching process .
we can further check whether a pulse and a delay ( @xmath33 ) brings the system into the same state as a single @xmath34 duration pulse .
fig .
[ fig : fp](b ) shows the occupation probability as a function of the polar angle for different delay times .
the distributions for a pulse and delay @xmath87 ( the solid light blue curves in fig .
[ fig : fp](b ) ) overlap with those of a single @xmath34 pulse ( dashed curves in fig .
[ fig : fp](b))which is consistent with scaling of the experimental data shown in fig .
[ fig:1](b )
. the fokker - planck simulation results also agree quantitatively with the experimental data .
as shown in fig .
[ fig : fp](c ) the calculated switching probability distributions have the same form independent of @xmath11 and can be shifted in time to coincide .
the results shifts , @xmath34s , are in good agreement with the experimental results , as indicated by the red dashed curve in fig .
[ fig:2 ] .
the difference between the llg and the fokker - planck analysis clearly shows the influence of the thermal fluctuations , even though the total process ( excitation plus relaxation ) lasts less than . in summary , we have presented a method ideally suited to study excitation and relaxation in a magnetic nanostructure .
this method only requires the generation of pulses with short durations and delays and measurement of the switching probability . it does not require resolving small signal changes at high bandwidth and it can also be used to study dynamics after a pulse , with no current applied .
our method provides a direct ( model independent ) measurement of the relaxation timescale by considering when two events are correlated .
we have used this method to study excitation and relaxation in a model system a thin film nanomagnet with uniaxial anisotropy excited by spin - current pulses .
we have shown that a single pulse followed by a delay leaves a nanomagnet in a state with the same ensemble average as that of a single pulse with a shorter duration @xmath34 .
this finding is also model - independent .
the @xmath34s are seen to decrease exponentially , following the form of eq .
[ eq : fitting ] , and giving a relaxation time of @xmath88{ps}$ ] .
we note that the relaxation time is more typically obtained experimentally by fitting to an arrhenius - nel s model with a variable prefactor ( see , for example , @xcite ) . in general
, the decay of the magnetization might be expected to be more complex than a simple exponential ( eq . [ eq : fitting ] ) , including multiple relaxation pathways with different timescales . however , even in such a case the directly measured relationship between @xmath34 and @xmath11 would provide information that is important for understanding such relaxation processes .
using eq . with an anisotropy field of @xmath89 t
( see ref .
@xcite ) and a gyromagnetic ratio of @xmath90(ts ) ( i.e. a land g - facto of @xmath91 ) , we find a gilbert damping of @xmath92 , a value that is larger than that determined from thin film ferromagnetic resonance measurements of the damping ( @xmath93 @xcite ) .
the larger damping inferred from the data may be a consequence of the large - angle excitation of the magnetization , including non - linear effects @xcite , or to changes of the magnet s material properties associated with device nanofabrication @xcite .
interestingly , our results can be understood within a simple macrospin model that incorporates thermal fluctuations . in fitting our data to this model
we obtain a dimensionless energy barrier of @xmath78 , which is within 25% of that expected based on the volume , magnetization and anisotropy of the nanomagnet ( @xmath94 ) . our previous results , as well as
those of many other groups , have found energy barriers to reversal for long field @xcite and long current pulses ( @xmath95 ns ) @xcite to be only a small fraction of that expected for a macrospin , suggesting the reversal proceeds by subvolume nucleation @xcite . for similar samples , we found that the energy barrier to reversal under long current pulses to be @xmath96{k_bt}$ ] @xcite , @xmath97 the macrospin value .
this suggests that the reversal modes in the short - time nonequilibrium limit are distinct from those in the long - time limit , as the system may not have time to follow minimum energy paths or be driven from these paths by the spin transfer torque .
the fundamental explanation is an open question .
we expect that the dependency of the effective energy barrier on pulse amplitude can be further studied by our method with different pulse conditions .
our double pulse method , which provides access to the intrinsic time scales of the excitation and relaxation dynamics , can be used in other cases where the readout signal is also small , such as current induced domain wall motion . |
during past decade , fluorescence of molecules placed near a metal nanostructure supporting surface plasmon ( sp ) has attracted renewed interest due to biosensing applications.@xcite while early studies mainly focused on fluorescence near rough metal films , @xcite more recent advances in near - field optics and in chemical control of molecule - nanostructure complexes spurred a number of fluorescence measurements from dye molecules and semiconductor quantum dots ( qds ) linked to metal nanoparticles ( np).@xcite fluorescence quantum efficiency of such conjugate systems is determined by competition between strong local field enhancement of molecule s dipole moment near sp resonance@xcite and , hence , of radiative decay rate @xmath2 , and nonradiative energy transfer from excited dipole to optically - inactive electronic excitations in the metal characterized by nonradiative decay rate @xmath3.@xcite the interplay between these two mechanisms depends on separation , @xmath0 , of the emitter from the metal surface.@xcite as a result , fluorescence is enhanced at some optimal @xmath0 and is quenched close to np due to the suppression of quantum efficiency , @xmath4 , by non - radiative processes . both enhancement and quenching
were widely observed in fluorescence experiments on au and ag nanoparticles.@xcite when a dipole with excitation energy @xmath5 is situated near a spherical np of radius @xmath6 its energy is transferred to via coulomb interactions to electronic excitations in np ( single - particle or collective ) with angular momentum @xmath7 . within classical approach,@xcite
these are collective modes described by resonances of np @xmath7-pole dynamic polarizabilities @xmath8 expressed in terms of _ local _ dielectric function @xmath9 , i.e. , bulk metal dielectric function @xmath10 for @xmath11 and dielectric constant of outside medium @xmath12 for @xmath13 .
the energy dissipation in np is then determined by @xmath14 which comes mainly from electron - phonon interactions and , for high @xmath5 , from optical transitions between d - band and sp - band .
the distance dependence of energy transfer rate @xmath3 stems from electromagnetic interaction of molecular dipole with its image that , at small @xmath0 ( i.e. , @xmath15 ) , yields@xcite @xmath16 , similar to flat metal surface.@xcite recent fluorescence measurements from molecules near ( but not very close to ) relatively large ( tens of nm ) nps @xcite indicated a good agreement with classical electromagnetic models.@xcite with decreasing distance between dipole and metal surface , the energy transfer rate gets significantly modified due to spatial dispersion of @xmath17 . for @xmath18 , where @xmath19 is fermi velocity , the high fourier harmonics of dipole s coulomb potential are no longer screened by conduction electrons in the metal so that the phase space restriction on excitation of electron - hole pairs over fermi level is lifted.@xcite for small @xmath0 , electron - hole generation becomes the dominant energy transfer mechanism that leads to a sharper distance dependence of energy transfer rate , @xmath20 ; the precise coefficient calculated within random phase approximation ( rpa ) differs in the literature.@xcite while fluorescence quenching from a molecule near a relatively large np is essentially similar to that near flat metal surface,@xcite the energy transfer to _ small _ nps is significantly modified due to quantum - size effects .
fluorescence near nanometer - sized au nps with diameter as small as 1.4 nm has attracted significant interest due to their large surface to volume ratio and , hence , lower ohmic losses , which is important for biosensing applications.@xcite for such nps , a spillover of a substantial fraction of electron density beyond np classical radius can significantly affect spatial dispersion of metal dielectric function @xcite and hence modify the dipole coulomb potential near np surface .
the presence of strongly localized d - band with abrupt nearly classical density profile gives rise to a surface layer where coulomb potential is not screened by interband transitions.@xcite in small nps , the size - dependent landau damping of plasmons excited by the molecular dipole can increase the rate of energy dissipation.@xcite incorporation of all the above effects in a single model requires a microscopic approach of energy transfer rates near small nps . in this paper , we present such a microscopic model , based on time - dependent local density approximation ( tdlda ) , for calculation of radiative and nonradiative decay rates of a molecular dipole near a small noble metal np . here
, the molecule is represented by a point - like dipole while we focus on non - local and quantum - size effects in dielectric function of nanometer - sized metal np . to this end , we adopt quantum three - region model we developed recently@xcite that accounts for different density profiles of sp - band and d - band electrons @xcite in small noble metal np embedded in a dielectric medium .
we find that calculated radiative decay rates , @xmath2 , are not significantly affected by non - local effects and are comparable to those calculated using semiclassical approaches@xcite with quantum corrections . at the same time , the energy transfer rate @xmath3 is significantly enhanced due to the generation of electron - hole pairs in metal at small molecule - np separations .
we find , however , that quantum - size and non - local effects strongly affect the distance dependence of @xmath3 close to np surface and that @xmath1 behavior no longer holds ; namely , the energy transfer rate is significantly slower than that predicted by rpa calculations.@xcite the paper is organized as follows . in section [ sec : dipole ] we set up the expressions for decay rates to be evaluated and derive the classical limit . in section [ sec : tdlda ] we outline the quantum three - region model used for evaluation of np polarizabilities . in section [ sec : numeric ] our numerical results are presented and discussed , and section [ sec : conc ] concludes the paper .
consider a molecule with a dipole moment @xmath21 located at position @xmath22 measured from the center of a metal np of radius @xmath6 placed in a medium with dielectric constant @xmath12 .
its decay rate is given by a general expression @xcite @xmath23 where @xmath24 is the electric field green dyadic in the presence of np and @xmath25 is the wave vector of light . the full decay rate eq .
( [ green ] includes contributions coming from two distinct processes , @xmath26 : emission of a photon by np - dipole system with the rate @xmath2 , and energy transfer to electronic excitations in the np with the rate @xmath3 . if the system size is sufficiently small , @xmath27 , then @xmath3 can be found in a standard way from the longwave limit of the green dyadic,@xcite @xmath28 where @xmath29 % \nonumber\\ = \sum_{lm } \frac{4\pi}{2l+1 } \left[\frac{r_{<}^{l}}{r_{>}^{l+1 } } -\frac{\alpha_l}{\left(rr'\right)^{l+1 } } \right ] y_{lm}(\hat{\bf r})y_{lm}^{\ast}(\hat{\bf r}'),\ ] ] is expansion of coulomb potential in the presence of np over spherical harmonics @xmath30 , and @xmath31 and @xmath32 are the larger and smaller values of @xmath33 and @xmath34 , respectively . the second term in eq .
( [ coulomb ] ) contains np @xmath7-pole dynamic polarizability , @xmath8 , given by @xmath35 where @xmath36 is the induced density due to potential @xmath37 ( hereafter we suppress @xmath5-dependence to alleviate expressions ) . from eqs .
( [ green ] ) and ( [ coulomb ] ) , @xmath3 can be calculated for arbitrary dipole orientation with respect to the np surface ; for perpendicular and parallel orientations , it takes the form @xcite @xmath38 where @xmath39 is radiative decay rate of an isolated dipole .
the radiative decay rates in the presence of np are similarly obtained using green dyadic s far field asymptotics as @xcite @xmath40 and they depend on np dipole polarizability , @xmath41 , that peaks at surface plasmon frequency , @xmath42 .
note that the expression ( [ polar ] ) for np polarizability that determines decay rates ( [ nonrad ] ) and ( [ rad ] ) is valid as long as there is no direct overlap between molecule orbitals and electron states in the metal.@xcite in classical approximation of a sharp np boundary at @xmath6 , the induced density @xmath43 peaks at @xmath44 , yielding standard expression for np polarizability , @xmath45 where @xmath46 is the metal dielectric function . as the dipole distance to np surface , @xmath47 , decreases , @xmath2 and @xmath3 exhibit significantly different behavior .
the distance dependence of @xmath2 is determined largely by the interference between direct and np - induced terms , but it remains finite for @xmath15 .
in contrast , @xmath3 sharply increases for small @xmath0 due to dominant contribution of high-@xmath7 terms in eq .
( [ nonrad ] ) . indeed , using drude form for bulk metal dielectric function ,
the large @xmath7 approximation for np polarizability ( [ polar - classical ] ) is @xmath48 , where @xmath49 and @xmath50 are , respectively , bulk plasmon energy and damping rate . for @xmath15 ,
the estimate of @xmath3 then takes the form @xmath51 where @xmath7 is restricted by @xmath52 , @xmath53 being the electron fermi momentum . in the local approximation when the spatial dispersion of metal dielectric function @xmath46 is neglected ,
@xmath50 is determined by electron bulk scattering length , @xmath54 due to mainly electron - phonon and impurity scattering , @xmath55 , resulting in @xmath16 distance dependence .
the rpa calculations non - local corrections at small @xmath0 due to electron - hole pairs generation@xcite change this behavior to @xmath20 . in the rest of the paper , we present microscopic calculation of the decay rates and fluorescence quantum efficiency .
for calculation of np polarizabilities eq .
( [ polar ] ) , we adopt quantum three - region model developed by us previously @xcite that combines a quantum - mechanical description for _ sp_-band electrons within tdlda and phenomenological treatment of _ d_-electrons with dielectric function @xmath56 in the region confined by @xmath57 and of outside medium with dielectric constant @xmath12 .
this model represents an extension of two - region model ( i.e. , without outside dielectric ) that is known to describe reasonably well screening effects for symmetric ( closed shell ) spherical clusters with electron numbers ranging from several dozens to several thousands.@xcite in the longwave approximation , local fields at point @xmath58 outside np are determined by quasistatic potential @xmath59 , where @xmath60 is induced potential .
the induced density is decomposed as @xmath61 , where @xmath62 , @xmath63 , and @xmath64 are , respectively , contributions from _ sp_-electrons , _ d_-electrons , and surrounding medium .
the _ sp_-electrons contribution is determined from tdlda equation @xcite @xmath65\delta n_s ( { \bf r}')\bigr],\end{aligned}\ ] ] where @xmath66 is polarization operator for noninteracting _ sp_-electrons ,
@xmath67 $ ] is the ( functional ) derivative of exchange - correlation potential,@xcite @xmath68 being the ground - state electron density ; @xmath66 and @xmath68 are calculated in a standard way using kohn - sham equations.@xcite the system is closed by expressing @xmath69 in eq .
( [ tdlda ] ) via @xmath62 .
this is accomplished by relating @xmath63 and @xmath64 in eq .
( [ poisson ] ) back to @xmath69 as @xcite @xmath70 , \nonumber\\ e^2 \delta n_m({\bf r})= \nabla \bigl[\chi_m(r)\nabla \phi({\bf r})\bigr ] , \end{aligned}\ ] ] where @xmath71 \,\theta(r_d - r)$ ] is interband susceptibility with step function @xmath72 enforcing the boundary conditions and , correspondingly , @xmath73\,\theta(r - r)$ ] is the susceptibility of surrounding medium . using spherical harmonics expansions @xmath74 and @xmath75 as well as the continuity of @xmath69 at @xmath44 and @xmath76
, we obtain decomposition @xmath77 , where @xmath78 is potential of external field , @xmath79 is contribution from d - band and outside medium , and @xmath80 is sp - band contribution .
the latter is related to the sp - band induced density as@xcite @xmath81 where the kernel @xmath82 is given by @xmath83 \end{aligned}\ ] ] with @xmath84 , 1 , and @xmath85 for @xmath33 in regions @xmath86 , @xmath87 , and @xmath88 .
the functions @xmath89 and @xmath90 are , respectively , the @xmath7-th harmonics of direct and image coulomb potential : @xmath91 and @xmath92 -\beta_l(r / r_d)\frac{\lambda_d}{\eta}\bigl[b_{0}^{(l)}(r_d , r ' ) \nonumber\\ & \qquad \qquad \qquad \qquad \qquad -a^{l}(l+1 ) \lambda_m b_{0}^{(l)}(r , r')\bigr],\end{aligned}\ ] ] where @xmath93 is the derivative of coulomb potential at the boundary , @xmath94 and coefficients @xmath95 , @xmath96 , @xmath97 , and @xmath98 are given by @xmath99 the combined d - band and outside medium contribution has the form @xmath100 \bigr].\end{aligned}\ ] ] we then obtain a closed equation for @xmath101 as , @xmath102 , \nonumber\end{aligned}\ ] ] where the polarization operator @xmath103 and the exchange potential@xcite @xmath104 are calculated from kohn - sham eigenfunctions.@xcite finally , np polarizability ( [ polar ] ) can be derived from the induced potential ( [ poisson ] ) outside np using the relation @xmath105 . using eqs .
( [ dws ] ) and ( [ dw0 ] ) , we obtain @xmath106 , where @xmath107/\eta,\ ] ] and @xmath108 with @xmath109 b_{0}^{(l)}(r , r).\end{aligned}\ ] ] equations ( [ a])([varphi ] ) determine self - consistently quantum polarizabilities @xmath8 of noble metal np in dielectric medium which , in turn , determine the decay rates ( [ nonrad ] ) and ( [ rad ] ) . in the next section
, we present the results of our numerical calculations of @xmath2 and @xmath3 and of fluorescence quantum efficiency for a molecule near a small np .
( color online ) calculated radiative ( a ) and nonradiative ( b ) decay rates of a molecule near @xmath110 nm diameter au np for normal and parallel dipole orientation are compared to semiclassical ( gersten - nitzan ) decay rates . ]
( color online ) same as in fig .
[ fig:1100rate ] , but for @xmath111 nm diameter au np . ]
calculations of decay rates ( [ nonrad ] ) and ( [ rad ] ) were carried out for a point dipole situated near au np in a medium with dielectric constant @xmath112 .
we chose dipole s emission frequency to coincide with that of cy5 fluorophore , 1.83 ev , which lies at low energy side of au np sp resonance ( 1.98 ev ) , so that quantum size and non - local effects we are interested in are not being obscured by interband absorption . to elucidate the role of quantum - size effects , the decay rates @xmath2 and @xmath3
were calculated for _ two _ spherical au clusters with electron numbers @xmath113 and @xmath114 corresponding to np diameters @xmath115 nm and @xmath116 nm , respectively , and to highlight the role of non - local and screening effects the results of microscopic calculations are compared to those obtained using semiclassical polarizabilities , eq . ( [ polar - classical ] ) with the landau damping of plasmons @xmath117 included in _ both _ quantum and classical np polarizabilities .
both normal and parallel dipole orientations were considered and multipolar polarizabilities @xmath8 with angular momenta up to @xmath118 were included .
the intrinsic decay rates , @xmath119 s@xmath120 and @xmath121 s@xmath120 of cy5 dye were included in the full decay rates , and the bulk interband dielectric function@xcite @xmath56 was used for determination of three - region model parameters.@xcite in figs .
[ fig:1100rate ] and [ fig:92rate ] , we show calculated radiative decay rate @xmath2 and energy transfer rate @xmath3 for @xmath110 nm and @xmath111 nm au nps , respectively , together with the corresponding semiclassical rates based on gersten - nitzan ( gn ) model@xcite defined by np polarizabilities eq .
( [ polar - classical ] ) with size - dependent landau damping correction included .
for radiative decay rate @xmath2 , the non - local effects are not significant except in narrow surface region ( denoted by dashed line ) were np radius is not well defined due to spillover effects .
this result is expected because only @xmath122 excitations contribute to @xmath2 through np dipole polarizability @xmath41 which does not account , due to the phase space restrictions , for direct excitation of electron - hole pairs .
a slightly larger quantum @xmath2 for normal dipole orientation is due to the difference in d - band and sp - band electron density profiles in the np surface layer ; the spillover of conduction electrons ( as opposed to strongly localized d - band electrons ) leads to reduction interband screening and hence larger local fields in the vicinity of np boundary.@xcite this effect is stronger for small np due to larger volume fraction of underscreened region.@xcite for parallel dipole orientation , the overall @xmath0-dependence of @xmath2 is dominated by destructive interference between direct and scattered waves.@xcite note that for small nps , the minimum is reached in the spillover region , but moves towards outer region with increasing np size .
( color online ) calculated fluorescence quantum efficiencies near @xmath110 nm diameter
au np for normal and parallel dipole orientation are compared to semiclassical ( gersten - nitzan ) quantum efficiencies .
inset : blow - up for small molecule - np distances . ]
( color online ) same as in fig .
[ fig:1100q ] , but for @xmath111 nm diameter au np . ]
( color online ) normalized nonradiative decay rates for @xmath111 nm ( a ) and @xmath110 nm diameter au np are shown together with @xmath123 fit . ] in contrast , the calculated quantum energy transfer rate @xmath3 is significantly larger than its semiclassical counterpart for either dipole orientation .
close to np surface , @xmath3 for the larger @xmath110 nm np is enhanced by factor @xmath124 due to direct excitation of fermi see electron - hole pairs by dipole s coulomb potential@xcite ( see fig .
[ fig:1100rate ] ) .
note that the role of non - local effects in @xmath3 considerably exceeds that of sp landau damping which is included in both calculations .
however , landau damping is more important in small @xmath111 nm np , as it leads to smaller difference between quantum and semiclassical @xmath3 while increasing the overall magnitude of @xmath3 ( see fig . [ fig:92rate ] ) .
this results in the overall larger fluorescence quantum efficiency for larger nps , as shown in figs .
[ fig:1100q ] and [ fig:92q ] .
although for relatively large np - dipole distances ( @xmath125 ) quantum efficiencies @xmath126 of 3.4 nm and 1.4 .
nm nps are comparable , being determined mainly by intrinsic molecular decay rates , for small distances ( @xmath15 ) the larger np @xmath126 is significantly higher ( see insets in figs .
[ fig:1100q ] and [ fig:92q ] ) . at the same time , the difference between quantum and semiclassical calculations of @xmath126 is also larger for @xmath110 nm np , indicating a larger role of non - local and screening effects .
the latter point is illustrated in fig .
( [ fig : power ] ) that compares the distance dependence of quantum @xmath3 with @xmath1 behavior near a flat surface predicted by rpa calculations.@xcite in fact , the raise of @xmath3 near np surface is significantly slower for either np size , but for the smaller @xmath111 nm np the deviation from @xmath1 behavior is considerably larger . as a final remark ,
the theory presented here is valid if there is no significant overlap between molecular orbitals and electron wave functions in the metal .
this condition no longer holds when molecule - np distance becomes several , i.e. , in the spillover region ( vertical dashed lines in figs .
[ fig:1100rate ] and [ fig:92rate ] ) . in this region ,
fluorescence quenching is dominated by electron tunneling rather than energy transfer .
we also did not attempt to quantify the effect of metal np on molecules ( or qds ) internal transitions and relaxation times which may play important role at very close distances.@xcite
in summary , were performed microscopic calculations of radiative and nonradiative decay rates of a fluorophore molecule situated near a small au nanoparticle in dielectric medium within quantum three - region model that incorporates non - local and quantum - size effects and accounts for different density profiles of d - band and sp - band electrons .
we found that , close to metal surface , the energy transfer rate from molecular dipole to nanoparticle is significantly ( by order of magnitude ) higher than that predicted by semiclassical electromagnetic models due to direct excitation of electron - hole pairs by dipole s coulomb potential .
however , the non - local and finite - size effects lead to a considerably slower dependence of energy transfer rate on molecule - np distance than @xmath1 behavior predicted by rpa calculations .
o. kulakovich , n. strekal , a. yaroshevich , s. maskevich , s. gaponenko , i. nabiev , u. woggon , and m. artemyev , nano lett . * 2 * , 1449 ( 2002 ) . z. gueroui , and a. libchaber , phys . rev
. lett . * 93 * , 166108 ( 2004 ) .
e. dulkeith , m. ringler , t. a. klar , j. feldmann , a. m. javier , and w. j. parak , nano lett .
* 5 * , 585 ( 2005 ) .
n. liu , b. s. prall , and v. i. klimov , j. am .
soc . * 128 * , 15362 ( 2006 ) .
p. p. pompa , l. martiradonna , a. d. torre , f. d. sala , l. manna , m. de vittorio , f. calabi , r. cingolani , and r. rinaldi , nature nanotech .
* 1 * , 126 ( 2006 ) .
p. anger , p. bharadwaj , and l. novotny , phys .
96 * , 113002 ( 2006 ) .
s. kuhn , u. hakanson , l. rogobete , and v. sandoghdar , phys .
lett . * 97 * , 017402 ( 2006 ) .
p. bharadwaj and l. novotny , opt .
express * 17 * , 14266 ( 2007 ) .
j. seelig , k. leslie , a. renn , s. k1uhn , v. jacobsen , m. van de corput , c. wyman , and v. sandoghdar , nano lett .
* 7 * , 685 ( 2007 ) .
k. aslan , m. wu , j. r. lakowicz , and c. d. geddes , j. am .
soc . * 129 * , 1524 ( 2007 ) . | we develop a microscopic model for fluorescence of a molecule ( or semiconductor quantum dot ) near a small metal nanoparticle . when a molecule is situated close to metal surface , its fluorescence is quenched due to energy transfer to the metal .
we perform quantum - mechanical calculations of energy transfer rates for nanometer - sized au nanoparticles and find that non - local and quantum - size effects significantly enhance dissipation in metal as compared to those predicted by semiclassical electromagnetic models
. however , the dependence of transfer rates on molecule s distance to metal nanoparticle surface , @xmath0 , is significantly weaker than the @xmath1 behavior for flat metal surface with a sharp boundary predicted by previous calculations within random phase approximation . |
it is well known that the theory of the classical local hardy spaces , originally introduced by goldberg @xcite , plays an important role in partial differential equations and harmonic analysis ; see , for example , @xcite and their references . in particular , pseudo - differential operators are bounded on local hardy spaces @xmath23 with @xmath24 $ ] , but they are not bounded on hardy spaces @xmath25 with @xmath24;$ ] see @xcite ( also @xcite ) . in @xcite , bui studied the weighted version @xmath26 of the local hardy space @xmath27 with @xmath28 , where and in what follows , @xmath29 for @xmath30 $ ] denotes the _ class of muckenhoup s weights _ ; see , for example , @xcite for their definitions and properties .
rychkov @xcite introduced and studied a class of local weights , denoted by @xmath31 ( see also definition [ d2.1 ] below ) , and the weighted besov - lipschitz spaces and triebel - lizorkin spaces with weights belonging to @xmath31 , which contains @xmath32 weights and exponential weights introduced by schott @xcite as special cases .
in particular , rychkov @xcite generalized some of the theory of weighted local hardy spaces developed by bui @xcite to @xmath31 weights .
in fact , rychkov established the local vertical and the local nontangential maximal function characterizations of weighted local hardy spaces with @xmath31 weights . very recently , tang @xcite established the weighted atomic decomposition characterization of the weighted local hardy space @xmath26 with @xmath33 via the local grand maximal function .
motivated by @xcite , tang also established some criterions for boundedness of @xmath34-sublinear operators on @xmath26 ( see section [ s6 ] for the notion of @xmath34-sublinear operators , which was first introduced in @xcite ) . as applications ,
tang @xcite proved that some strongly singular integrals , pseudo - differential operators and their commutators are bounded on @xmath26 .
it is worth pointing out that in recent years , many papers are focused on criterions for boundedness of ( sub)linear operators on various hardy spaces with different underlying spaces ( see , for example , @xcite ) , and on entropy and approximation numbers of embeddings of function spaces with muckenhoupt weight ( see , for example , @xcite ) .
it is also well known that the classical @xmath35 space ( the _ space of functions with bounded mean oscillation _ ) originally introduced by john and nirenberg @xcite and the classical morrey space originally introduced by morrey @xcite play an important role in the study of partial differential equations and harmonic analysis ; see , for example , @xcite .
in particular , fefferman and stein @xcite proved that @xmath36 is the dual space of the hardy space @xmath37 .
moveover , goldberg @xcite introduced the space @xmath38 and proved that @xmath38 is the dual space of the local hardy space @xmath39 . on the other hand , as the generalization of @xmath40 , the orlicz space was introduced by birnbaum - orlicz in @xcite and orlicz in @xcite , since then , the theory of the orlicz spaces themselves has been well developed and these spaces have been widely used in probability , statistics , potential theory , partial differential equations , as well as harmonic analysis and some other fields of analysis ; see , for example , @xcite .
moreover , orlicz - hardy spaces are also suitable substitutes of the orlicz spaces in dealing with many problems of analysis ; see , for example , @xcite .
recall that orlicz - hardy spaces and their dual spaces were studied by janson @xcite on @xmath41 and viviani @xcite on spaces of homogeneous type in the sense of coifman and weiss @xcite .
recently , orlicz - hardy spaces associated with some differential operators and their dual spaces were introduced and studied in @xcite .
let @xmath33 , @xmath42 be a concave function on @xmath43 of strictly lower type @xmath44 $ ] ( see below for the definition ) and @xmath45 for all @xmath46 , where @xmath47 is the inverse function of @xmath42 .
a typical example of such orlicz functions is @xmath48 for all @xmath46 and @xmath24 $ ] .
motivated by @xcite , in this paper , we introduce the weighted local orlicz - hardy space @xmath4 via the local grand maximal function .
we also introduce the @xmath7-type space @xmath8 and establish the duality between @xmath4 and @xmath8 .
characterizations of @xmath4 , including the atomic characterization , the local vertical and the local nontangential maximal function characterizations , are presented . using the atomic characterization , we prove the existence of finite atomic decompositions achieving the norm in some dense subspaces of @xmath4 ,
from which , we further deduce that for a given admissible triplet @xmath9 and a @xmath10-quasi - banach space @xmath11 with @xmath12 $ ] , if @xmath13 is a @xmath11-sublinear operator , and maps all @xmath9-atoms and @xmath14-single - atoms with @xmath15 ( or all continuous @xmath9-atoms with @xmath16 ) into uniformly bounded elements of @xmath11 , then @xmath13 uniquely extends to a bounded @xmath11-sublinear operator from @xmath4 to @xmath11 . as applications ,
we show that the local riesz transforms are bounded on @xmath4 , the local fractional integrals are bounded from @xmath49 to @xmath50 when @xmath19 and from @xmath49 to @xmath51 when @xmath21 , and some pseudo - differential operators are also bounded on both @xmath4 and @xmath8 .
we point out that the orlicz - hardy spaces @xmath52 include the classical local hardy spaces of goldberg @xcite , the weighted local hardy spaces of bui @xcite and the weighted local hardy spaces of tang @xcite as special cases .
moreover , the method of obtaining atomic decompositions used in this paper ( see the proof of theorem [ t5.6 ] below ) is different from the classical methods in @xcite . indeed , following bownik @xcite
( see also @xcite ) , we give a direct proof for the weighted atomic characterization of @xmath52 , without invoking the atomic characterization of @xmath53 .
all results of this paper for any general @xmath0 even when @xmath54 are new . precisely , this paper is organized as follows . in section [ s2 ] , we first recall some definitions and notation concerning local weights introduced in @xcite , then describe some basic assumptions and present some properties of orlicz functions considered in this paper . in section [ s3 ] ,
we first introduce the weighted local orlicz - hardy space @xmath55 via the local grand maximal function , and then the weighted atomic local orlicz - hardy space @xmath56 for any admissible triplet @xmath57 ( see definition [ d3.4 ] below ) .
we point out that when @xmath48 for all @xmath46 and @xmath24 $ ] , the weighted local orlicz - hardy space @xmath55 is just the weighted local hardy space @xmath58 introduced by tang in @xcite .
next , we establish the local vertical and the local nontangential maximal function characterizations of @xmath55 via a local caldern reproducing formula and some useful estimates established by rychkov @xcite , which generalizes ( * ? ? ?
* theorem 2.24 ) by taking @xmath48 for all @xmath46 and @xmath24 $ ] ; see theorems [ t3.12 ] and [ t3.14 ] and remark [ r3.13 ] below .
finally , we present some properties of these weighted local orlicz - hardy spaces @xmath55 and weighted atomic local orlicz - hardy spaces @xmath56 . throughout the whole paper , as usual , @xmath59 denotes the _ set of all @xmath60 functions on @xmath41 with compact support _ , endowed with the inductive limit topology , and @xmath61 its _ topological dual space _ , endowed with the weak @xmath62-topology .
let @xmath63 for any @xmath64 denote the _ maximal integer not more than @xmath65_. in section [ s4 ] , for any given @xmath66 , integer @xmath67 and @xmath68 , where @xmath69 , @xmath70 and @xmath71 are respectively as in , and below , and @xmath72 , following @xcite , via a whitney decomposition of @xmath73 in , we obtain the caldern - zygmund decomposition @xmath74 in @xmath61 of degree @xmath75 and height @xmath76 associated with the local grand maximal function @xmath71 .
the main task of section [ s4 ] is to establish some subtle estimates for @xmath77 and @xmath78 .
precisely , lemmas [ l4.2 ] through [ l4.5 ] are estimates on @xmath78 , the bad part of @xmath79 , while lemmas [ l4.6 ] and [ l4.7 ] on @xmath77 , the good part of @xmath79 . as an application of these estimates
, we obtain the density of @xmath80 in @xmath55 , where @xmath81 ( see corollary [ c4.8 ] below ) . differently from the proof of (
* lemma4.9 ) , via an application of the boundedness of the local vector - valued hardy - littlewood maximal operator obtained by rychkov @xcite ( see also lemma [ l3.10 ] below ) , our lemma [ l4.7 ] below improves ( * ? ? ?
* lemma4.9 ) even when @xmath48 for all @xmath46 and @xmath24 $ ] , which is necessary for corollary [ c4.8 ] . in section [ s5 ]
, we prove that for any given admissible triplet @xmath57 , @xmath82 with equivalent norms when positive integer @xmath83 ( see below for the definition of @xmath84 ) , by using the caldern - zygmund decomposition and some subtle estimates obtained in section [ s4 ] , which completely covers ( * ? ? ?
* theorem5.1 ) by taking @xmath48 for all @xmath24 $ ] and @xmath46 ; see theorem [ t5.6 ] and remark [ r5.7 ] below .
it is worth pointing out that we show theorem [ t5.6 ] by a way different from the methods in @xcite , but close to those in @xcite .
for simplicity , in the rest of this introduction , we denote by @xmath85 the _ weighted local orlicz - hardy space @xmath55 with _ @xmath83 .
assume that @xmath57 is an admissible triplet .
let @xmath86 be the _ space of all finite linear combinations of @xmath57-atoms and @xmath87-single - atoms _ ( see definition [ d6.1 ] below ) , and @xmath88 the _ space of all @xmath89 with compact support_. in section [ s6 ] , we prove that @xmath90 and @xmath91 are equivalent quasi - norms on @xmath86 when @xmath92 , and @xmath93 and @xmath91 are equivalent quasi - norms on @xmath94 when @xmath95 ( see theorem [ t6.2 ] below ) .
as an application , we prove that for a given admissible triplet @xmath9 and a @xmath10-quasi - banach space @xmath11 with @xmath12 $ ] , if @xmath42 has a upper type @xmath96 satisfying @xmath97 , and @xmath13 is a @xmath34-sublinear operator mapping all @xmath57-atoms and @xmath87-single - atoms with @xmath81 ( or all continuous @xmath9-atoms with @xmath16 ) into uniformly bounded elements of @xmath34 , then @xmath13 uniquely extends to a bounded @xmath34-sublinear operator from @xmath85 to @xmath34 which coincides with @xmath13 on these @xmath57-atoms and @xmath87-single - atoms ; see theorem [ t6.4 ] below .
we remark that this extends both the results of meda - sjgren - vallarino @xcite and yang - zhou @xcite to the setting of weighted local orlicz - hardy spaces .
we also point out that theorem [ t6.2](i ) and theorem [ t6.4](i ) below completely cover ( * ? ? ?
* theorem6.1 ) and ( * ? ? ?
* theorem 6.2 ) , respectively , by taking @xmath48 for all @xmath46 and @xmath24 $ ] ; and theorem [ t6.2](ii ) and theorem [ t6.4](ii ) below are new even for @xmath48 with @xmath46 and @xmath24;$ ] see remark [ r6.5 ] below .
let @xmath57 be an admissible triplet , @xmath98 the dual exponent of @xmath99 and @xmath69 the critical index of @xmath100 . in section [ s7 ]
, we introduce the @xmath35-type space @xmath101 and prove that @xmath102^{\ast}=\bmo_{\rz,\,{\overline}}^{q ' } ( \rn),\ ] ] where @xmath103^{\ast}$ ] denotes the dual space of @xmath104 ; see theorem [ t7.5 ] below .
denote @xmath105 simply by @xmath106 . as applications of theorems [ t5.6 ] and [ t7.5 ]
, we see that for @xmath107 , @xmath108 with equivalent norms and @xmath109^{\ast}=\bmo_{\rz,\,{\overline}}(\rn);\ ] ] see corollaries [ c7.6 ] and [ c7.7 ] below . in section [ s8 ] , as applications of theorem [ 6.2 ]
, we obtain the boundedness of some operators from @xmath85 to some @xmath10-quasi - banach space @xmath11 with @xmath110 $ ] .
first , we prove that the local riesz transforms are bounded on @xmath85 if @xmath111 and holds for @xmath112 with @xmath113 ( see section [ s2 ] below for the definitions of @xmath112 ) , which completely covers ( * ? ? ?
* lemma8.3 ) by taking @xmath114 for all @xmath46 ; see theorem [ t8.2 ] and remark [ r8.4 ] below .
then we introduce the local fractional integral operator @xmath115 and show that @xmath115 is bounded from @xmath49 to @xmath50 when @xmath116 , @xmath117 $ ] , @xmath118 , and @xmath119 for some @xmath120 and @xmath121^p\,dx=\fz$ ] ( see theorem [ t8.10 ] below ) ; furthermore , when @xmath122 , @xmath123 $ ] , @xmath118 , and @xmath100 satisfies the same conditions , we prove that @xmath115 is bounded from @xmath49 to @xmath51 ( see theorem [ t8.11 ] below ) . to the best of our knowledge ,
theorems [ t8.10 ] and [ t8.11 ] are new even when @xmath124 .
finally , let @xmath125 which was introduced by tang @xcite ( see also definition [ d8.13 ] below ) and @xmath13 be an @xmath126 pseudo - differential operator .
we prove that @xmath13 is bounded on @xmath85 if @xmath111 and holds for @xmath112 with @xmath113 ; see theorem [ t8.18 ] below .
we point out that @xmath127 but @xmath128 .
we also remark that theorem [ t8.18 ] below extends ( * ? ? ?
* theorem4 ) to the setting of weighted local orlicz - hardy spaces , and completely covers ( * ? ? ?
* theorem7.3 ) by taking @xmath48 for all @xmath46 and @xmath24 $ ] and also ( * ? ? ?
* theorem2 ) by taking @xmath129 for all @xmath46 and @xmath130 ; see remark [ r8.19 ] below . as an application of theorems [ t7.5 ] and [ t8.18 ]
, we also obtain that @xmath13 is bounded on @xmath106 ; see corollary [ c8.20 ] below .
a main motivation of this paper is to pave the way for the study of weighted orlicz - hardy spaces associated with divergence operators on strongly lipschitz domains of @xmath41 .
the corresponding hardy spaces associated with divergence operators on strongly lipschitz domains of @xmath41 were originally studied by auscher and russ @xcite , where the atomic characterization of the classical hardy spaces plays a key role .
earlier works on hardy spaces on domains can be found , for example , in @xcite .
it was shown in these papers that the theory of hardy spaces on domains plays an important role in partial differential equations and harmonic analysis .
finally we make some conventions on notation . throughout the whole paper ,
we denote by @xmath131 _ a positive constant _ which is independent of the main parameters , but it may vary from line to line .
we also use @xmath132 to denote _ a positive constant depending on the indicated parameters @xmath133 , @xmath134 , @xmath135_. the _ symbol _ @xmath136 means that @xmath137 . if @xmath138 and @xmath139 , then we write @xmath140 . the _ symbol _
@xmath141 for @xmath142 denotes the maximal integer not more than @xmath75 . for any given normed spaces @xmath143 and @xmath144 with the corresponding norms @xmath145 and @xmath146 , the _ symbol _
@xmath147 means that for all @xmath148 , then @xmath149 and @xmath150 . for any subset @xmath151 of @xmath41 , we denote by @xmath152 the _ set _ @xmath153 ; for a measurable set @xmath154 , denote by @xmath155 the _ characteristic function of _ @xmath154 .
we also set @xmath156 and @xmath157 .
for any @xmath158 , let @xmath159 and @xmath160 . given a function @xmath77 on @xmath41 , if @xmath161 , let @xmath162 ; otherwise , let @xmath163 be the _ maximal integer _ such that @xmath77 has vanishing moments up to order @xmath164 , namely , @xmath165 for all multi - indices @xmath166 with @xmath167 .
in this section , we first recall some notions and notation concerning local weights introduced in @xcite , then describe some basic assumptions and present some properties of orlicz functions considered in this paper . in this subsection , we recall some notions and properties of the local weights .
let @xmath169 be a cube in @xmath41 and we denote its lebesgue measure by @xmath170 . throughout the whole paper , _ all cubes
are assumed to be closed and their sides parallel to the coordinate axes_. [ d2.1 ] let @xmath171 .
the _ weight class @xmath168 _ is defined to be the set of all nonnegative locally integrable functions @xmath100 on @xmath41 such that @xmath172^{-p'/p}\,dy{\right})^{p / p'}<\fz,\ ] ] where the supremum is taken over all cubes @xmath173 with @xmath174 and @xmath175 . when @xmath176 , the _ weight class @xmath177 _ is defined to be the set of all nonnegative locally integrable functions @xmath100 on @xmath41 such that @xmath178^{-1}{\right})<\fz,\ ] ] where the supremum is taken over all cubes @xmath173 with @xmath174 .
when @xmath179 , the _ weight class @xmath180 _ is defined to be the set of all nonnegative locally integrable functions @xmath100 on @xmath41 such that for any @xmath122 , @xmath181<\fz,\ ] ] where @xmath182 runs through all measurable sets in @xmath41 with the indicated properties , the supremum is taken over all cubes @xmath173 with @xmath174 and @xmath183 . [ r2.2 ] ( i ) we point out that the weight class @xmath168 for @xmath184 $ ] was introduced by rychkov @xcite and @xmath185 by tang @xcite . by hlder s inequality , we see that @xmath186 , if @xmath187 .
conversely , it was proved in ( * ? ? ?
* lemma 1.3 ) that if @xmath188 , then @xmath189 for some @xmath171 .
thus , we have @xmath190 .
\(ii ) for any constant @xmath191 , the condition @xmath174 can be replaced by @xmath192 in , and ; see ( * ? ? ?
* remark1.5 ) . in this case , @xmath193 with @xmath194 and @xmath195 depend on @xmath196 . in what follows
, @xmath197 denotes the _ closed cube centered at @xmath198 and of the sidelength @xmath199_. similarly , given @xmath200 and @xmath201 , we write @xmath202 for the _ @xmath76-dilated cube _ , which is the cube with the same center @xmath198 and with sidelength @xmath203 . given a lebesgue measurable set @xmath154 and a weight @xmath204 , let @xmath205 . for any @xmath206
, the _ space _ @xmath207 with @xmath208 denotes the set of all measurable functions @xmath79 such that @xmath209 and @xmath210 .
the _ space @xmath211 _ denotes the set of all measurable functions @xmath79 such that @xmath212 for a positive constant @xmath196 , any locally integrable function @xmath79 and @xmath213 , the _ local hardy - littlewood maximal function @xmath214 _ is defined by @xmath215 where the supremum is taken over all cubes @xmath173 such that @xmath216 and @xmath192 .
if @xmath217 , we denote @xmath214 simply by @xmath218 .
next , we recall some properties for weights in @xmath31 and @xmath219 , where and in what follows , @xmath219 for @xmath194 denotes the classical _ muckenhoupt weights _ ; see @xcite for their definitions .
[ l2.3 ] @xmath220
let @xmath221 , and @xmath169 be a unit cube , namely , @xmath222 .
then there exist an @xmath223 such that @xmath224 on @xmath169 , and a positive constant @xmath131 independent of @xmath169 such that @xmath225 , where @xmath226 denotes the weight constant of @xmath227 , which is as in and by removing the restriction @xmath228 .
@xmath229 if @xmath230 with @xmath171 , then there exist @xmath231 such that @xmath232 with @xmath233 , and @xmath234 .
@xmath235 if @xmath187 , then @xmath236 .
@xmath237 for @xmath171 , @xmath238 if and only if @xmath239 , where @xmath240 @xmath241 for @xmath33 and @xmath242 , there exists a positive constant @xmath131 such that @xmath243 when @xmath244 , and @xmath245 when @xmath246 .
@xmath247 if @xmath171 and @xmath238 , then the local hardy - littlewood maximal operator @xmath248 is bounded on @xmath207 .
@xmath249 if @xmath250 , then @xmath248 is bounded from @xmath251 to @xmath211 .
@xmath252 if @xmath253 with @xmath194 , then there exists a positive constant @xmath131 such that for all cubes @xmath254 with @xmath255 , @xmath256 lemma [ l2.3](i ) is just ( * ? ? ?
* lemma1.1 ) .
the statements ( ii ) through ( vii ) of lemma [ l2.3 ] are just lemma 2.1 and corollary 2.1 in @xcite , which are deduced from lemma [ l2.3](i ) and the properties of @xmath219 ; see the proofs of ( * ? ? ?
* lemma2.1,corollary2.1 ) .
lemma [ l2.3](viii ) is included , for example , in @xcite .
[ r2.4 ] let @xmath196 be a positive constant .
it was pointed out in ( * ? ? ?
* remark1.5 ) and @xcite that ( i ) through ( vii ) of lemma [ l2.3 ] are also true , if @xmath222 , @xmath257 @xmath258 @xmath259 and @xmath248 are respectively replaced by @xmath260 , @xmath261 @xmath262 @xmath263 and @xmath264 . in this case , the constants appearing in ( i ) , ( vi ) and ( vii ) of lemma [ l2.3 ] depend on @xmath196 . for
any given @xmath33 , define the _
critical index of @xmath100 _ by @xmath265 [ r2.5 ] obviously , @xmath266 . if @xmath267 , by lemma [ r2.2](ii ) , it is easy to know that @xmath268 .
moreover , there exists an @xmath269 such that @xmath270 .
indeed , for @xmath271 , let @xmath272^{-1}$ ] .
johnson and neugebauer ( * ? ? ?
* , remark ) showed that @xmath273 . by the fact that @xmath274 for all @xmath194 , which is obvious by their definitions
, we see that @xmath275 .
we claim that @xmath276 .
in fact , taking @xmath277 , then we have @xmath278^{-1}\,dt\equiv\fz.\end{aligned}\ ] ] moreover , it is easy to see that @xmath279 as @xmath280 .
thus , by , we know that @xmath276 . for @xmath281 and @xmath282 , we have the following conclusions . [ l2.6 ]
let @xmath33 , @xmath69 be as in and @xmath283 $ ] .
@xmath220 if @xmath175 , then @xmath284 .
@xmath229 @xmath285 and the inclusion is continuous .
let @xmath286 and @xmath287 .
if @xmath81 , then for any @xmath288 , @xmath289 in @xmath282 as @xmath290 , where and in what follows , @xmath291 for all @xmath46 and @xmath213 .
we remark that ( i ) and ( ii ) of lemma [ l2.6 ] , and lemma [ l2.6](iii ) are , respectively , lemma 2.2 and proposition 2.1 in @xcite .
let @xmath42 be a positive function on @xmath292 .
the function @xmath42 is said to be of _ upper type @xmath293 ( resp .
lower type @xmath293 ) _ for some @xmath294 , if there exists a positive constant @xmath131 such that for all @xmath113 ( resp .
@xmath295 $ ] ) and @xmath296 , @xmath297
obviously , if @xmath42 is of lower type @xmath293 for some @xmath208 , then @xmath298 . so for the sake of convenience , if it is necessary , we may assume that @xmath299 .
if @xmath42 is of both upper type @xmath300 and lower type @xmath301 , then @xmath42 is said to be of _ type @xmath302_. let @xmath303 and @xmath304\,\ , \text{and}\,\,s\in(0,\fz)\}.\end{aligned}\ ] ] the function @xmath42 is said to be of _ strictly lower type @xmath293 _ if for all @xmath305 and @xmath296 , @xmath306 , and we define @xmath307 it is easy to see that @xmath308 for all @xmath42 . in
what follows , @xmath309 and @xmath310 are respectively called the _ strictly critical lower type index _ , the _ critical lower type index _ and the _ critical upper type index _ of @xmath42 .
we point out that if @xmath70 is defined as in , then @xmath42 is also of strictly critical lower type @xmath70 ; see @xcite for the proof . throughout the whole paper
, we always assume that @xmath42 satisfies the following assumption .
assumption ( a ) .
let @xmath42 be a positive function defined on @xmath311 , which is of strictly lower type and its strictly critical lower type index @xmath44 $ ] .
also assume that @xmath42 is continuous , strictly increasing , subadditive and concave .
notice that if @xmath42 satisfies assumption ( a ) , then @xmath299 and @xmath42 is obviously of upper type 1 . for any concave and positive function @xmath312 of strictly lower type @xmath293 ,
if we set @xmath313 for @xmath314 , then by ( * ? ? ?
* proposition 3.1 ) , @xmath42 is equivalent to @xmath312 , namely , there exists a positive constant @xmath131 such that @xmath315 for all @xmath314 ; moreover , @xmath42 is strictly increasing , concave , subadditive and continuous function of strictly lower type @xmath293 .
notice that all our results are invariant on equivalent functions satisfying assumption ( a ) . from this
, we deduce that all results in this paper with @xmath42 as in assumption ( a ) also holds for all concave and positive functions @xmath312 of the same strictly critical lower type @xmath70 as @xmath42 .
let @xmath42 satisfy assumption ( a ) and @xmath316 .
a measurable function @xmath79 on @xmath41 is called to belong to the _ space _ @xmath317 if @xmath318 .
moreover , for any @xmath319 , define @xmath320 since @xmath42 is strictly increasing , we define the _ function @xmath321 _ on @xmath311 by setting , for all @xmath46 , @xmath322 where @xmath47 is the inverse function of @xmath42 . then the types of @xmath42 and @xmath321 have the following relation .
let @xmath323 and @xmath42 be an increasing function , then @xmath42 is of type @xmath324 if and only if @xmath321 is of type @xmath325 see @xcite for its proof .
moreover , it is easy to see that for all @xmath46 , @xmath326 which is used in what follows .
in this section , we introduce the weighted local orlicz - hardy space @xmath55 via the local grand maximal function and establish its local vertical and nontangential maximal function characterizations via a local caldern reproducing formula and some useful estimates obtained by rychkov @xcite .
we also introduce the weighted atomic local orlicz - hardy space @xmath56 and give some basic properties of these spaces .
[ d3.1 ] let @xmath327 and @xmath328 .
for any @xmath66 , the _ local nontangential grand maximal function _
@xmath330 of @xmath79 is defined by setting , for all @xmath213 , @xmath331 and the _ local vertical grand maximal function _
@xmath332 of @xmath79 is defined by setting , for all @xmath213 , @xmath333 for convenience s sake , when @xmath334 , we denote @xmath335 , @xmath330 and @xmath336 simply by @xmath337 , @xmath338 and @xmath339 , respectively ; when @xmath340 , we denote @xmath335 , @xmath330 and @xmath336 simply by @xmath341 , @xmath342 and @xmath343 , respectively . for any @xmath327 and @xmath213 , obviously , @xmath344 [ d3.3 ] let @xmath42 satisfy assumption ( a ) , @xmath33 , @xmath69 and @xmath70 be respectively as in and , and @xmath357 .
for each @xmath358 with @xmath359 , the _ weighted local orlicz - hardy space _ is defined by @xmath360 moreover , let @xmath361 .
we remark that when @xmath48 for all @xmath46 and @xmath24 $ ] , @xmath55 above is the weighted local hardy space @xmath58 introduced by tang @xcite .
obviously , for any integers @xmath362 and @xmath363 with @xmath364 , @xmath365 and the inclusions are continuous .
we also point out that theorem [ t3.14 ] below further implies that @xmath366 for all @xmath358 with @xmath367 ( see for the definition of @xmath84 ) .
[ d3.4 ] let @xmath42 satisfy assumption @xmath368 , @xmath369 and @xmath69 , @xmath321 be respectively as in and .
a triplet @xmath57 is called _ admissible _ if @xmath370 $ ] , @xmath371 and @xmath372 . a function @xmath373 on @xmath41
is called a _ @xmath57-atom _ if there exists a cube @xmath173 such that @xmath235 @xmath376 for all @xmath377 with @xmath378 , when @xmath244 .
+ moreover , a function @xmath373 on @xmath41 is called a _
@xmath87-single - atom _ with @xmath370 $ ] , if @xmath379^{\frac{1}{q}-1 } [ \rz({\overline}(\rn))]^{-1}.\ ] ] [ d3.5 ] let @xmath42 satisfy assumption ( a ) , @xmath33 , @xmath69 and @xmath321 be respectively as in and , and @xmath57 be admissible .
the _ weighted atomic local orlicz - hardy space _
@xmath56 is defined be the set of all @xmath66 satisfying that @xmath382 in @xmath61 , where @xmath383 are @xmath57-atoms with @xmath384 , @xmath385 is a @xmath87-single - atom , @xmath386 , and @xmath387 moreover , letting @xmath388 the _ quasi - norm _ of @xmath389 is defined by @xmath390 where the infimum is taken over all the decompositions of @xmath79 as above .
[ r3.6 ] ( i ) notice that the definition of @xmath391 above is different from @xcite .
in fact , if @xmath24 $ ] and @xmath48 for all @xmath46 , then @xmath392 coincides with @xmath393 and @xmath394 satisfy @xmath395 , where @xmath396 . by the fact that @xmath42 is subadditive and of strictly lower type @xmath70
, we have @xmath397^{p_{\bfai}}\le\sum_{k=1}^2 \left[\blz(\{\lz^k_i a^k_i\}_{i\in\zz_+}){\right}]^{p_{\bfai}}.\ ] ] next , we introduce some local vertical , tangential and nontangential maximal functions , and then establish the characterizations of the weighted local orlicz - hardy space @xmath55 on these local maximal functions . [ d3.7 ] let @xmath401 for @xmath402 , @xmath403 and @xmath404 , let @xmath405 .
the _ local vertical maximal function _
@xmath406 of @xmath79 associated to @xmath407 is defined by setting , for all @xmath213 , @xmath408 the _ local tangential peetre - type maximal function @xmath409 _ of @xmath79 associated to @xmath407 is defined by setting , for all @xmath213 , @xmath410 and the _ local nontangential maximal function @xmath411 _ of @xmath79 associated to @xmath407 is defined by setting , for all @xmath213 , @xmath412 where and in what follows , for all @xmath213 , @xmath413 for all @xmath402 and @xmath414 for all @xmath46 . in order to establish the local vertical and the local nontangential maximal function characterizations of @xmath55
, we first establish some relations in the norm of @xmath317 of the local maximal functions @xmath416 and @xmath417 , which further imply the desired characterizations .
we begin with some technical lemmas .
[ r3.9 ] let @xmath423 and @xmath424 be as in lemma [ l3.8 ] . from the proof of (
* theorem 1.6 ) , it is easy to deduce that for any @xmath402 and @xmath66 , @xmath425 in @xmath61 ( see also ( * ? ? ? * ( 2.11 ) ) ) .
we omit the details .
[ l3.10 ] let @xmath171 , @xmath430 $ ] , and @xmath431 .
then there exists a positive constant @xmath131 such that for any sequence @xmath432 of measurable functions , @xmath433 where and in what follows , @xmath434 also , there exists positive constants @xmath131 and @xmath435 such that for all @xmath436 , @xmath437 lemma [ l3.10 ] is just ( * ? ? ?
* lemma2.11 ) .
moreover , from the proof of ( * ? ? ?
* lemma2.11 ) , it is easy to deduce that also holds for @xmath264 with any given positive constant @xmath196 . in this case
, the positive constant @xmath131 in lemma [ l3.10 ] depends on @xmath196 .
[ l3.11 ] let @xmath407 be as in and @xmath438 .
then there exists a positive constant @xmath439 depending only on the support of @xmath407 such that for any @xmath440 and @xmath427 , there exists a positive constant @xmath131 , depending only on @xmath441 and @xmath442 , satisfying that for all @xmath66 , @xmath213 and @xmath402 , @xmath443^r&\le & c\sum_{k = j}^{\fz } 2^{(j - k)(ar - n)}\left\{m^{\loc}(|(\pz_0 ) _
k \ast f|^r)(x){\right}.\\ & & + k_{br}(|(\pz_0 ) _ k \ast f|^r)(x)\big\},\end{aligned}\ ] ] where @xmath444 for all @xmath213 .
lemma [ l3.11 ] is a modified version of ( * ? ? ?
* lemma2.10 ) , and was essentially obtained by rychkov in the proof of ( * ? ? ?
* theorem2.24 ) .
let @xmath445 be as in lemma [ l3.8 ] .
indeed , rychkov @xcite showed lemma [ l3.11 ] under the assumption that @xmath446 , namely , there exist positive constant @xmath447 and nonnegative integer @xmath448 such that for all @xmath449 , @xmath450 which guarantees that for all @xmath213 and @xmath402 , @xmath451 by ( * ? ? ?
* proposition2.3.4(a ) ) , we have that for any @xmath66 , @xmath452 for all @xmath213 and @xmath402 , provided @xmath453 , where @xmath439 is a positive constant depending only on the support of @xmath407 .
thus , lemma [ l3.11 ] is true for all @xmath66 .
this finishes the proof of lemma [ l3.11 ] .
[ t3.12 ] let @xmath42 satisfy assumption @xmath368 , @xmath369 , @xmath328 , @xmath454 and @xmath70 be respectively as in , and @xmath455 , @xmath456 , and @xmath330 be respectively as in , and .
let @xmath457 @xmath458 and integer @xmath459 , where @xmath439 and @xmath460 are respectively as in lemmas [ 3.3 ] and [ 3.2 ] .
then for any @xmath461 , @xmath462 and integer @xmath463 , there exists a positive constant @xmath131 , depending only on @xmath464 and @xmath465 , such that for all @xmath66 , @xmath466 and @xmath467 let @xmath66 .
first , we prove .
let @xmath461 and @xmath462 .
by @xmath468 and @xmath469 , we know that there exists @xmath470 such that @xmath471 and @xmath472 , where @xmath439 and @xmath460 are respectively as in lemmas [ 3.3 ] and [ l3.10 ] . thus , by lemma [ l3.11 ] , for all @xmath213 , we have @xmath473^{r_0}&\ls&\sum_{k = j}^{\fz } 2^{(j - k)(ar_0 -n)}\big\{m^ { \loc}\left(|(\pz_0)_k\ast f|^{r_0}{\right})(x ) \nonumber\\ & & + k_{br_0}\left(|(\pz_0)_k\ast f|^{r_0}{\right})(x)\big\}.\end{aligned}\ ] ] let @xmath474 and @xmath409 be respectively as in and .
we notice that for any @xmath213 and @xmath475 , @xmath476 which together with implies that for all @xmath213 , @xmath477^{r_0}\ls m^{\loc}\left([\pz^{+}_0 ( f)]^{r_0})(x)+k_{br_0}([\pz^{+}_0 ( f)]^{r_0}{\right})(x).\end{aligned}\ ] ] by and the subadditivity of @xmath42 , we have @xmath478^{r_0}{\right})(x){\right}\}^{1/{r_0}}{\right}){\overline}(x)\,dx\nonumber\\ & & \hs\hs+\int_{\rn}\bfai\left(\left\{k_{br_0}\left([\pz^{+}_0 ( f)]^{r_0}{\right})(x){\right}\}^{1/{r_0}}{\right}){\overline}(x)\,dx \equiv\mathrm{i_1}+\mathrm{i_2}.\end{aligned}\ ] ] first , we estimate @xmath479 .
by @xmath480 , we know that there exists @xmath81 such that @xmath481 and @xmath482 . for any @xmath483 and @xmath484 , let @xmath485 it is easy to see that @xmath486 which together with lemma [ l2.3](vi ) implies that @xmath487^q{\overline}(x)\,dx\nonumber\\ & & \hs\ls\frac{1}{\az^q}\int_{\rn}|g_2 ( x)|^q{\overline}(x)\,dx \sim\frac{1}{\az^q}\int_{\{x\in\rn:\,|g(x)|>\az\}}|g(x)|^q{\overline}(x)\,dx.\end{aligned}\ ] ] thus , for any @xmath483 , by , we have @xmath488^{r_0}{\right})(x){\right}]^{1/{r_0}}>\az{\right}\}{\right})\nonumber\\ & & \hs\ls\frac{1}{\az^{r_0 q}}\int_{\{x\in\rn:\,[\pz^{+}_0 ( f)(x)]^{r_0}>\frac{\az^{r_0}}{2}\}}[\pz^{+}_0 ( f)(x)]^{r_0 q}{\overline}(x)\,dx\nonumber\\ & & \hs\sim\sz_{\pz^+_0 ( f)}\left(\frac{\az}{2^{1/{r_0}}}{\right})+\frac{1}{\az^{r_0 q}}\int^{\fz}_{\frac{\az}{2^{1/{r_0}}}}r_0 qs^{r_0 q-1}\sz_{\pz^+_0 ( f)}(s)\,ds,\end{aligned}\ ] ] where and in what follows , @xmath489 from the fact that @xmath42 is concave and of lower type @xmath70 , we infer that @xmath490 for all @xmath46 . by this , and the lower type @xmath70 property of @xmath42 , the fact @xmath481 and fubini s theorem , we have @xmath491^{r_0})(x){\right}\}^{1/{r_0}}}_0\frac{\bfai(t)}{t}\,dt{\right}\}{\overline}(x)\,dx \nonumber\\ & \sim&\int^{\fz}_0\frac{\bfai(t)}{t}\sz_{\left\{m^{\loc}([\pz^{+}_0 ( f)]^{r_0}){\right}\}^{1/{r_0}}}(t)\,dt\nonumber\\ & \ls&\int^{\fz}_0\frac{\bfai(t)}{t}\left\{\sz_{\pz^+_0 ( f)}\left(\frac{t}{2^{1/{r_0}}}{\right})+\frac{1}{t^{r_0 q } } \int^{\fz}_{\frac{t}{2^{1/{r_0}}}}r_0 qs^{r_0 q-1}\sz_{\pz^{+}_0 ( f)}(s)\,ds{\right}\}\,dt\nonumber\\ & \ls & \mathrm{j}_f+\int^{\fz}_0 r_0 q s^{r_0 q-1}\sz_{\pz^{+}_0 ( f)}(s ) \left\{\int_0^{2^{\frac{1}{r_0 } } s}\frac{\bfai(t)}{t}\frac{1}{t^{r_0 q}}\,dt{\right}\}\,ds\nonumber\\ & \sim&\mathrm{j}_f + \int^{\fz}_0 r_0 qs^{r_0 q-1}\sz_{\pz^{+}_0 ( f)}(s)\bfai(2^{\frac{1}{r_0 } } s)\left\{\int_0^{2^{\frac{1}{r_0 } } s}\left(\frac{t}{2^{\frac{1}{r_0 } } s}{\right } ) ^{p_{\bfai}}\frac{1}{t^{r_0 q+1}}\,dt{\right}\}\,ds\nonumber\\ & \sim&\mathrm{j}_f\sim\int_{\rn}\bfai\left(\pz^+_0 ( f)(x){\right}){\overline}(x)\,dx,\end{aligned}\ ] ] where @xmath492 .
next , we estimate @xmath493 . for any @xmath483 and @xmath494 , let @xmath495 and @xmath496 be as above . for @xmath497 ,
let @xmath498 .
it is easy to see that for all @xmath213 , @xmath499 , which implies that @xmath500 where @xmath501 is as in .
thus , by lemma [ l3.10 ] , we have @xmath502 similarly to , from the above estimate , @xmath503 and lemma [ 3.2 ] , we also deduce that @xmath504^{r_0})(x){\right}]^{1/{r_0}}>\az{\right}\}{\right})\\ & & \hs\ls\sz_{\pz^+_0 ( f)}\left(\frac{\az}{(c_{br_0}+1)^{1/r_0}}{\right})+\frac{1}{\az^{r_0 q}}\int^{\fz}_{\frac{\az}{(c_{br_0}+1)^{1/r_0}}}r_0 qs^{r_0 q-1}\sz_{\pz^+_0 ( f)}(s)\,ds.\end{aligned}\ ] ] by this , similarly to the estimate of @xmath479 , we also have @xmath505 thus , we deduce from , and that @xmath506 replacing @xmath79 by @xmath507 with @xmath201 in the above inequality , and noticing that @xmath508 and @xmath509 , we have @xmath510 which together with the arbitrariness of @xmath201 implies .
now , we prove .
by @xmath511 , we know that there exists @xmath461 such that @xmath512 . in the rest of this proof , we fix @xmath461 satisfying @xmath512 and @xmath462 .
let integer @xmath463 and @xmath328 .
for any @xmath513 , @xmath305 and @xmath402 , from lemma [ l3.8 ] and remark [ r3.9 ] , it follows that @xmath514 where @xmath420 with @xmath515 and @xmath445 is as in lemma [ l3.8 ] . to estimate @xmath528 , by the definition of @xmath445 , it is easy to know that for any @xmath529 , @xmath530 by the definition of @xmath409 and the facts that @xmath531 for any @xmath529 and @xmath532 , we conclude that @xmath533 similarly , we also have @xmath534 thus , @xmath535 from and the above estimates of @xmath521 and @xmath528 , it follows that @xmath536 assume that @xmath537
. then @xmath538 for all @xmath402 .
moreover , by @xmath539 and @xmath516 , we see that @xmath540 from this , we further deduce that @xmath541 and @xmath542 which implies that latexmath:[\ ] ] which together with implies in case 2 ) .
this finishes the proof of theorem [ t8.18 ] .
[ r8.19 ] let @xmath24 $ ] .
theorem [ t8.18 ] when @xmath124 and @xmath48 for all @xmath46 was obtained by goldberg ( * ? ? ? * theorem4 ) ; moreover , theorem [ t8.18 ] when @xmath48 for all @xmath46 was obtained by tang ( * ? ? ?
* theorem7.3 ) . also , theorem [ t8.18 ] when @xmath130 and @xmath129 for all @xmath46 was obtained by lee , c - c .
lin and y .- c .
lin ( * ? ? ?
* theorem2 ) .
[ c8.20 ] let @xmath13 be an @xmath126 pseudo - differential operator , @xmath42 satisfy assumption @xmath368 , @xmath1724 and @xmath70 be as in .
if @xmath111 and @xmath42 is of upper type @xmath112 , then there exists a positive constant @xmath1768 such that for all @xmath1769 , @xmath1770 | let @xmath0 be a concave function on @xmath1 of strictly lower type @xmath2 $ ] and @xmath3 ( the class of local weights introduced by v. s. rychkov ) .
we introduce the weighted local orlicz - hardy space @xmath4 via the local grand maximal function .
let @xmath5 for all @xmath6 .
we also introduce the @xmath7-type space @xmath8 and establish the duality between @xmath4 and @xmath8 .
characterizations of @xmath4 , including the atomic characterization , the local vertical and the local nontangential maximal function characterizations , are presented . using the atomic characterization
, we prove the existence of finite atomic decompositions achieving the norm in some dense subspaces of @xmath4 , from which , we further deduce that for a given admissible triplet @xmath9 and a @xmath10-quasi - banach space @xmath11 with @xmath12 $ ] , if @xmath13 is a @xmath11-sublinear operator , and maps all @xmath9-atoms and @xmath14-single - atoms with @xmath15 ( or all continuous @xmath9-atoms with @xmath16 ) into uniformly bounded elements of @xmath11 , then @xmath13 uniquely extends to a bounded @xmath11-sublinear operator from @xmath4 to @xmath11 . as applications ,
we show that the local riesz transforms are bounded on @xmath4 , the local fractional integrals are bounded from @xmath17 to @xmath18 when @xmath19 and from @xmath17 to @xmath20 when @xmath21 , and some pseudo - differential operators are also bounded on both @xmath4 .
all results for any general @xmath0 even when @xmath22 are new .
[ multiblock footnote omitted ] |
we use boyer - lindquist coordinates @xmath34 , choose units so that the mass of the black hole is @xmath35 , and choose the value of the spin parameter to be @xmath36 .
the event @xmath33 is @xmath37 , and @xmath25 is @xmath38 .
the contravariant components of @xmath22 s four - velocity at @xmath33 are @xmath39 while those of @xmath23 are @xmath40 these four - velocities were chosen to be close to the case where @xmath41 and re - intersection is guaranteed ( although one can show that @xmath42 in this special case ) .
we choose the masses of @xmath22 and @xmath23 to be unity , so that @xmath43 .
the interval of proper time from @xmath33 to @xmath25 for @xmath22 is @xmath44 , while for @xmath23 it is @xmath45 .
our numerical integration code conserves the quantities @xmath46 , @xmath47 , and @xmath48 along each geodesic to within one part in @xmath49 , and each geodesic reaches the event @xmath50 to within one part in @xmath49 as well . with these choices
the initial carter constant is @xmath51 , while the final one is @xmath52 .
the fractional difference is @xmath53 which is nonzero .
we have also constructed several other numerical examples .
those examples are sufficient to prove a slight generalization of the result in the body of the paper : there is no general conservation law for conserved stress - energy tensors associated with any killing tensor of the form @xmath54 for any value of @xmath55 .
in this appendix we show that the expression ( [ ans ] ) reduces to the carter constant of a point particle when the stress energy tensor is taken to be that of a point particle .
if the particle s worldline is written as @xmath56 , then this stress energy tensor is t^(x ) = d p^ ( ) p^ ( ) , where @xmath57 .
we now insert this into the expression ( [ ans ] ) .
we use gaussian normal coordinates adapted to the surface @xmath17 , so that the surface is @xmath58 , the metric is @xmath59 , and @xmath60 with @xmath61 the unit normal .
the result is @xmath62 \nonumber \\ \fl \ \ \ \
\times p^\alpha(\lambda ) p^\beta(\lambda ) p^{\alpha'}(\lambda ' ) p^{\beta'}(\lambda ' ) \delta[t - t(\lambda ) ] \delta[t ' - t(\lambda ' ) ] \frac{\delta^3[{\bf x } - { \bf z}(\lambda)]}{\sqrt{h } } \frac{\delta^3[{\bf x } ' - { \bf z}(\lambda')]}{\sqrt{h'}}. \label{expr1}\end{aligned}\ ] ] evaluating the time integrals of the delta functions give factors of @xmath63 , and so @xmath16 reduces to @xmath64 evaluated at the point where the worldline crosses the surface . since we are assuming that the bitensor reduces to the killing tensor in the coincidence limit , this is just the carter constant .
10 leor barack .
gravitational self - force : orbital mechanics beyond geodesic motion . in ji
bik and tom ledvinka , editors , _ general relativity , cosmology and astrophysics _ , volume 177 of _ fundamental theories of physics _ , pages 147168 .
springer international publishing , 2014 . | freely falling point particles in the vicinity of kerr black holes are subject to a conservation law , that of their carter constant .
we consider the conjecture that this conservation law is a special case of a more general conservation law , valid for arbitrary processes obeying local energy momentum conservation . under some fairly general assumptions we prove that the conjecture is false : there is no conservation law for conserved stress - energy tensors on the kerr background that reduces to conservation of carter for a single point particle .
the validity of conservation laws in physics can depend on the type of interactions being considered .
for example , baryon number minus lepton number , @xmath0 , is an exact conserved quantity in the standard model of particle physics .
however , it is no longer conserved when one enlarges the set of interactions under consideration to include those of many grand unified theories @xcite , or those that occur in quantum gravity @xcite .
the subject of this note is the conserved carter constant for freely falling point particles in black hole spacetimes @xcite .
as is well known , this quantity is not associated with a spacetime symmetry or with noether s theorem .
instead it is obtained from a symmetric tensor @xmath1 which obeys the killing tensor equation @xmath2 @xcite , via @xmath3 , where @xmath4 is four - momentum .
is this conservation law specific to freely falling particles , or does it persist in the presence of more general interactions ?
the carter constant conservation law has already been generalized in a variety of directions : * it applies to charged particles in rotating charged black hole spacetimes @xcite . *
it has been generalized to spinning test particles , to linear order in the spin @xcite . * for a free scalar field on the kerr background , carter showed that the differential operator @xmath5 commutes with the dalembertian , which implies the existence of a conserved quantity for the field @xcite . for a free complex scalar field @xmath6 ,
the charge associated with the conserved current @xmath7/2 $ ] is a generalization of the carter constant for particle motion , in the following sense .
solutions of the form @xmath8 $ ] in the eikonal limit @xmath9 can be interpreted as streams of particles , and the conserved charge is just the sum of the carter constants of the particles .
this is valid for both massive and massless fields .
* there is also a conserved quantity for spin @xmath10 fields related to the killing tensor . on the kerr spacetime
the killing tensor can be expressed as the square of an antisymmetric killing - yano tensor @xmath11 , for which @xmath12 and @xmath13 @xcite .
carter and mclenaghan @xcite showed that the operator @xmath14 commutes with the dirac operator and so gives rise to a conserved quantity .
a similar construction for spin-1 fields can be found in ref .
@xcite . *
recently , ashtekar and kesavan have shown that in spacetimes which settle down at late times to a kerr black hole , the killing tensor at future null infinity can be expressed as a linear combination of products of asymptotic symmetry vector fields ( bms generators ) , allowing them to compute a charge associated with any cut and derive an asymptotic conservation law @xcite .
* there is no known general local conservation law associated with the killing tensor for spin 2 fields in kerr , that is , for linearized vacuum perturbations .
however , there are hints that a conserved current may exist in this case .
specifically , for the radiation - reaction inspirals of point particles into black holes , it is possible to compute the time averaged time derivative of the particle s carter constant @xcite .
the result consists of two terms , a term involving the amplitudes of the gravitational wave modes at future null infinity that is naturally interpreted as a flux to infinity , and a term involving the amplitudes of the modes at the horizon that is naturally interpreted as a flux down the horizon . in this note
we consider a different possible type of generalization of the conservation law , the possibility that the carter quantity may be conserved under local interactions between particles that obey local stress energy conservation .
specifically , suppose we are given a conserved symmetric tensor @xmath15 on the kerr spacetime , with compact spatial support .
does there exist a quantity @xmath16 which can be computed from @xmath15 and its derivatives on any cauchy surface @xmath17 , which has the properties that ( i ) @xmath16 is independent of @xmath17 , and ( ii ) @xmath16 reduces to the carter constant for a single point particle ?
for example , one could consider quadratic functionals of the form ) is intended as an illustrative example only ; the argument below is not restricted to functionals of this form . ] _
= _ d^3 _
a(x ) _ d^3 _ a(x ) t^ab(x ) k_bb(x , x ) t^ab(x ) .
[ ans ] such functionals satisfy property ( ii ) if the bitensor @xmath18 reduces to the killing tensor @xmath1 in the coincidence limit @xmath19 ( see [ explicit ] ) .
of course , they do not necessarily satisfy property ( i ) .
if a conserved quantity of this type existed , one could say that the carter constant is conserved not just for freely falling , non - interacting particles , but also under general processes that obey local stress energy conservation .
no such conservation law exists .
there is a very simple argument which shows this , which is as follows .
consider a freely falling particle @xmath20 in the kerr spacetime .
suppose that at some point @xmath21 , it splits into two particles @xmath22 and @xmath23 , obeying energy - momentum conservation .
thus the four - momenta @xmath24 of the particles are related by _
a(p ) = k_b(p ) + k_c(p ) .
suppose that the initial four - momenta and masses of @xmath22 and @xmath23 are carefully chosen so that these two particles , after falling freely for some time , come together again at some other spacetime point @xmath25 .
at that point the two particles combine to form a fourth particle , @xmath26 , again obeying energy - momentum conservation : _
d(q ) = k_b(q ) + k_c(q ) . between the collisions , the carter constants of the individual particles
@xmath27 etc , of the individual particles are locally conserved . now
, if a general conservation law of the type discussed above existed , then at early times the putative conserved quantity @xmath28 would be simply the carter constant of @xmath20 , i.e @xmath29 similarly at late times would depend on the details of how the conserved quantity is supposed to be computed .
for example , in the model ( [ ans ] ) it would depend on values of the bitensor @xmath30 away from the coincidence limit .
this is why we need to consider a process that begins and ends with a single particle , in order to get a clean test of the conjecture . ]
it would be @xmath31 however , by numerically integrating the geodesic equations in kerr , it is easy to find examples of scenarios of this kind where the initial and final values ( [ eq : initial ] ) and ( [ eq : final ] ) do _ not _ coincide ( see [ numerical ] ) .
this occurs whenever the cross term @xmath32 has different values at @xmath33 and @xmath25 . in conclusion ,
generalizations of the carter constant conservation law are relevant to the program of computing gravitational wave signals from point particles inspiralling into spinning black holes@xcite .
one might have hoped for a unified conservation law for `` carter '' , analogous to that for energy , that some combination of the carter constant of the particle and the `` carter '' in the gravitational wave field be conserved .
the result presented here suggests that no such unified law exists .
this research was supported in part by nsf grants phy-1404105 and phy-1068541 .
we thank scott hughes , justin vines , david nichols and leo stein for helpful conversations . |
let @xmath1 be the free group of rank @xmath2 with the set of generators @xmath3 .
assume further that @xmath4 is the automorphism group of @xmath1 .
the artin braid group @xmath5 can be represented as a subgroup of @xmath4 .
this representation , due to artin himself , is defined associating to any generator @xmath6 , for @xmath7 , of @xmath5 the following automorphism of @xmath1 : @xmath8 moreover ( see for instance ( * ? ? ?
* theorem 5.1 ) ) , any automorphism @xmath9 of @xmath4 corresponds to an element of @xmath5 if and only if @xmath9 satisfies the following conditions : @xmath10 where @xmath11 is a permutation from the symmetric group @xmath12 and @xmath13 .
generalisations of artin s representation have been provided by wada @xcite and further by crisp and paris @xcite , in order to construct group invariants of oriented links .
another interesting representation of @xmath5 is the _ braid monodromy representation _ of @xmath5 into @xmath14 ( see last section ) , that was proven to be faithful in @xcite and @xcite . in this paper , we extend artin s and braid monodromy representations to some generalisations of braid groups . in the case of braid groups of oriented surfaces with boundary components our representations are faithful ( theorem [ theorem3annexe ] ) .
moreover , in the case of surfaces of genus @xmath15 , the induced representations of surface braid groups as outer automorphisms hold faithful ( theorem [ out ] ) . in the case of closed surfaces ,
as earlier remarked in @xcite for braid groups of the sphere , we can not extend artin s representation and we provide a representation in the outer automorphism group of a finitely generated free group . in the last section we consider the braid monodromy representation of @xmath5 in @xmath14 and we provide a faithful representation of the @xmath2-th artin - tits group of type @xmath0 in @xmath16 ( proposition [ thmartind ] ) .
* acknoledgments . *
the research of the first author has been supported by the scientific and research council of turkey ( tbitak ) .
the first author would like to thank prof .
hasan gmral , prof .
vladimir tolstykh and other members of the department of mathematics of yeditepe university ( istanbul ) for their kind hospitality .
we thank prof .
warren dicks for pointing out a doubtful argument on a previous version of proposition [ thmartind ] and for suggesting us a different approach for an easier proof .
* surface braids as collections of paths .
* let @xmath17 be a set of @xmath2 distinct points ( _ punctures _ ) in the interior of a surface @xmath18 .
a _ geometric braid _ on @xmath18 based at @xmath19 is a collection @xmath20 of @xmath2 disjoint paths ( called _ strands _ ) on @xmath21 $ ] which run monotonically with @xmath22 $ ] and such that @xmath23 and @xmath24 .
two braids are considered to be equivalent if they are isotopic relatively to the base points .
the usual product of paths defines a group structure on the equivalence classes of braids .
this group , denoted usually by @xmath25 , does not depend on the choice of @xmath19 and it is called braid group on @xmath2 strands of @xmath18 .
the @xmath2th braid group of the disk @xmath26 , @xmath27 , is isomorphic to @xmath5 . in the following
we will denote by @xmath28 the braid group on @xmath2 strands of an orientable surface of genus @xmath29 with @xmath30 boundary components ( we set @xmath31 ) and by @xmath32 the braid group on @xmath2 strands of a non - orientable surface of genus @xmath29 with @xmath30 boundary components ( we set @xmath33 ) . in this section
we will consider an orientable surface @xmath34 of genus @xmath35 and with @xmath36 boundary components .
we set also @xmath37 .
we denote by @xmath38 the standard generators of the braid group @xmath5 .
since @xmath36 , we can embed a disk in @xmath34 and therefore we can consider @xmath38 as elements of @xmath28 .
let also @xmath39 be the generators of @xmath40 , where @xmath41 s denote loops around the holes .
assume that the base point of the fundamental group is the startpoint of the first strand .
then each element @xmath42 determines an element denoted also by @xmath43 in @xmath28 , by considering the braid whose first strand is describing the curve @xmath43 and other strands are constant .
let us set @xmath44 and @xmath45 for @xmath46 .
according to @xcite we have that the group @xmath47 admits a presentation with generators : @xmath48 and defining relations : braid relations : @xmath49 @xmath50 mixed relations : ( r1 ) @xmath51 @xmath52 @xmath53 ( r2 ) @xmath54 @xmath53 ( r3 ) @xmath55 , @xmath56 , @xmath57 , ( r4 ) @xmath58 @xmath59 ( r5 ) @xmath60 @xmath52 @xmath61 ( r6 ) @xmath62 @xmath63 @xmath64 ( r7 ) @xmath65 @xmath66 ( r8 ) @xmath67 @xmath68 associating to any surface braid the corresponding permutation one obtains a surjective homomorphism @xmath69 such that @xmath70 , @xmath71 for @xmath72 and @xmath73 . in @xcite
the second author considered the subgroup @xmath74 and found its generators .
we provide a set of defining relations of @xmath75 using the well - known reidemeister - schreier s method ( see ( * ? ? ?
2 ) ) . let @xmath76 be the set defined as follows : @xmath77 it is easy to prove ( see @xcite ) that @xmath78 and that @xmath79 is a schreier set of coset representatives of @xmath75 in @xmath80 define the map @xmath81 which takes an element @xmath82 into the representative @xmath83 from @xmath79 .
the element @xmath84 belongs to @xmath75 and , by theorem 2.7 from @xcite , the group @xmath75 is generated by @xmath85 where @xmath86 runs over the set @xmath79 and @xmath87 runs over the set of generators of @xmath28 . + * case 1 . * if @xmath88 , then we find the generators @xmath89 + * case 2 . * if @xmath90 , then we find the generators @xmath91 + * case 3 . * if @xmath92 , then we find the generators @xmath93 + to find defining relations of @xmath75 we define a rewriting process @xmath94 .
it allows us to rewrite a word which is written in the generators of @xmath28 and to present an element in @xmath75 as a word in the generators of @xmath75 .
let us associate to the reduced word @xmath95 where @xmath96 the word @xmath97 in the generators of @xmath75 , where @xmath98 is a representative of the ( @xmath99)th initial segment of the word @xmath100 if @xmath101 and @xmath98 is a representative of the @xmath102th initial segment of the word @xmath100 if @xmath103 . by ( * ?
* theorem 2.9 ) , the group @xmath75 is defined by relations @xmath104 where @xmath105 is a defining relation of @xmath28 .
[ propdn1 ] the group @xmath75 admits a presentation with the generators + @xmath106 @xmath107 and relations : + braid relations ( b1 ) @xmath108 ( b2 ) @xmath109 ( b3 ) @xmath110 ( b4 ) @xmath111 ( b5 ) @xmath112 mixed relations ( r1.1 ) @xmath51 @xmath113 @xmath53 ( r1.2 ) @xmath114 @xmath115 ( r1.3 ) @xmath116 @xmath53 @xmath117 ( r2.1 ) @xmath54 @xmath53 ( r2.2 ) @xmath118 @xmath53 ( r2.3 ) @xmath119 @xmath53 ( r3.1 ) @xmath55 , @xmath56 , @xmath120 ( r3.2 ) @xmath121 , @xmath56 , @xmath57 , ( r3.3 ) @xmath122 @xmath123 @xmath120 ( r4.1 ) @xmath58 @xmath59 ( r4.2 ) @xmath124 @xmath59 ( r4.3 ) @xmath125 @xmath59 ( r5.1 ) @xmath60 @xmath113 @xmath61 ( r5.2 ) @xmath126 @xmath115 @xmath61 ( r5.3 ) @xmath127 @xmath61 @xmath128 ( r6.1 ) @xmath62 @xmath63 @xmath61 ( r6.2 ) @xmath129 @xmath63 @xmath61 ( r6.3 ) @xmath130 @xmath63 @xmath61 ( r7.1 ) @xmath65 @xmath66 ( r7.2 ) @xmath131 @xmath66 ( r7.3 ) @xmath132 @xmath66 ( r8.1 ) @xmath133 @xmath134 ( r8.2 ) @xmath135 @xmath134 ( r8.3 ) @xmath136 @xmath68 the generators @xmath137 generate a group isomorphic to @xmath138 ( see also remark 3.1 from @xcite ) and it is easy to see that the relations ( b1 ) , ( b2 ) , ( r1.1 ) , ( r2.1 ) , , ( r8.1 ) are a complet set of relations for @xmath138 . from the other relations we can find the following conjugacy formulae : + ( s1 ) @xmath139
@xmath140 ( s2 ) @xmath141 ( s3 ) @xmath142
@xmath143 ( s4 ) @xmath144
@xmath115 ( s5 ) @xmath145 @xmath53 @xmath146 ( s6 ) @xmath147 ( s7 ) @xmath148 ( s8 ) @xmath149},$ ] @xmath123 @xmath120 ( s9 )
@xmath150 @xmath123 @xmath120 ( s10 ) @xmath151 w_{2m-1 } \tau_1,$ ] ( s11 ) @xmath152 ( s12 ) @xmath153 @xmath115 @xmath61 ( s13 ) @xmath154 @xmath128 @xmath61 ( s14 ) @xmath155},$ ] @xmath53 @xmath61 ( s15 ) @xmath156
@xmath53 @xmath61 ( s16 ) @xmath157},$ ] @xmath158 ( s17 ) @xmath159
@xmath158 ( s18 ) @xmath160 \tau_1,$ ] @xmath61 ( s19 ) @xmath161
@xmath61 where @xmath162 and @xmath163 = a^{-1 } b^{-1 } a b.$ ] let @xmath164 be the subgroup of @xmath75 generated by @xmath165 @xmath166 @xmath167 @xmath168 . [ un ]
the group @xmath164 is a normal subgroup of @xmath75 and it is a free group of rank @xmath169 .
the statement was proven in ( * ? ? ?
* section.2 ) using the interpretation of @xmath164 as the fundamental group of the surface @xmath34 with @xmath170 points removed . from relations ( s1) ...
(s19 ) we deduce that @xmath138 acts on @xmath164 by conjugacy .
[ theorem3 ] the group @xmath138 with @xmath171 , @xmath35 and @xmath36 acts by conjugacy on the free group @xmath164 .
therefore we have a representation @xmath172 defined algebraically as follows : generators @xmath173 @xmath174 : @xmath175 generators @xmath176 @xmath177 : @xmath178},\ , s < r , \ , ( s , r ) \neq ( 2m-1 , 2 m ) \ , ; & \\ w_{r-1 } \longmapsto [ \tau_1 , w_r^{-1 } ] w_{r-1 } \tau_1 , \ ; \mbox{if } \ ; r = 2 m \ , ; & \\ w_r \longmapsto w_{r}^{\tau_1 } \ , ; & \\ w_s \longmapsto w_{s},\ , r < s,\ , ( r , s ) \neq ( 2m-1 , 2 m ) \ , ; & \\
w_{r+1 } \longmapsto \tau_1^{-1 } w_{r+1 } , \ ; \mbox{if } \ ; r = 2m-1 \ , ; & \\ \xi_j \longmapsto \xi_{j}^{[w_r^{-1 } , \tau_1 ] } , \ , 1 \leq j \leq p-1\ , .
\end{array } \right.\ ] ] generators @xmath179 @xmath180 : @xmath181 z_j : \ { ll _ 1 _ 1^_j^-1 _ 1 ; & + _ i _ i , 2 i n ; & + w_r^z_j w_r,@xmath1821 r 2 g ) ; & + _ l^z_j _ l^[_j^-1 , _ 1 ] , 1 l < j p-1 ; & + _ j^z_j _ j^_1 ; & + _ l^z_j _ l , 1 j < l p-1 . .
@xmath181 in the following we outline a proof of the faithfulness of the representation @xmath183 of @xmath184 given in theorem [ theorem3 ] using the interpretation of surface braids as mapping classes .
first , we recall that the mapping class group of a surface @xmath34 , let us denote it by @xmath185 , is the group of isotopy classes of orientation - preserving self - homeomorphisms which fix the boundary components pointwise .
let @xmath186 be a set of @xmath2 distinct points ( _ punctures _ ) in the interior of the surface @xmath34 .
the punctured mapping class group of @xmath34 relative to @xmath19 is defined to be the group of isotopy classes of orientation - preserving self - homeomorphisms which fix the boundary components pointwise , and which fix @xmath19 setwise .
this group , that we will denote by @xmath187 , does not depend on the choice of @xmath19 , but just on its cardinal .
we recall also that a simple closed curve @xmath188 is _ essential _ if either it does not bound a disk or it bounds a disk containing at least two punctures . finally , we denote @xmath189 as a dehn twist along a simple closed curve @xmath188 .
let @xmath188 and @xmath190 be two simple closed curves bounding an annulus containing only the puncture @xmath191 .
we shall say that the multitwist @xmath192 is a _ @xmath102-bounding pair braid _ , also called _ spin map _ in @xcite . * surface braids as mapping classes .
* let @xmath193 and let @xmath194 be the homomorphism induced by the map which forgets the set @xmath19 .
when @xmath195 , according to a well - known result of birman ( * ? ? ?
* chapter 4.1 ) , the group @xmath196 is isomorphic to @xmath197 if @xmath198 . the statement of birman s theorem concerns the case of closed surfaces , but the proof extends naturally to the case of surfaces with boundary components and the group @xmath28 is isomorphic to @xmath197 if @xmath15 and @xmath36 .
geometrically the correspondence between @xmath197 and @xmath28 is realized as follows : given a homeomorphism @xmath199 of @xmath34 isotopic to the identity @xmath200 of @xmath34 and fixing boundary components pointwise and @xmath19 setwise , the track of the punctures @xmath201 under an isotopy from @xmath199 to @xmath200 is the geometric braid corresponding to the homeomorphism @xmath199 .
[ theorem3annexe ] let @xmath202 , @xmath36 and @xmath171 .
the representation @xmath203 is faithful .
first we prove the claim when @xmath29 is greater or equal then @xmath204 . according to birman s result
we can represent the group @xmath205 as a ( normal ) subgroup of @xmath187 , more precisely as the subgroup of mapping classes in @xmath197 sending the @xmath2th puncture into itself ( see also remark 18 from @xcite ) .
in particular generators @xmath206 of @xmath205 ( see proposition [ propdn1 ] ) correspond to @xmath2-bounding pair braids in @xmath187 .
let @xmath207 be the homomorphism forgetting the last puncture .
now , let us consider the exact sequence associate to the restriction of @xmath208 to @xmath205 . the image of @xmath205 by @xmath208 coincides with @xmath209 which is isomorphic to @xmath138 .
on the other hand , @xmath210 is the subgroup of @xmath187 generated by @xmath2-bounding pair braids @xmath211 and therefore it is isomorphic to @xmath164 . the group @xmath138 embeds naturally in @xmath212 by sending generators of @xmath138 in corresponding ones of @xmath212 and therefore @xmath138 can be considered also as a subgroup of @xmath187 .
moreover , the action considered in theorem [ theorem3 ] corresponds to the action by conjugacy of @xmath138 , seen as a subgroup of @xmath187 , on @xmath164 . since @xmath164 is a normal subgroup of @xmath187 we can define a map @xmath213 .
we prove that @xmath214 is injective and therefore that the action by conjugacy of @xmath138 on @xmath164 is faithful .
let @xmath215 such that @xmath216 and @xmath188 be an essential curve .
we can associate to the curve @xmath188 another simple closed curve @xmath190 such that they bound an annulus containing only the puncture @xmath217 .
the mapping class @xmath218 is then a @xmath2-bounding pair braid . since @xmath216 , then @xmath219 and from a simple argument on the index of intersection of curves
( see for instance proposition 2.10 of @xcite ) one can easily deduce that @xmath220 . since @xmath220 for any essential curve @xmath188 it follows that @xmath29 is isotopic to the identity ( lemma 5.1 and theorem 5.3 of @xcite ) .
therefore @xmath214 is injective and in particular the representation defined in theorem [ theorem3 ] is faithful when @xmath221 . the only case left is when the genus is equal to zero .
we recall that , for @xmath36 , the group @xmath222 is isomorphic to the subgroup @xmath223 of @xmath224 fixing the last @xmath225 strands . in this case
our representation coincides with artin representation of @xmath224 in @xmath226 restricted to the subgroup @xmath223 .
the group @xmath227 is residually finite for @xmath228 , @xmath35 and @xmath229 .
baumslag and smirnov proved ( see for instance ( * ? ? ?
* theorem 4.8 ) ) that any finitely presented group which is isomorphic to a subgroup of automorphisms of a free group of finite rank is residually finite .
therefore for @xmath230 the statement is a corollary of theorem [ theorem3annexe ] .
in the case @xmath231 , the group @xmath232 is isomorphic to the fundamental group of @xmath34 which is free and therefore residually finite .
remark also that from theorem [ theorem3annexe ] one can derive , using fox derivatives , a burau representation for @xmath28 .
since the restriction on @xmath5 of such representation coincides with the usual burau representation of @xmath5 , the burau representation for @xmath28 obtained from theorem [ theorem3annexe ] is not faithful for @xmath233 .
as recalled in the introduction , any element @xmath234 of @xmath235 fixes the product @xmath236 of generators of @xmath1 .
we prove a similar statement for the group @xmath28 , for @xmath229 .
[ prop1 ] let @xmath36 and let @xmath164 be the free group of rank @xmath169 defined above .
any element @xmath234 in @xmath237 fixes the product @xmath238 \ldots [ w_{2 g - 1}^{-1 } , w_{2g}].\ ] ] in order to prove the claim it suffices to verify that any generator of @xmath239 , considered as an automorphism of @xmath164 , fixes the element @xmath240 .
+ _ case 1 : _ generators @xmath6 , @xmath241 .
the group @xmath242 is a subgroup of @xmath138 and by artin s theorem @xmath242 is a subgroup of @xmath243 @xmath244 and fixes the product @xmath245 .
hence , the generator @xmath6 also fixes the product @xmath246 . from theorem [ theorem3 ]
it follows that @xmath247 hence , @xmath248 + _ case 2 : _ generators @xmath249 , @xmath250 . by theorem [ theorem3 ]
we have @xmath251 and @xmath252}.\ ] ] let us consider @xmath253 \ldots [ w_{2 g - 1}^{-1 } , w_{2g}])^{x_r}.\ ] ] we will distinguish two cases : @xmath254 and @xmath255 where @xmath256 in the first case we have the following equalities : @xmath253 \ldots [ w_{2 m - 3}^{-1 } , w_{2m-2}])^{x_{2m-1 } } = ( [ w_1^{-1 } , w_2 ] \ldots [ w_{2 m - 3}^{-1 } , w_{2m-2}])^{[w_{2m-1}^{-1 } , \tau_1]},\ ] ] @xmath257^{x_{2m-1 } } = [ w_{2m-1}^{-\tau_1 } , \tau_1^{-1 } w_{2 m } ] = [ \tau_1 , w_{2m-1}^{-1 } ] [ w_{2 m - 1}^{-1 } , w_{2m}],\ ] ] @xmath258 \ldots [ w_{2 g - 1}^{-1 } , w_{2g}])^{x_{2m-1 } } = [ w_{2m+1}^{-1 } , w_{2m+2 } ] \ldots [ w_{2 g - 1}^{-1 } , w_{2g}].\ ] ] in the second case we have the following equalities : @xmath253 \ldots [ w_{2 m - 3}^{-1 } , w_{2m-2}])^{x_{2 m } } = ( [ w_1^{-1 } , w_2 ] \ldots [ w_{2 m - 3}^{-1 } , w_{2m-2}])^{[w_{2m}^{-1 } , \tau_1]},\ ] ] @xmath257^{x_{2 m } } = [ \tau^{-1}_1 w_{2m-1}^{-1 } [ w_{2m}^{-1 } , \tau_1 ] , w_{2m}^{\tau_1 } ] = [ \tau_1 , w_{2m}^{-1 } ] [ w_{2 m - 1}^{-1 } , w_{2m}],\ ] ] @xmath258 \ldots [ w_{2 g - 1}^{-1 } , w_{2g}])^{x_{2 m } } = [ w_{2m+1}^{-1 } , w_{2m+2 } ] \ldots [ w_{2 g - 1}^{-1 } , w_{2g}].\ ] ] in both cases we have @xmath253 \ldots [ w_{2 g - 1}^{-1 } , w_{2g}])^{x_r } = [ \tau_1 , w_r^{-1 } ] [ w_1^{-1 } , w_2 ] \ldots [ w_{2 g - 1}^{-1 } , w_{2g}].\ ] ] hence , the element @xmath249 acts by conjugation on @xmath240 as follows @xmath259 ( \xi_1 \ldots \xi_{p-1})^{[w_r^{-1 } , \tau_1 ] } [ \tau_1 , w_r^{-1 } ] [ w_1^{-1 } , w_2 ] \ldots\ ] ] @xmath260 , $ ] and then it is easy to check that @xmath261 + _ case 3 : _ generators @xmath262 , @xmath263 . by theorem [ theorem3 ]
we have @xmath264,\ ] ] @xmath265 } \xi_j^{\tau_1 } \xi_{j+1 } \ldots \xi_{p-1 } = [ \xi_j^{-1 } , \tau_1]^{-1 } \xi_{1 } \ldots \xi_{p-1},\ ] ] and @xmath253 \ldots [ w_{2 g - 1}^{-1 } , w_{2g}])^{z_j } = [ w_1^{-1 } , w_2 ] \ldots [ w_{2 g - 1}^{-1 } , w_{2g}].\ ] ] hence , @xmath266 and the proposition follows .
we recall that @xmath267 is the representation of @xmath138 defined in theorem [ theorem3 ] and let @xmath268 be the canonical projection . [ out ] the representation @xmath269 is faithful for @xmath36 when @xmath270 and for @xmath271 when @xmath272 .
since the representation @xmath267 is faithful ( theorem [ theorem3annexe ] ) , we can identify @xmath138 with @xmath273 .
now suppose that there exists @xmath274 . from proposition [ prop1 ]
one deduces that @xmath234 is a conjugation by a power @xmath275 of @xmath238 \ldots [ w_{2 g - 1}^{-1 } , w_{2g}].\ ] ] now , let @xmath29 be a generator of @xmath138 . since all elements of @xmath138 fix the element @xmath240 we deduce the following equalities : @xmath276 for any @xmath277 .
one deduces that @xmath278 for any generator @xmath29 of @xmath138 and therefore @xmath234 belongs to the center of @xmath138 . since @xmath138 ( with @xmath37 ) has trivial center for @xmath36 when @xmath270 @xcite and for @xmath271 when @xmath272 , the intersection @xmath279 is trivial and the claim follows .
[ outrem ] let @xmath280 the representation obtained composing artin representation of @xmath5 in @xmath281 with the canonical projection of @xmath281 in @xmath282 .
such representation is not faithful and it is easy to see that the kernel is the center of @xmath5 .
let @xmath283 be a non - orientable surface of genus @xmath284 with @xmath36 boundary components .
let @xmath285 be the usual generators of @xmath5 and @xmath286 be the usual generators of the fundamental group of @xmath283 .
as in previous section we can consider @xmath285 and @xmath286 as elements of @xmath287 . according to @xcite the group @xmath287 admits a presentation with generators : @xmath288 and relations : braid relations : @xmath49 @xmath50 mixed relations : ( r1 ) @xmath289 @xmath52 @xmath290 ( r2 ) @xmath291 @xmath290 ( r3 ) @xmath292 , @xmath293 ( r4 ) @xmath60 @xmath52 @xmath61 ( r5 ) @xmath294 , @xmath295 , @xmath296 , @xmath297 ( r6 ) @xmath65 @xmath66 ( r7 ) @xmath298 , @xmath68 + as in previous section
let us consider the natural projection of @xmath299 , which associates to any braid the corresponding permutation .
this projection map @xmath6 in the corresponding transposition and generators @xmath300 into the identity . as before , let @xmath301 and let @xmath302 @xmath303 , @xmath304 .
the set @xmath76 is a schreier set of coset representatives of @xmath305 in @xmath32 and the group @xmath305 is generated by @xmath85 where @xmath86 runs over the set @xmath79 and @xmath87 runs over the set of generators of @xmath32 . + * case 1 . * if @xmath88 , then we find the generators @xmath89 + * case 2 . * if @xmath306 , then we find the generators @xmath307 + * case 3 . *
if @xmath92 , then we find the generators @xmath93 using the same argument as in the orientable case one can find the following group presentation for @xmath308 . [ theorem4 ] the group @xmath309 @xmath310 admits a presentation with the generators @xmath311 and the following relations : braid relations : + ( b1 ) @xmath108 ( b2 ) @xmath109 ( b3 ) @xmath110 ( b4 ) @xmath111 ( b5 ) @xmath112 mixed relations : + ( r1.1 ) @xmath289 @xmath113 @xmath290 ( r1.2 ) @xmath312 @xmath115 ( r1.3 ) @xmath116 @xmath290 @xmath128 ( r2.1 ) @xmath313 @xmath290 ( r2.2 ) @xmath314 @xmath290 ( r2.3 ) @xmath315 @xmath290 ( r3.1 ) @xmath316 @xmath293 ( r3.2 ) @xmath317 @xmath293 ( r3.3 ) @xmath318 @xmath293 ( r4.1 ) @xmath60 @xmath113 @xmath61 ( r4.2 ) @xmath126 @xmath115 @xmath61 ( r4.3 ) @xmath127 @xmath61 @xmath128 ( r5.1 ) @xmath319 @xmath320 @xmath61 ( r5.2 ) @xmath321 @xmath320 @xmath61 ( r5.3 ) @xmath130 @xmath320 @xmath61 ( r6.1 ) @xmath65 @xmath66 ( r6.2 ) @xmath131 @xmath66 ( r6.3 ) @xmath132 @xmath66 ( r7.1 ) @xmath133 @xmath134 ( r7.2 ) @xmath135 @xmath134 ( r7.3 ) @xmath136 @xmath68 it is easy to see that the relations ( b1 ) , ( b2 ) , ( r1.1 ) , ( r2.1 ) , , ( r7.1 ) are defining relations of @xmath322 . from the other relations we can find the following conjugacy formulae : ( s1 ) @xmath139
@xmath140 ( s2 ) @xmath141 ( s3 ) @xmath142
@xmath143 ( s4 ) @xmath323 @xmath115 ( s5 ) @xmath145 @xmath290 @xmath146 ( s6 ) @xmath324 ( s7 ) @xmath325 @xmath290 ( s8 ) @xmath326 @xmath293 ( s9 ) @xmath327
@xmath293 ( s10 ) @xmath153 @xmath115 @xmath61 ( s11 ) @xmath154
@xmath128 @xmath61 ( s12 ) @xmath328 @xmath290 @xmath61 ( s13 ) @xmath156 @xmath290 @xmath61 ( s14 ) @xmath157},$ ] @xmath158 ( s15 ) @xmath159
@xmath158 ( s16 ) @xmath160 \tau_1,$ ] @xmath61 ( s17 ) @xmath161
@xmath329 + consider the subgroup @xmath330 of @xmath331 .
the group @xmath332 is free ( see @xcite ) and from ( s1)(s17 ) we see that @xmath333 acts on @xmath332 by conjugacy .
[ theorem5a ] the group @xmath287 , for @xmath334 , @xmath335 and @xmath221 acts by conjugation on the free group @xmath332 and the induced representation @xmath336 is defined as follows : generators @xmath173 @xmath337 : @xmath181 _ i : \ { ll _ i _ i+1^_i^-1 ; & + _ i+1 _ i ; & + _ l _ l,@xmath182l i , i+1 ; & + w_r w_r , 1 r g ; & + _ j _
j,@xmath1821 j p-1 . .
@xmath181 generators @xmath338 @xmath339 : @xmath181 a_r : \
{ ll _ 1 ( _ 1 ^ -1)^w_r^-1 _ 1 ; & + _ i _ i , 2 i n ; & + w_s w_s^w_r _ 1 w_r^-1 _ 1,@xmath1821 s < r g ; & + w_r _ 1 ^ -1 w_r ; & + w_s w_s,@xmath182 1 r
< s g ; & + _ j _ j^w_r _ 1 w_r^-1 _
1,@xmath1821 j p-1 . .
@xmath181 generators @xmath179 @xmath180 : @xmath181 z_j : \ { ll _ 1 _ 1 ; & + _ i _
i,@xmath1822 i n ; & + w_r w_r , 1 r g ; & + _ l _ l^[_j^-1 , _ 1 ] , 1 l < j p-1 ; & + _ j _ j^_1 ; & + _ l _ l , 1 j < l p-1 . .
@xmath181 we do nt know if the representation @xmath336 is faithful or not .
[ prop2 ] let @xmath234 be in @xmath287 .
the element @xmath340 in @xmath341 fixes the product @xmath342 as above it is enough to prove that images of generators of @xmath287 fix @xmath240 .
_ case 1 : _ generators @xmath173 @xmath343 .
the group @xmath5 is a subgroup of @xmath32 and by artin s theorem @xmath5 is a subgroup of @xmath4 , @xmath344 and fixes the product @xmath245 .
hence , @xmath345 also fixes the product @xmath246 . from theorem [ theorem5a ]
it follows that @xmath346 hence , @xmath248 + _ case 2 : _ generators @xmath338 @xmath347 . by theorem [ theorem5a ]
we have that @xmath348 and @xmath349 and @xmath350 hence , the element @xmath351 acts on @xmath240 as follows : @xmath352 and @xmath353 _ case 3 .
_ generators @xmath179 @xmath263 . by theorem [ theorem5a ]
we deduce that @xmath354 @xmath355 and @xmath356 hence , @xmath266 .
let @xmath357 be a closed orientable surface of genus @xmath358 .
the group @xmath359 @xmath334 admits a group presentation with generators : @xmath360 and relations : braid relations : @xmath49 @xmath50 mixed relations : ( r1 ) @xmath51 @xmath52 @xmath53 ( r2 ) @xmath54 @xmath53 ( r3 ) @xmath361 @xmath123 @xmath120 ( r4 ) @xmath58 @xmath59 ( tr5 ) @xmath362 [ x_3^{-1 } , x_4 ] \ldots [ x_{2g-1}^{-1 } , x_{2 g } ] = \sigma_1 \sigma_2 \ldots \sigma_{n-1}^2 \ldots \sigma_2 \sigma_1.$ ] + as before , let @xmath363 be the natural projection of @xmath196 in @xmath12 , let @xmath364 and @xmath365 @xmath366 @xmath367 .
the set @xmath76 is a schreier set of coset representatives of @xmath368 in @xmath196 and @xmath368 is generated by @xmath369 where @xmath370 @xmath91 [ theorem6 ] the group @xmath371 @xmath372 admits a group presentation with generators : @xmath373 and relations : braid relations : ( b1 ) @xmath108 ( b2 ) @xmath109 ( b3 ) @xmath110 ( b4 ) @xmath111 ( b5 ) @xmath112 mixed relations : ( r1.1 ) @xmath51 @xmath113 @xmath53 ( r1.2 ) @xmath114 @xmath115 ( r1.3 ) @xmath116 @xmath53 @xmath128 ( r2.1 ) @xmath54 @xmath53 ( r2.2 ) @xmath118 @xmath53 ( r2.3 ) @xmath119 @xmath53 ( r3.1 ) @xmath361 @xmath123 @xmath120 ( r3.2 ) @xmath374 @xmath123 @xmath120 ( r3.3 ) @xmath122 @xmath123 @xmath120 ( r4.1 ) @xmath58 @xmath59 ( r4.2 ) @xmath124 @xmath59 ( r4.3 ) @xmath125 @xmath59 ( rt.1 ) @xmath362 [ x_3^{-1 } , x_4 ] \ldots [ x_{2g-1}^{-1 } , x_{2 g } ] = \sigma_1 \sigma_2 \ldots \sigma_{n-3 } \sigma_{n-2}^{2 } \sigma_{n-3 } \ldots \sigma_{2 } \sigma_1 \tau_1,$ ] ( rt.2 ) @xmath375 [ w_3^{-1 } , w_4 ] \ldots [ w_{2g-1}^{-1 } , w_{2 g } ] = \tau_1 \tau_2 \ldots \tau_{n-1}.$ ] consider the subgroup @xmath376 of @xmath377 generated by @xmath378 @xmath379 .
the group @xmath376 is free and normal in @xmath377 @xcite . using relation ( rt.2 )
one finds that @xmath380 [ w_3^{-1 } , w_4 ] \ldots [ w_{2g-1}^{-1 } , w_{2 g } ] \tau_{n-1}^{-1 } \ldots \tau_{2}^{-1}\ ] ] and therefore @xmath381 . also , consider the subgroup @xmath382 of @xmath368 .
the group @xmath383 is not isomorphic to @xmath384 .
in fact , one can prove that the relations ( b1 ) , ( b2 ) , ( r1.1 ) , ( r2.1 ) , ( r3.1 ) , ( r4.1 ) and ( rt.1 ) are defining relations of @xmath385 and therefore we can deduce that @xmath386 . from the other relations we can see that @xmath383 acts on @xmath387 by conjugacy .
hence , the following theorem holds .
[ theorem6b ] there exists a representation @xmath388 , induced by the action by conjugacy of @xmath383 on the free group @xmath389 @xmath390 given algebraically as follows : + action on generators @xmath391 : ( s1 ) @xmath139 @xmath140 + ( s2 ) @xmath141 + ( s3 ) @xmath142 @xmath143 + ( s4 ) @xmath144 @xmath115 @xmath53 + action on generators @xmath392 : ( s5 ) @xmath145 @xmath53 @xmath146 + ( s6 ) @xmath147 @xmath53 + ( s7 ) @xmath148 @xmath53 + ( s8 ) @xmath149},$ ] @xmath123 @xmath120 + ( s9 ) @xmath150 @xmath123 @xmath120 + ( s10 ) @xmath151 w_{2m-1 } \tau_1,$ ] + ( s11 ) @xmath152 + where @xmath380 [ w_3^{-1 } , w_4 ] \ldots [ w_{2g-1}^{-1 } , w_{2 g } ] \tau_{n-1}^{-1 } \ldots \tau_{2}^{-1}.\ ] ] from proposition [ theorem6b ] we get a representation of @xmath393 in @xmath394 , where @xmath395 is the conjugation by the word @xmath396 \ldots [ w^{-1}_{2g-1 } , w_{2g}]\tau_{n-1}^{-1 } \ldots \tau_2^{-1}.\ ] ] therefore , we have the following result : [ theorem7 ] there exists a representation @xmath397 , given algebraically as follows : + action ( up to conjugacy ) on generators @xmath391 : ( s1 ) @xmath139 @xmath140 + ( s2 ) @xmath141 + ( s3 ) @xmath142 @xmath143 + ( s4 ) @xmath144 @xmath115 @xmath53 + action ( up to conjugacy ) on generators @xmath392 : ( s5 ) @xmath145 @xmath53 @xmath146 + ( s6 ) @xmath147 @xmath53 + ( s7 ) @xmath148 @xmath53 + ( s8 ) @xmath149},$ ] @xmath123 @xmath120 + ( s9 ) @xmath150
@xmath123 @xmath120 + ( s10 ) @xmath151 w_{2m-1 } \tau_1,$ ] + ( s11 ) @xmath152 + where @xmath380 [ w_3^{-1 } , w_4 ] \ldots [ w_{2g-1}^{-1 } , w_{2 g } ] \tau_{n-1}^{-1 } \ldots \tau_{2}^{-1}.\ ] ]
we do nt know if the representation @xmath398 of @xmath384 is faithful or not .
we recall that classical braid groups are also called artin - tits groups of type @xmath399 .
more precisely , let @xmath400 be a coxeter system and let us denote the order of the element @xmath401 in @xmath402 by @xmath403 ( for @xmath404 ) .
let @xmath405 be the group defined by the following group presentation : @xmath406 the group @xmath405 is the artin - tits group associated to @xmath402 .
the group @xmath405 is said to be of spherical type if @xmath402 is finite .
there exists three infinite families of finite coxeter groups usually denoted respectively as of type @xmath399 , @xmath407 and @xmath0 .
artin braid groups correspond to the family of artin - tits groups associated to coxeter groups of type @xmath399 and braid groups of the annulus coincide with artin - tits groups associated to coxeter groups of type @xmath407 . in this section
we give a faithful representation of the remaining infinite family of spherical artin - tits groups , associated to coxeter groups of type @xmath0 .
first let us denote by @xmath408 the @xmath2-th artin - tits group of type @xmath0 with the group presentation provided by its coxeter graph ( see figure 1 ) , where vertices @xmath409 are the generators of the group and two generators @xmath410 verify the relation @xmath411 if they are related by an edge and elsewhere they commute . 1 .
the representation @xmath419 given algebraically by : + @xmath420 + and for @xmath421 , + @xmath422 + is well defined and faithful .
the group @xmath408 is isomorphic to @xmath423 , where the projection on the second factor is @xmath424 and the section @xmath425 is just @xmath426 .
in particular , @xmath427 is freely generated by @xmath428 , where @xmath429 and @xmath430 for @xmath431 . from proposition [ artind ]
it follows that @xmath423 is isomorphic to @xmath408 through the morphism @xmath437 defined algebraically as follows : @xmath438 for any generator @xmath439 of @xmath440 and @xmath441 for any generator @xmath6 of @xmath442 .
remark that @xmath423 is a subgroup of @xmath443 .
now , considering the natural inclusion of @xmath14 into @xmath16 leaving the generator @xmath444 invariant and the action by conjugacy of @xmath440 on @xmath1 , we obtain a morphism @xmath445 , which can be easily proved to be injective . by direct calculation
one can verify that @xmath446 is the conjugacy by @xmath439 for @xmath447 ; moreover , since @xmath448 restricted to @xmath5 coincides with the representation of @xmath5 into @xmath16 obtained composing the map @xmath449 with the natural inclusion of @xmath14 into @xmath16 leaving the generator @xmath444 invariant , we can deduce that @xmath450 and therefore the claim follows . | we define and study extensions of artin s representation and braid monodromy representation to the case of topological and algebraical generalisations of braid groups .
in particular we provide faithful representations of braid groups of oriented surfaces with boundary components as ( outer ) automorphisms of free groups .
we give also similar representations for braid groups of non oriented surfaces with boundary components and we show a representation of braid groups of closed surfaces as outer automorphisms of free groups .
finally , we provide faithful representations of artin - tits groups of type @xmath0 as automorphisms of free groups . |
the problem of index coding with side information was introduced by birk and kol @xcite and bar - yossef _ et al . _
@xcite studied the class of index coding problems in which each receiver demands only one single message and the number of receivers equals number of messages . ong and ho @xcite classify the binary index coding problem depending on the demands and the side information possessed by the receivers .
an index coding problem is unicast if the demand sets of the receivers are disjoint .
if the problem is unicast and if the size of each demand set is one , then it is said to be single unicast .
any unicast index problem can be equivalently reduced to an single unicast problem discussed in @xcite . for this canonical unicast index coding problem
, it was shown that the length of the optimal linear index code is equal to the minrank of the side information graph of the index coding problem but finding the minrank is np hard .
+ maleki _ et al .
_ @xcite found the capacity of symmetric multiple unicast index problem with neighboring antidotes ( side information ) . in a symmetric multiple unicast index coding problem with equal number of @xmath0 messages and source - destination pairs , each destination has a total of @xmath6 antidotes , corresponding to the @xmath7 messages before ( up " from ) and @xmath8 messages after ( down " from ) its desired message . in this
setting , the @xmath9 receiver @xmath10 demands the message @xmath11 having the antidotes @xmath12 the symmetric capacity @xmath13 of this index coding problem setting is shown to be as follows : @xmath14 @xmath15 + @xmath16 , + @xmath6 , + @xmath17per
message.@xmath18 + the above expression for capacity per message can be equivalently expressed as : @xmath19 in the setting of @xcite with one sided antidote cases , i.e. , the cases where @xmath7 or @xmath8 is zero , without loss of generality , we can assume that @xmath20 and @xmath21 ( all the results hold when @xmath22 ) . in this setting , the @xmath9 receiver @xmath10 demands the message @xmath11 having the antidotes , @xmath23 for which reduces to @xmath24 symbols per message .
+ in the capacity expression given in if @xmath25 divides @xmath26 , capacity can be achieved by using scalar linear codes . in the scalar linear coding one packs @xmath0 messages in @xmath27 dimensions ( code symbols ) .
if @xmath25 does not divide @xmath26 , vector linear coding can only achieve capacity . in the vector linear code
, one needs to pack @xmath28 message symbols corresponding to @xmath0 users in @xmath29 dimensions .
+ in @xcite maleki @xmath30 proved that vector linear coding exists for any arbitrary @xmath7 and @xmath8 over a sufficiently large field size by imposing conditions on encoding and decoding matrices @xmath31 and @xmath32 . in section
ii we prove that vector linear solution exists for a two - sided symmetric antidote problem with @xmath7 antidotes above and @xmath8 antidotes below if a scalar linear solution exists for one - sided antidote problem with same number of messages and number of one - sided antidotes @xmath33 .
we give a construction procedure which constructs a sequence of multiple unicast index problems with two - sided antidotes starting from a given multiple unicast index coding problem of one - sided antidote .
it is shown that if there is an optimal scalar linear index code for the starting problem then this code can be used to construct an optimal vector linear index code for all the extended problems . in @xcite
the authors proposed optimal scalar linear index codes for ten classes of one sided ( not necessarily adjacent ) symmetric multiple unicast index coding problems with optimal scalar linear index codes .
the antidotes assumed in this work is a proper subset of for eight classes .
these optimal codes continue to be optimal if the antidotes are taken to be adjacent as given in .
we illustrate our construction to some of the ten cases of the symmetric multiple unicast problems studied in @xcite and demonstrate the new classes of symmetric multicast problems created for all of which an optimal vector linear code is exhibited .
+ in @xcite , we proposed a lifting construction which constructs a sequence of multiple unicast index problems with one - sided antidotes with a scalar linear index code starting from a given multiple unicast index coding problem with a known scalar linear index code .
the construction in this paper is different in the following two respects : 1 .
the construction in @xcite starts from a problem with @xmath0 messages and gives a sequence of problems with @xmath34 number of messages for @xmath35 i.e. , the number of messages goes on increasing as the index @xmath2 moves .
whereas in this work the number of messages remains @xmath0 and only the size of the antidote sets increase .
2 . in @xcite new scalar linear index codes
are obtained starting from a scalar linear index code for problems of different source sizes whereas in this paper new vector linear index codes are obtained starting from a scalar linear index code for the problem with the same source size .
+ 3 . the lifting construction in @xcite results in index coding problems with one sided antidotes where as the construction in this paper results in index coding problems with two sided antidotes . throughout the paper wlog we consider the case @xmath36 .
all the codes discussed in this paper also applicable for @xmath37 .
the decoding procedure in this paper is considered for the binary field .
however the index codes considered in this paper works for any finite field .
let the messages symbols be @xmath38 for every @xmath39 when we deal with vector linear index codes the different messages symbols corresponding to @xmath40 are denoted by @xmath41 etc .
if @xmath42 do not divide @xmath29 , scalar linear codes can not achieve capacity . in this case capacity
can be achieved by using vector linear coding . in the vector linear code , we require to pack @xmath28 message symbols corresponding to @xmath0 users in @xmath43 dimensions .
we prove that we can pack @xmath28 message symbols into @xmath0 symbols @xmath44 and then convert these @xmath0 symbols into @xmath45 code symbols by using one side adjacent antidote scalar linear code .
define the symbol @xmath46 for @xmath47 as @xmath48 the symbol @xmath46 comprises of @xmath49 message symbols and each of the @xmath28 message symbols appear exactly once in one of the @xmath46 . in the symbol @xmath46
, there exists @xmath49 message symbols and these @xmath49 message symbols are required by the receivers @xmath50 , for @xmath51 + [ fig : ifgraphexampletheorem ] [ thm1 ] for a multiple unicast index coding problem with @xmath0 messages @xmath52 and the same number of receivers with the receiver @xmath53 wanting the message @xmath54 and having a symmetric antidote pattern @xmath55 given by @xmath56 let @xmath57 be a scalar linear code of length @xmath58 define @xmath59 as @xmath9 vector message symbol for @xmath47 . for an arbitrary positive integer @xmath7 consider the index coding problem with @xmath0 number of messages @xmath60 and the number of receivers being @xmath61 and receiver @xmath53 ( @xmath62 ) having antidote pattern given by @xmath63 for this index coding problem the code @xmath64 is obtained by replacing every message symbol @xmath54 in the code symbols of @xmath1 with @xmath65 for @xmath66 i.e. , by making the substitution @xmath67 the given scalar linear code @xmath57 of length @xmath68 for a multiple unicast index coding problem with @xmath0 messages @xmath52 and the same number of receivers with the receiver @xmath53 wanting the message @xmath54 and having a symmetric antidote pattern @xmath55 given in enables the decoding of @xmath0 messages @xmath52 .
that is , with linear decoding there exist combination of code symbols @xmath69 to get the sum of the form given in @xmath70 for @xmath71 such that @xmath72 are in antidotes of receiver @xmath10 .
since @xmath73 are less than or equal to @xmath8 for @xmath47 from the sum @xmath74 , receiver @xmath10 can decode its wanted message @xmath46 .
given an arbitrary positive integer @xmath7 consider the index coding problem with @xmath0 number of vector message symbols @xmath60 and the number of receivers being @xmath61 and the receiver @xmath53 ( @xmath62 ) having antidote pattern given by .
the code @xmath64 is the index code obtained by making the substitution @xmath75 in the available code @xmath1 .
we prove that the receiver @xmath10 can decode its wanted message set @xmath76 ( @xmath42 message symbols @xmath77 ) by using the code @xmath64 . for the code @xmath64
, the sum in can be written by + @xmath78 we describe the entire decoding process in the following three steps .
the fig.1 will be helpful to trace these steps .
+ _ step 1 .
decoding of @xmath79 message symbol by receiver @xmath10 _ + + the sum @xmath80 similar to can be written as @xmath81 in @xmath80 , only one message symbol is present which is required by @xmath82 we have @xmath83 and thus all other message symbols present in @xmath80 are in antidotes of receiver @xmath10 according to antidote pattern as in .
( note that if @xmath84 comprise @xmath85 message symbols which belong to @xmath86 and @xmath10 requires at least two of them , then messages interfere and @xmath10 can not decode any of them ) .
thus by using @xmath80 , receiver @xmath10 can decode its @xmath79 message symbol @xmath87 .
the code @xmath1 enables the decoding of @xmath44 from the sums @xmath88 , which implies that code @xmath64 enables the the decoding of @xmath89 + + _ step 2 .
decoding of @xmath90 message symbol by receiver @xmath10 _
+ + receiver @xmath10 uses sum @xmath91 to decode its @xmath90 message symbol @xmath92 .
the sum @xmath91 similar to can be written as + @xmath93 + where @xmath94 if @xmath95 , else @xmath96 . that is , depending on the value of @xmath97 , sum @xmath91
comprises of either one message symbol or two message symbols belongs to vector message symbol @xmath86 .
+ if @xmath98 , @xmath91 comprises only one message symbol ( @xmath92 ) which belongs to @xmath99 .
all other messages present in @xmath91 are antidotes to @xmath10 .
thus receiver @xmath10 can decode message symbol @xmath92 .
+ if @xmath95 , @xmath91 comprises two message symbols ( @xmath100 ) belongs to vector message symbol @xmath101 receiver @xmath10 has already decoded the message symbol @xmath87 and all other messages present in @xmath91 are antidotes to @xmath10 .
thus receiver @xmath10 can decode message symbol @xmath92 .
+ + _ step l. decoding of @xmath102 message symbol by receiver @xmath10 _
+ + receiver @xmath10 uses sum @xmath103 to decode its @xmath104 message symbol @xmath105 .
the sum @xmath103 similar to can be written as + @xmath106 where @xmath107 is the indicator function such that @xmath108 if @xmath109 , else it is zero .
@xmath110 if @xmath111 , else @xmath112 for @xmath113 . in the above sum @xmath105
is the required message . + the quantity @xmath114 for @xmath113 is the interference to the receiver @xmath10 from its wanted vector message symbol @xmath86 .
receiver @xmath10 already knows the @xmath115 message symbols @xmath116 and thus the interference from the wanted message symbol @xmath86 in the sum @xmath103 can be canceled .
+ the message symbols in the terms @xmath117 for @xmath113 is in antidotes for receiver @xmath10 .
thus receiver @xmath10 can decode @xmath105 from @xmath103 .
+ this procedure continued until all receiver @xmath10 decode its @xmath42 wanted messages @xmath118 .
thus receiver @xmath10 decodes its wanted vector message symbol @xmath86 .
this completes the proof for decoding .
[ thm2 ] in the construction of theorem [ thm1 ] , if the given code @xmath1 has optimal length @xmath119 then all extended vector linear codes of given code are of optimal length and hence capacity achieving .
the number of code symbols in the extended code is equal to the number of code symbols in @xmath1 which is equal to @xmath45 = @xmath29 . by using @xmath29 code symbols
, every receiver gets @xmath42 of its wanted messages .
the capacity achieved by this code is @xmath120 per message , which is equal to the capacity mentioned in .
it follows that the extended vector linear codes are of optimal length .
the following examples illustrates theorem [ thm1 ] and theorem [ thm2 ] @xmath121 + + we have one - sided antidote scalar code @xmath1=@xmath122 .
+ + @xmath123 : @xmath124 + + @xmath125 , capacity=@xmath126 .
define @xmath127 for @xmath128 we have + + @xmath129 , + @xmath130 , + @xmath131 , + @xmath132 , + @xmath133 , + @xmath134 , + @xmath135 , + @xmath136 , + @xmath137 , + @xmath138 .
+ we get the code @xmath139 given by + @xmath140 .
+ + @xmath141 : @xmath142 + + @xmath125 , capacity=@xmath143 . defining @xmath144 for @xmath145
we get + + @xmath146 , + @xmath147 , + @xmath148 , + @xmath149 , + @xmath150 , + @xmath151 , + @xmath152 , + @xmath153 , + @xmath154 , + @xmath155 . + for this case we get the extended code @xmath156 to be + @xmath157 .
+ + @xmath158 : @xmath159 + + @xmath125 , capacity=@xmath143 . with defining @xmath160 for @xmath161 we end with + + @xmath162 , + @xmath163 , + @xmath164 , + @xmath165 , + @xmath166 , + @xmath167 , + @xmath168 , + @xmath169 , + @xmath170 , + @xmath171 .
+ the resulting extended vector linear code is + @xmath172 .
[ thm3 ] consider a multiple unicast index coding problem with @xmath0 messages and the same number of receivers with the receiver @xmath53 wanting the message @xmath40 and having a symmetric antidote pattern @xmath55 given in .
for this index coding problem vector linear solution exists if scalar linear solution exists for one - sided antidote problem as mentioned in with same number of messages and number of one - sided antidotes @xmath33 . the vector linear index code is optimal if the scalar linear code is optimal .
+ proof follows from theorem [ thm1 ] and theorem [ thm2 ] .
[ thm4 ] proposed codes in @xcite can be extended for all @xmath7 and @xmath8 with @xmath173 = @xmath174 for the following problem instances : 1 .
@xmath173 divides @xmath0 2 .
@xmath45 divides @xmath0 3 .
@xmath175 divides @xmath173 4 .
@xmath176 divides @xmath177 5 .
@xmath173 divides @xmath178 and @xmath179 divides @xmath173 where @xmath179 is an integer 6 .
@xmath45 divides @xmath178 and @xmath179 divides @xmath45 7 .
@xmath180 divides @xmath0 and @xmath179 divides @xmath173 8 .
@xmath181 divides @xmath0 and @xmath179 divides @xmath45 9 .
@xmath173 divides @xmath182 and @xmath179 divides @xmath173 10 .
@xmath45 divides @xmath182 and @xmath179 divides @xmath45 in @xcite codes for symmetric instances of one side antidote problem for a given @xmath0 and @xmath8 satisfying the above mentioned conditions with @xmath173 replaced with @xmath8 were proposed .
the proposed codes in @xcite can also be used for the one - sided adjacent antidotes as given in and these codes achieve the capacity in .
then the proof follows from theorem [ thm1 ] .
_ corollary 1 .
_ if @xmath8 divides @xmath0 then the proposed optimal length scalar linear code is @xmath183 @xmath142 + + @xmath184 , capacity=@xmath143 .
+ let @xmath144 for @xmath128 + + @xmath146 , + @xmath147 , + @xmath148 , + @xmath149 , + @xmath150 , + @xmath151 , + @xmath152 , + @xmath153 , + @xmath154 , + @xmath155 .
+ the proposed code is @xmath185 .
@xmath186 .
_ corollary 2 .
_ if divides @xmath0 , then the proposed code is @xmath187 where @xmath188 and .
+ @xmath189 .
+ + @xmath190 , capacity=@xmath191 .
+ let @xmath127 for @xmath128 + + @xmath129 , + @xmath130 , + @xmath131 , + @xmath132 , + @xmath133 , + @xmath134 , + @xmath135 , + @xmath136 , + @xmath137 , + @xmath138 .
+ the proposed code is , + @xmath192 + + @xmath193 + _ corollary 3 .
_ for the case @xmath194 divides @xmath8 , the proposed scalar linear code is , + @xmath195 where @xmath196 , @xmath197 and @xmath198 .
+ @xmath199 .
+ + @xmath200 , capacity=@xmath201 .
+ let @xmath127 for @xmath128 + + @xmath202 , + @xmath203 + @xmath204 + @xmath205 + @xmath206 , + @xmath207 , + @xmath208 .
+ + the proposed code is @xmath209 + @xmath210 .
+ @xmath211 _ corollary 4 . _
if @xmath8 divides @xmath178 and @xmath179 divides @xmath8 , then the scalar linear code is given by @xmath212 } for @xmath213 and @xmath214 .
@xmath215 + + @xmath216 , capacity=@xmath217 .
+ let @xmath144 for @xmath218 + + @xmath219 , + @xmath220 , + @xmath148 , + @xmath149 , + @xmath150 , + @xmath151 , + @xmath152 , + @xmath153 , + @xmath154 , + @xmath155 , + and @xmath221 + the proposed code is @xmath222 + + @xmath223 _ corollary 5 . _
if @xmath224 divides @xmath225 @xmath179 divides @xmath226 , then the proposed scalar linear code is @xmath227 .
+ + @xmath229 capacity=@xmath191 .
+ let @xmath127 for @xmath218 + + @xmath230 , + @xmath130 , + @xmath131 , + @xmath132 , + @xmath133 , + @xmath134 , + @xmath135 , + @xmath136 , + @xmath137 , + @xmath138 , + and @xmath231 + + the proposed code is @xmath232 + + @xmath233 + _ corollary 6 . _ for the case @xmath234 divides @xmath61 @xmath179 divides @xmath8 , the scalar linear code is + @xmath235 } where @xmath236 and @xmath237 @xmath238 + + @xmath239 , capacity=@xmath240 .
+ let @xmath127 for @xmath241 + + @xmath242 , + @xmath243 , + @xmath244 , + @xmath245 , + @xmath246 , + @xmath247 , + @xmath248 , + @xmath249 , + @xmath250 , + + the proposed code is @xmath251 .
+ + @xmath252 _ corollary 7 . _
if divides @xmath0 and @xmath179 divides , then the scalar linear code is given by @xmath253 } where @xmath254 and @xmath255 .
@xmath256 + + @xmath257 , capacity=@xmath191 .
+ let @xmath127 for @xmath258 + + + @xmath259 , + @xmath260 , + @xmath261 , + @xmath262 , + @xmath263 , + @xmath264 , + @xmath265 , + @xmath266 , + @xmath267 , + @xmath268 , + @xmath269 , + @xmath270 .
+ + the proposed code is @xmath271 .
+ + @xmath272 + _ corollary 8 .
_ if @xmath8 divides @xmath182 and @xmath179 divides @xmath8 , then the scalar linear code is given by @xmath273 where @xmath274 , @xmath275 mod @xmath8 = @xmath276 .
@xmath277 + + @xmath278 , capacity=@xmath279 .
+ let @xmath144 for @xmath280 + + @xmath281 , + @xmath147 , + @xmath148 , + @xmath149 , + @xmath150 , + @xmath151 , + @xmath152 , + @xmath153 , + @xmath154 , + @xmath282 + + the proposed code is @xmath283 @xmath284 _ corollary 9 .
_ if @xmath224 divides @xmath182 and @xmath179 divides @xmath224 , then the scalar linear code + @xmath1=@xmath285 + where @xmath188 , @xmath286 , @xmath287 and .
@xmath288 + + @xmath289 , capacity=@xmath290 .
+ let @xmath127 for @xmath291 + + @xmath292 , + @xmath293 , + @xmath294 , + @xmath295 , + @xmath296 , + @xmath297 , + @xmath298 , + @xmath299 , + @xmath300 , + @xmath301 , + @xmath302 , + @xmath303 , + @xmath304 , + @xmath305 .
+ + the proposed code is @xmath306 + + @xmath307 +
in this paper a construction is given for vector linear index codes of multiple unicast index problems from scalar linear codes which results in a sequence of index coding problems with same number of messages and receivers and two sided antidote patterns .
moreover , it is shown that if the problem with which the construction begins has an optimal linear index code then it induces an optimal linear index code for the vector index coding problem .
this construction has been used on few classes of index coding problems given in @xcite for which optimal linear index codes are known and new classes codes have been obtained starting from these classes of codes .
+ another interesting direction of further research is to study the suitability of the new classes of codes presented in this paper for application to noise broadcasting problem .
recently , it has been observed that in a noisy index coding problem it is desirable for the purpose of reducing the probability of error that the receivers use as small a number of transmissions from the source as possible and linear index codes with this property have been reported in @xcite , @xcite . while the report @xcite considers fading broadcast channels , in @xcite awgn channels
are considered and it is reported that linear index codes with minimum length ( capacity achieving codes or optimal length codes ) help to facilitate to achieve more reduction in probability of error compared to non - minimum length codes for receivers with large amount of side - information .
these aspects remain to be investigated for the new classes of sequences of vector codes presented in this paper .
l ong and c k ho , `` optimal index codes for a class of multicast networks with receiver side information '' , in _ proc .
ieee icc _ , 2012 , pp .
2213 - 2218 .
roop kumar bhattaram , mahesh babu vaddi and b. sundar rajan , `` a lifting construction for scalar linear index codes , '' arxiv:1510.08592 [ cs.it ] 29 , oct 2015 .
anoop thomas , kavitha radhakumar , attada chandramouli and b. sundar rajan , `` optimal index coding with min - max probability of error over fading , '' pimrc 2015 , hong kong , august 2015 . | this paper deals with vector linear index codes for multiple unicast index coding problems where there is a source with @xmath0 messages and there are @xmath0 receivers each wanting a unique message and having symmetric ( with respect to the receiver index ) two - sided antidotes ( side information ) . optimal scalar linear index codes for several such instances of this class of problems for one - sided antidotes(not necessarily adjacent ) have been reported in @xcite .
these codes can be viewed as special cases of the symmetric unicast index coding problems discussed in @xcite with one sided adjacent antidotes . in this paper , starting from a given multiple unicast index coding problem with with @xmath0 messages and one - sided adjacent antidotes for which a scalar linear index code @xmath1 is known , we give a construction procedure which constructs a sequence ( indexed by @xmath2 ) of multiple unicast index problems with two - sided adjacent antidotes ( for the same source ) for all of which a vector linear code @xmath3 is obtained from @xmath4 also , it is shown that if @xmath1 is optimal then @xmath3 is also optimal for all @xmath5 we illustrate our construction for some of the optimal scalar linear codes of @xcite though the construction is applicable for all the codes of @xcite . |
the twin spacecraft making up the stereo ( solar terrestrial relations observatory ) mission have already provided a rich dataset since launch in october 2006 , with images , both in the visible and in the extreme ultraviolet ( euv ) @xcite .
these images are collected by the euvi ( extreme ultraviolet imager ) a and b imager and cor ( coronagraph ) 1 and cor2 a and b coronagraphs on the stereo a and b satellites as part of the secchi ( sun - earth connection coronal and heliospheric investigation ) instrument packages @xcite .
each secchi suite consists of the following five telescopes : the euvi imager and cor1 and cor2 coronagraphs comprise the scip package pointing to the sun , complemented by the off - pointing hi ( heliospheric imager ) 1 and hi2 imagers observing the heliosphere .
all imagers together cover the observational range from the sun to beyond 1 au . besides the classical coronal mass ejections ( cmes ) that are identified in the fovs of the cor imagers ( _ e.g. _ ) , small scale collimated ejections of solar plasma are another common feature ( _ e.g. _ ) .
these coronal jets are best observed inside polar coronal holes at euv wavelengths when the plasma beams are seen in emission against the dark background and are not obscured by bright ambient coronal structures .
very early observations of coronal jets were provided by skylab and , in more details , later by yohkoh in the soft x - rays @xcite .
polar coronal jets were studied by using images of the lasco ( large angle and spectrometric coronagraph ) and eit ( extreme ultraviolet imaging telescope ) instruments on board soho ( solar and heliospheric observatory ) .
further observations were reported by with the trace ( transition region and coronal explorer ) spacecraft .
more recently , also hinode has provided important data on polar jet parameters @xcite .
x - ray jets have typical lengths of @xmath2@xmath3 km , widths of @xmath4@xmath5 km , and speeds ranging from 10 to 1000 km / s @xcite .
it is usually assumed that the jet is the result of magnetic reconnection phenomena happening in the solar corona @xcite .
coronal jets can be observed both in active regions , quiet sun , and coronal holes @xcite .
a polar coronal jet has been studied for the first time stereoscopically with observations from the stereo / secchi imagers by , demostrating the existence of helical magnetic field lines in this event .
the two points of view provided by the twin stereo satellites are very helpful for the identification of coronal hole jets .
indeed , for the first time it is possible with stereo / secchi data to assess what is the three dimensional ( 3d ) structure of the jet , to understand what projection effects are present in single point observations , and to try to establish the true 3d velocity for the jet @xcite . here
we present the first catalogue of stereo polar coronal jets , comprising 79 events , which were observed by both euvi imagers in the ultraviolet , and contemporarily by the cor1 coronagraph in white - light , during the period at solar minimum at the end of solar cycle 23 from march 2007 to april 2008 .
this time interval , during which the separation angle between the two spacecraft increased from 2 to 48 degrees with respect to the sun , includes two time periods of high time cadence observations in may 2007 and january 2008 . the criteria according to which coronal jet events were identified and selected from the secchi dataset are explained in section 2 of this paper , including some brief summary of their basic characteristics .
prototype events which also emphasize the difference of the viewpoints when seen either from stereo a or from stereo b , are shown in section 3 . in section 4 the lifetime distributions of jets in the different wavelengths in the euvi and in cor1
fov are discussed .
the conclusions are given in section 5 , followed by the catalogue of polar coronal jets in the appendix .
secchi images are taken at four euv wavelengths centered at 171 , 195 , 284 with a time cadence of 2.5 , 10 , 20 minutes , and at 304 with a time cadence of 10 minutes . during high time cadence observations on may 519 , 2007 , and on january 720 , 2008
the time cadences was 75 s in the 171 channel .
the pixel size of the euvi ccd on the solar disk has a spatial resolution of 1.6 arcsec . because of the low time cadence of the 284 observations , in this jet study more events
are seen in the 171 , 195 and 304 wavelength ranges .
we began our systematic study of coronal jets in the secchi euvi a and b data after stereo s completed in - orbit insertion and the official science mission start on january 22 , 2007 , and after some follow on period of routine imaging , in march 2007 .
for the search of coronal jets euvi images were used that are provided at the secchi website http://secchi.nrl.navy.mil/. secchi euvi a and b daily movies are available at http://cdaw.gsfc.nasa.gov / stereo / daily_movies/. the existence of subsequent jets in the cor1 a and b fovs were studied in the daily movies in intensity running difference available at http://cor1.gsfc.nasa.gov / dailymov / mpeg/. the running difference movies allow to identify faint transient coronal structures , _
e.g. _ those hidden by coronal streamers .
it is a pleasure to thank all the stereo staff , without which this work would not have been possible .
g. n. acknowledges support from an erasmus grant during his stay in goettingen .
v.b acknowledges the support of the project stereo / corona by the german bundesministerium fr bildung und forschung through the deutsche zentrum fr luft - und raumfahrt e.v .
( dlr , german space agency ) as a collaborative effort with the max - planck - institut fr sonnensystemforschung ( mps ) under grant 50 0c 0904 .
stereo / corona is a science and hardware contribution to the optical image package secchi , developed for the nasa stereo mission launched in 2006 .
the secchi data used here were produced by an international consortium of the naval research laboratory ( usa ) , lockheed martin solar and astrophysics lab ( usa ) , nasa goddard space flight center ( usa ) , rutherford appleton laboratory ( uk ) , university of birmingham ( uk ) , max - planck - institut for solar system research ( germany ) , centre spatiale de lige ( belgium ) , institut doptique thorique et applique ( france ) , and institut dastrophysique spatiale ( france ) . g. z. was supported in part by the italian inaf and by the italian space agency , contract asi n. i/015/07/0 `` esplorazione del sistema solare '' .
we would like to thank the referee for his / her careful and constructive criticism which has allowed us to substantially improve the paper .
cirtain , j. ; golub , l. ; lundquist , l. ; ballegooijen , van a. ; savcheva , a. ; shimojo , m. ; deluca , e. ; tsuneta , s. ; sakao , t. ; revves , k ; weber , m. ; kano , r. ; nakurage , n. ; shibasaki , k. : 2007 , _ science _ * 318 * , 1580 .
howard , r.a . ,
moses , j.d .
, vourlidas , a. , newmark , j.s . ,
socker , d.g . ,
plunkett , s.p . ,
korendyke , c.m . ,
cook , j.w . ,
hurley , a. , davila , j.m . ,
thompson , w.t . , st .
cyr , o.c .
, mentzell , e. , mehalick , k. , lemen , j.r . ,
welser , j.p . ,
duncan , d.w . ,
tarbell , t.d .
, wolfson , c.j . , moore , a. , harrison , r.a . ,
waltham , n.r .
, lang , j. , davis , c.j . ,
eyles , c.j .
, mapson - menard , h. , simnett , g.m . ,
halain , j.p . , defise , j.m . ,
mazy , e. , rochus , p. , mercier , r. , ravet , m.f . , delmotte , f. , auchre , f. , delaboudinire , j.p . , bothmer , v. , deutsch , w. , wang , d. , rich , n. , cooper , s. , stephens , v. , maahs , g. , baugh , r. , mcmullin , d. , carter , t. : 2008 , _ space sci .
rev . _ * 136 * , 67 .
savcheva , a. ; cirtain , j. ; deluca , e. e. ; lundquist , l. l. ; golub , l. ; weber , m. ; shimojo , m. ; shibasaki , k. ; sakao , t. ; narukage , n. ; tsuneta , s. ; kano , r. : 2007 , _ pub .
. japan _ * 59 * , s771 .
shibata , k. , ishido , y. , acton , l. w. , strong , k. t. , hirayama , t. , uchida , y. , mcallister , a. h. , matsumoto , r. , tsuneta , s. , shimizu , t. , hara , h. , sakurai , t. , ichimoto , k. , nishino , y. , ogawara , y. : 1992 , _ pub .
japan _ textbf44,l173 . | in this paper we present the first comprehensive statistical study of euv coronal jets observed with the secchi ( sun earth connection coronal and heliospheric investigation ) imaging suites of the two stereo spacecraft .
a catalogue of 79 polar jets is presented , identified from simultaneous euv and white - light coronagraph observations , taken during the time period march 2007 to april 2008 , when solar activity was at minimum . the twin spacecraft angular separation increased during this time interval from 2 to 48 degrees .
the appearances of the coronal jets were always correlated with underlying small - scale chromospheric bright points .
a basic characterisation of the morphology and identification of the presence of helical structure were established with respect to recently proposed models for their origin and temporal evolution . though each jet appeared morphologically similar in the coronagraph field of view , in the sense of a narrow collimated outward flow of matter , at the source region in the low corona the jet showed different characteristics which may correspond to different magnetic structures .
a classification of the events with respect to previous jet studies shows that amongst the 79 events there were 37 _
eiffel tower_-type jet events commonly interpreted as a small - scale ( @xmath0 arcsec ) magnetic bipole reconnecting with the ambient unipolar open coronal magnetic fields at its looptops , and 12 _ lambda_-type jet events commonly interpreted as reconnection with the ambient field happening at the bipoles footpoints .
five events were termed _
micro - cme _ type jet events because they resembled the classical coronal mass ejections ( cmes ) but on much smaller scales .
the remainig 25 cases could not be uniquely classified .
thirty - one of the total number of events exhibited a helical magnetic field structure , indicative for a torsional motion of the jet around its axis of propagation .
a few jets are also found in equatorial coronal holes . in this study
we present sample events for each of the jet types using both , stereo a and stereo b , perspectives .
the typical lifetimes in the secchi / euvi ( extreme ultraviolet imager ) field of view between 1.0 to 1.7 @xmath1 and in secchi / cor1 field of view between 1.4 to 4 @xmath1 are obtained , and the derived speed are roughly estimated . in summary ,
the observations support the assumption of continuous small - scale reconnection as an intrinsic feature of the solar corona , with its role for the heating of the corona , particle acceleration , structuring and acceleration of the solar wind remaining to be explored in more details in further studies . |
this paper is concerned with invariants arising from witten - reshetikhin - turaev @xmath8 topological quantum field theories ( tqft ) following the skein theoretical approach of @xcite .
such a tqft defines for @xmath9 a oriented compact @xmath10-manifold without boundary and @xmath11 a banded link , a sequence of invariants @xmath12 indexed by even integers @xmath13 . for a given @xmath14 , the invariant @xmath12 belongs to the cyclotomic field @xmath15 / ( \phi_{2p}(a))$],where @xmath16 denotes the @xmath1-th cyclotomic polynomial . to have a numerical invariant
, one needs to specify an embedding of @xmath17 into @xmath18 or equivalently a @xmath1-th primitive root of unity @xmath19 .
we will denote by @xmath20 the associated numerical evaluation .
an interesting question is to understand the asymptotic of the quantity @xmath21 as @xmath22 and as @xmath5 converges to a given number on the unit circle . when @xmath23 , this problem is called the witten asymptotic expansion conjecture .
other limits have not been studied yet with the exception of some seifert spaces studied by lawrence and zagier , see @xcite . in this paper , we focus on the case @xmath24 where @xmath25 is a compact connected oriented closed surface .
we look for a formula for the quantum invariant @xmath26 where @xmath27 is a given banded link .
notice that the quantity @xmath28 is the dimension of @xmath29 : the @xmath17-vector space associated to @xmath25 by the witten - reshetikhin - turaev tqft .
moreover , @xmath30 is computed by the verlinde formula and is polynomial in @xmath14 with degree @xmath31 if the genus of @xmath25 is @xmath32 or @xmath33 respectively .
hence the asymptotics of the quantity @xmath34 is determined by the asymptotics of @xmath35 .
the main result of this paper is that the asymptotics of @xmath35 is almost determined by the evaluation of a laurent polynomial with integral coefficients depending only on @xmath0 .
[ thm1 ] let @xmath0 be a link in @xmath36 .
there exists a laurent polynomial @xmath37 $ ] and a finite set @xmath38 such that for any sequence @xmath39 such that @xmath40 , one has @xmath41 in particular the polynomial @xmath42 $ ] is well - defined and is a topological invariant of @xmath0 .
the polynomial @xmath3 can be viewed as a generalization of the kauffman bracket of a link in @xmath43 .
indeed if @xmath44 is a link inside a @xmath10-ball embedded in @xmath36 , the polynomial @xmath3 is nothing but the usual kauffman bracket of @xmath0 . for a more complicated link inside @xmath36
, this polynomial can be computed algorithmically , see section [ skein ] .
the existence of such a polynomial is not clear a priori , we note that costantino already built one in the case of a @xmath45 using other methods ( see @xcite ) . in order to prove theorem [ thm1 ] , we introduce the key notion of _ cyclic expansion_. we will denote by @xmath46 the vector space of maps @xmath47 which are continuous with compact support and piecewise polynomial . we will say that the sequence @xmath48 has a cyclic expansion if there exists @xmath37 $ ] , an integer @xmath49 and a family @xmath50 such that @xmath51 we need to explain the meaning of @xmath52 in equation ( [ cyclic ] ) .
indeed , we have to interpret both sides as elements of @xmath53 endowed with the norm @xmath54 .
the main technical result of this article is the following theorem .
[ thm existence ] for any banded link @xmath27 , the sequence @xmath48 admits a cyclic expansion .
the proof of this theorem consists in a careful counting of integral points in various polytopes related to tqft .
it is postponed to section [ proof cyclic ] .
the interest of having such a cyclic expansion is that the study of asymptotics is reduced to the following question .
let @xmath55 be in @xmath46 , @xmath56 be a non zero integer and @xmath5 be a convergent sequence of @xmath1-th primitive roots of unity .
what is the asymptotics of @xmath57 as @xmath14 tends to infinity ?
this problem can be solved with elementary analytic tools as follows .
[ limits ] let @xmath58 and @xmath56 be a positive integer .
let @xmath59 be a sequence of @xmath60-th primitive roots of unity with @xmath61 a strictly increasing sequence of even integers . 1 .
if @xmath62 with @xmath63 one has @xmath64 2 .
if @xmath62 with @xmath65 we write @xmath66 so that @xmath67 .
if @xmath68 diverges when @xmath69 goes to infinity then @xmath70 3 . in the same setting as ( 2 )
suppose that @xmath68 does not diverge .
as @xmath71 , the sequence @xmath72 takes discrete values and up to extracting a subsequence , one can suppose that it is constant equal to @xmath73 , i. e. @xmath74 for some odd integer @xmath75 .
then , we have @xmath76 we set @xmath77 .
we compute @xmath78 the last equality is obtained by applying a taylor expansion of @xmath55 away from a finite number of values : this is possible since @xmath55 is piecewise polynomial . in the first case , since @xmath63 , we deduce that @xmath79 . in the second case , we simply observe that @xmath80 hence @xmath81 and we can conclude . in the last case ,
we write @xmath82 and recognize a riemann sum .
this gives @xmath83 .
we observe that proposition [ limits ] and theorem [ thm existence ] imply directly theorem [ thm1 ] .
the polynomial @xmath3 associated to the cyclic expansion can be used for proving the amu conjecture .
let @xmath25 be a closed surface of genus at least @xmath84 , @xmath85 be an even integer and @xmath86 be the quantum representation - tqft but the exact same can be done in the @xmath8 setting .
] considered by koberda and the second author in @xcite . here
@xmath87 denotes the vector space associated by the @xmath8 witten - reshetikhin - turaev tqft to the surface @xmath25 equipped with a banded point colored by @xmath88 .
it corresponds to the @xmath8-tqft at level @xmath89 in the geometric setting . here
, we use the notation of @xcite with a shift of 1 concerning the color @xmath88 .
a consequence of the amu conjecture stated by andersen , masbaum and ueno in @xcite is the following ( see ( * ? ? ?
* section 7 ) for more detail on this implication ) . if @xmath90 is not a power of a simple element then @xmath91 has infinite order for all @xmath14 big enough . here
a non - trivial element of @xmath92 is said simple if it is freely homotopic to a simple closed curve .
we denote by @xmath93\right ) ^{\times } = \ { \pm a^m \ } _ { m \in { \mathbb{z}}}$ ] the group of units in @xmath94 $ ] . using cyclic expansions we define the following : [ gammachapeau ] let @xmath95 be represented by a loop @xmath96 .
we define @xmath97 as the knot @xmath98 with arbitrary banded structure . for @xmath99 , we denote by @xmath100 the banded link @xmath101 colored by @xmath88 and by @xmath102 its normalized trace ( observe that @xmath103 is the trivial color and @xmath104 the usual one ) . because the indeterminacy of the banded structure , the polynomial @xmath105 is well - defined up to the multiplication by a unit .
[ thm amu ] let @xmath95 . if for some @xmath88 , @xmath105 is neither zero nor a unit , then the amu conjecture for surface groups holds for @xmath106 .
we have by definition @xmath107 where @xmath108 is the @xmath88-th factor of the representation @xmath109 defined in equation [ rhop ] .
if @xmath110 has finite order , then its eigenvalues are roots of unity and we have for any primitive root of unity of order @xmath1 the inequality @xmath111 by proposition [ limits ] , for all but a finite number of @xmath112 , one has @xmath113 hence , if @xmath114 has finite order for a sequence @xmath61 going to infinity , then @xmath115 for all @xmath112 .
the theorem is then a direct consequence of lemma [ parseval ] .
[ parseval ] let @xmath116 $ ] such that @xmath117 .
then one has @xmath118 or @xmath119\right ) ^{\times}$ ] .
we can write @xmath120 with the @xmath121 s in @xmath122 .
let us define the continuous function @xmath123 .
one has @xmath124 since the @xmath121 s are in @xmath122 this implies that @xmath118 or @xmath125 for some @xmath126 .
notice that determining if the polynomials @xmath105 belong to @xmath93\right ) ^{\times } \cup \{0 \ } $ ] is similar to the problem of determining if a non - trivial knot in @xmath43 have non - trivial colored jones polynomials , which is believed to be true .
the following theorem [ incompressible ] will provide examples where the degree of the polynomial @xmath127 can be controlled .
this is reminiscent to the estimation of the colored jones polynomials of alternating knots .
we give here a formula for @xmath3 in the case where @xmath0 is a banded link in @xmath128 whose projection on the first factor is a multicurve .
[ algo ] let @xmath129 be the first projection map and consider a collection of disjoint and non - parallel annuli @xmath130 .
let @xmath131 be a banded link projecting to @xmath132 and set @xmath133 .
we have @xmath134 so that one can reduce to the case when @xmath0 is inside an annulus times a circle .
we put @xmath135 if @xmath0 is the empty link .
let @xmath0 be a banded link in @xmath136 $ ] which is a union of banded knots @xmath137 for @xmath138 .
let @xmath139 be the corresponding homology classes .
we have @xmath140 where @xmath141 is the area of the polygon in @xmath142 whose sides are the vectors @xmath143 .
this formula will be useful for proving that @xmath127 is non - trivial for some curves @xmath144 . in order to describe them ,
we introduce the notion of euler incompressibility .
a cycle in a graph @xmath145 is called _ eulerian _ if it visits every edge at most once .
an graph @xmath145 embedded in @xmath25 is said _ euler - incompressible _ when no eulerian cycle of @xmath145 bounds a disc in @xmath25 .
[ incompressible ] let @xmath25 be a closed surface of genus @xmath33 and @xmath146 be an embedding with @xmath147 transverse double points whose image @xmath148 is euler - incompressible .
then , up to a unit , we have @xmath149 where @xmath150 and @xmath151 are not zero . in particular , it is not trivial if @xmath152 and theorem [ thm amu ] applies .
figure [ filling ] shows an example of curve which fulfills the assumptions of theorem [ incompressible ] .
note that this loop is filling the surface in the sense that the complement of its image is a disjoint union of topological discs . according to kra s theorem (
* theorem 1.1 ) , this loop is sent to a pseudo - anosov element in the birman exact sequence . to the authors knowledge , this is the first example of a pseudo - anosov element in the mapping class group of a genus @xmath33 surface satisfying the amu conjecture .
moreover we note that any loop whose image s complementary is a single disc fulfills the assumptions of theorem [ incompressible ] . the standard root of unity @xmath23 is generally used in tqft as the vector space @xmath29 is hermitian in that case .
most of the results or conjectures about the asymptotics of tqft concern this setting .
the witten conjecture is the most well - known and is related to our work in the case when the underlying manifold is @xmath128 .
consider for more generality an odd integer @xmath75 and replace @xmath5 with @xmath153 .
observe that for these roots to be primitive of order @xmath1 we need that @xmath1 is coprime to @xmath75 which we assume from now on .
suppose that @xmath131 is a banded link having a cyclic expansion as in equation .
then this expansion governs the asymptotics in the sense that we have @xmath154 let @xmath155 be the space of conjugacy classes of irreducible representations @xmath156 su@xmath157 .
denoting by @xmath158 the generator of @xmath159 , such a representation has to satisfy @xmath160 as @xmath158 is central .
hence @xmath155 is a union of two copies of @xmath161 , defined in the same way .
this manifold has dimension @xmath162 and is endowed with a natural symplectic form @xmath163 and volume form @xmath164 .
we set @xmath165 .
when @xmath166 , the witten conjecture predicts the following asymptotics where @xmath167 .
@xmath168 this formula was proved in @xcite in the case when @xmath0 lies inside @xmath169\subset \sigma\times s^1 $ ] .
this formula has also an intersection with @xcite , where special links in finite order mapping tori were studied .
theorem [ witten - sigma ] will cover the general case where @xmath0 and @xmath75 are arbitrary .
unfortunately , the geometric meaning of the formula is less clear when @xmath170 .
[ witten - sigma ] let @xmath75 be an odd integer and set @xmath23 .
if @xmath131 projects to @xmath25 without crossings , we have the following formula .
@xmath171 this formula extends to all links using the kauffman relation , however we do not know a direct expression of @xmath172 for general banded links .
the fact the @xmath75 exponent moved from @xmath5 to @xmath173 is a striking phenomenon which deserve further study .
the theorem will be a rather direct consequence of theorem [ sigma ] proved in section [ geom ] .
the computation of the polynomial @xmath3 associated to a banded link @xmath27 is better understood in terms of skein modules . for any compact oriented manifold @xmath9 ( maybe with boundary ) , we denote by @xmath174 the kauffman bracket skein module with coefficients in @xmath94 $ ] .
we recall that it is the free @xmath175$]-module generated by the set of isotopy classes of banded links in the interior of @xmath9 quotiented by the following relations .
first @xmath176 where @xmath177 , @xmath178 , @xmath179 are any three banded links in @xmath9 which are the same outside a small @xmath10-ball but differ inside as in figure [ ktriple ] . in this case , the triple @xmath177 , @xmath178 , @xmath179 is called a kauffman triple .
the second relation satisfied in @xmath174 is @xmath180 where @xmath0 is any link in @xmath9 and @xmath181 is a trivial banded knot .
[ skein_cyclo ] let @xmath177 , @xmath178 , @xmath179 be a kauffman triple in @xmath36 .
if the sequences @xmath182 and @xmath183 have a cyclic expansion then the sequence @xmath183 has a cyclic expansion .
moreover @xmath184 remark that the witten - reshetikhin - turaev invariants satisfy the skein relation : @xmath185 .
therefore , it is enough to prove that cyclic expansions well behave under multiplication by @xmath186 and finite sum . for @xmath187 an integer , @xmath188 we define @xmath189
let @xmath190 be an integer and let @xmath191 . applying a taylor expansion to the functions @xmath192 we have the following @xmath193 . hence the sequence @xmath194 admits a cyclic expansion .
this says that if a sequence admits a cyclic expansion then the same sequence multiplied by @xmath195 also admits a cyclic expansion .
let @xmath196 be an integer divisible by @xmath56 and set @xmath197 .
one can make the euclidean division @xmath198 and write @xmath199by the previous argument and the estimation @xmath200 , we find that there exists @xmath201 such that @xmath202 . from this
we can deduce that the sum of two sequences admitting cyclic expansions also admits a cyclic expansion .
this proposition means that the polynomial @xmath3 extends to a @xmath94$]-linear map @xmath203 $ ] defined by @xmath204 .
if @xmath0 sits inside a ball @xmath205 , the polynomial @xmath3 is just the kauffman bracket of @xmath0 . hence the map @xmath206 defines a section of the natural inclusion map @xmath207 .
we will construct @xmath206 by giving its value on a @xmath94$]-span of @xmath208 . that this map is well - defined
is a non - trivial consequence of the existence of tqft invariants and properties of the cyclic expansions . before that , we need to recall the results concerning the multiplicative structure of the skein module of the torus times an interval .
we denote by @xmath209 the torus @xmath210 .
it was proven by frohman - gelca and sallenave ( see @xcite ) that the skein module of the torus @xmath209 is isomorphic to the symmetric part of the quantum torus .
formally , we define the quantum torus as the non - commutative @xmath94$]-algebra @xmath211\langle m^{\pm 1},l^{\pm 1}\rangle /(lm - a^2ml)$ ] .
let @xmath75 be the involution of @xmath212 defined by @xmath213 .
for any @xmath214 we set @xmath215 there is an isomorphism of algebras @xmath216)\overset{\sim}{\to } \mathcal{t}^{\sigma}$ ] which maps the standard banded curve in @xmath217 $ ] with slope @xmath218 to the element @xmath219 when @xmath220 .
we have for any @xmath221 the following product - to - sum formula : @xmath222 for technical reasons , we will also need the following normalization @xmath223 . [ remfrogel ]
this proposition implies that @xmath224)$ ] is generated as a @xmath175$]-module by @xmath225 .
we introduce weighted multicurves as a set of generators of @xmath226 over @xmath94 $ ] .
let @xmath227 , a @xmath69-multicurve is a collection of @xmath69 disjoint essential simple oriented and non pairwise parallel closed curves .
let @xmath228 be a @xmath69-multicurve .
a weight on @xmath106 is an element @xmath229 thought as an assignment of a pair @xmath230 for each connected component @xmath231 of @xmath106 .
a pair @xmath232 will be called a weighted multicurve .
let @xmath228 be a @xmath69-multicurve on @xmath25 with weight @xmath233 . for @xmath234 , we choose a diffeomorphism between @xmath7 and @xmath235 respecting the orientation .
denote by @xmath209 the torus @xmath236 : we can embed @xmath209 in @xmath128 by sending the first factor to @xmath235 .
this embedding extends to an embedding @xmath237\to \sigma\times s^1 $ ] respecting the orientation of @xmath128 .
the skein associated to the weighted multicurve @xmath232 is the element @xmath238=\bigcup_{j=1}^k \phi_j(\langle a_j , b_j\rangle_t)\in \mathcal{k}(\sigma\times s^1)$ ] where @xmath239 and @xmath240 is the skein element defined in subsection [ frogel0 ] .
[ multi_span ] the set of weighted multicurves @xmath238 $ ] spans the @xmath94$]-module @xmath226 .
let @xmath27 be a banded link
. set @xmath241 \}$ ] and @xmath242 .
there is a finite set @xmath243 of banded points in @xmath25 ( perhaps empty ) so that one has up to isotopy @xmath244 .
let @xmath245 be the intersection of @xmath0 with @xmath246 .
we have the following picture : > 0 18inkscape -z -d file = multi.svg export - pdf = multi.pdf export - latex applying the skein relations we can reduce to the case where the projection of @xmath245 on @xmath25 is finite disjoint union of banded simple closed curves and banded points .
hence we can find an integer @xmath227 and a @xmath69-multicurve @xmath247 in @xmath25 such that , up to isotopy , @xmath248 where @xmath249 is a tubular neighborhood of @xmath106 .
this says that @xmath0 is in the image of the canonical map @xmath250)^{\otimes k}\to \mathcal{k}(\sigma \times s^1)$ ] induced by the inclusion @xmath251 .
finally from remark [ remfrogel ] we conclude that @xmath0 is a @xmath175$]-linear combination of weighted multicurves where the underlying multicurve is @xmath106 .
theorem [ thm1 ] will follow from the following one which will be proved in section [ proof cyclic ] .
[ existence ] for any weighted curve @xmath238 $ ] the sequence @xmath252\right)$ ] has a cyclic expansion .
let @xmath69 be the number of connected components of @xmath106 .
one has @xmath253)=2^k$ ] if @xmath254 and @xmath255 otherwise .
this theorem conjugated with the following lemma proves proposition [ algo ] . [ area ]
let @xmath209 be the standard torus and @xmath256 be the 1-form on @xmath257 given by @xmath258 where @xmath259 stands for the intersection product .
let @xmath260 be vectors in @xmath261 .
we denote by @xmath262 the concatenation of the segments generated by the vectors @xmath260 . for @xmath263 ,
we denote by @xmath264 the corresponding vector in @xmath224)$ ] .
we have then the following formula : @xmath265 this is a generalization of the product - to - sum formula which can be proved by an immediate induction , observing that @xmath266 .
the main application of this lemma will concern the case when @xmath267 in which case stokes formula and the equality @xmath268 implies that @xmath269 . in this section ,
we prove theorem [ incompressible ] .
we define the degree of a non - zero laurent polynomial by the following formula : @xmath270 we fix a surface @xmath25 of genus @xmath271 and start the proof with two technical lemmas .
[ graph height ] let @xmath145 be a graph embedded in @xmath25 with one oriented edge . consider a tubular neighborhood @xmath272 of @xmath145 and a map @xmath273 which is constant equal to @xmath274 out of the labeled edge and makes one positive turn along the oriented edge .
we define the following banded graph : @xmath275 moreover we denote by @xmath276 the banded link in @xmath36 defined by the boundary of @xmath277 . given a banded graph @xmath278 .
we say that a component @xmath279 of @xmath145 bounds a disc if there is an embedded disc @xmath280 such that @xmath281 .
[ bound ] let @xmath145 be a graph embedded in @xmath25 with one oriented edge and suppose that the component of @xmath145 containing the arrow is not the boundary of a disc
. then the degree of @xmath282 is bounded by twice the number of components of @xmath276 bounding a disc in @xmath128 .
each trivial component produces a factor @xmath283 $ ] by the kauffman relations : this shows that the bound is optimal .
remove all these components from @xmath276 .
we are reduced to prove that the polynomial @xmath282 has degree zero .
represent the directed edge by 2 arrows on the multicurve @xmath284 .
the lemma follows from a case by case study of the possible configurations of arrows .
* case 1 : * the two arrows belong to the same component and cancel .
this component is non - trivial as we removed them before starting .
we conclude with the following observation : the polynomial associated to a multicurve in @xmath25 is a constant . indeed , the multicurve is a union of parallel copies of curves of type @xmath285 .
lemma [ area ] gives @xmath286 which is an integer .
* case 2 : * the two arrows belong to the same component @xmath106 and add .
this is indeed impossible by considering @xmath106 as a boundary curve of the component of the banded graph containing the arrows . by construction
, the two arrows are in opposite directions relatively to this orientation and hence can not add .
* case 3 : * the two arrows belong to two non parallel components . hence , one of them is non - trivial and its neighborhood has the form @xmath287 .
invoking lemma [ area ] , there are no closed path in the expansion of @xmath288 , hence @xmath289 .
* case 4 : * the two arrows belong to two parallel and trivial components .
then these curves bound a trivial circle containing the arrow which is forbidden by assumption .
* case 5 : * the two arrows belong to parallel and consecutive non - trivial components .
then this corresponds to @xmath290 .
again by lemma [ area ] , any closed path in the expansion of @xmath288 has vanishing area , hence @xmath291 is an integer . *
case 6 : * the two arrows belong to parallel and non - consecutive non - trivial components .
then , noticing that there exists an embedded arc which join them , we are in the configuration @xmath292 .
we conclude as in the case 5 .
[ inequalities ] let @xmath145 be an embedded graph with one oriented edge .
let @xmath88 and @xmath293 be respectively the number of vertices and edges of @xmath145 .
let * @xmath6 be the number of components of @xmath294 bounding simply connected components of @xmath277 * @xmath295 be the number of components of @xmath294 bounding a disc and not counted in @xmath6 .
suppose that @xmath145 is euler - incompressible and at most quadrivalent .
then we have the following inequalities
@xmath296 moreover @xmath297 implies that @xmath145 is a disjoint union of circles ( none of them bound a disc in @xmath25 since @xmath145 is euler - incompressible ) .
let @xmath298 be the connected components of @xmath145 with negative euler characteristic .
let @xmath299 ( resp .
@xmath300 ) be the number of vertices ( resp .
edges ) of @xmath301 . as @xmath301 is not a circle
, one has @xmath302 .
since @xmath301 is at most quadrivalent , we have @xmath303 and hence @xmath304 .
we compute @xmath305 form which we conclude @xmath306 .
now let @xmath307 be the number of boundary components of @xmath301 and @xmath308 be the number of boundary components of @xmath301 bounding a disc in @xmath25 .
since @xmath301 is euler - incompressible , we have @xmath309 .
consider the closed surface @xmath310 obtained by gluing discs to the boundary components of @xmath301 .
if @xmath310 is a sphere then any boundary component of @xmath301 is an eulerian cycle which implies @xmath311 by euler - incompressibility .
the equation @xmath312 follows in that case .
if @xmath310 is not a sphere , we have @xmath313 . from @xmath314 and @xmath315
we get @xmath316 in any cases , @xmath312 , and summing over @xmath317 we get @xmath318 . it remains to prove the last part of the lemma .
suppose that @xmath319 and fix @xmath320 .
we have @xmath321 .
if @xmath310 is a sphere we have @xmath322 : this is not possible since a graph should have at least one vertex .
therefore ( [ eq0 ] ) implies @xmath323 and @xmath324 .
hence @xmath301 is embedded in a torus @xmath310 in such a way that any of its boundary components bounds a disc in @xmath25 .
this allows to define an embedding @xmath325 : since the genus of @xmath25 is at least two , this is not possible .
we conclude that @xmath326 is empty and @xmath145 is a disjoint union of circles .
we denote by @xmath327\in \mathcal{k}(\sigma\times s^1)$ ] the skein element obtained by coloring @xmath101 ( see definition [ gammachapeau ] ) by the color 3 ( 2 in the @xcite setting ) . in the setting of @xcite ,
this corresponds to the insertion of the idempotent @xmath328 in @xmath329 as shown in figure [ idempotent ] .
18inkscape -z -d file = f2.svg export - pdf = f2.pdf export - latex for any banded graph @xmath330 , we will denote by @xmath331 $ ] the element of @xmath332 obtained by inserting the idempotent @xmath328 in @xmath333 at all edges of @xmath334 . setting @xmath335=a^2+a^{-2}$ ] ,
these skein elements satisfy the following skein relation : @xmath336{0.8 cm } \includegraphics[scale = 0.1]{cross4 } \end{minipage } = a^4 \ , \begin{minipage}[c]{0.8 cm } \includegraphics[scale = 0.1]{cross2 } \end{minipage}+ [ 2 ] \ , \ , \begin{minipage}[c]{0.8 cm } \includegraphics[scale = 0.1]{cross0 } \end{minipage } + a^{-4 } \ , \ , \begin{minipage}[c]{1 cm } \includegraphics[scale = 0.1]{cross3 } \end{minipage}\ ] ] we extend the map @xmath337 $ ] by tensoring with @xmath338 so that we can evaluate banded colored graphs .
the proof will consist in investigating all terms in the state sum suggested by equation .
set @xmath339 and orient the edge going through the base point .
let @xmath272 be the set of vertices of @xmath334 .
given @xmath340 , we define @xmath341 to be the graph obtained by transforming all vertices of @xmath334 as follows : if a vertex evaluates to @xmath274 under @xmath342 it is replaced by a positive smoothing , if a vertex evaluates to @xmath343 under @xmath342 it is replaced by a negative smoothing and if a vertex evaluates to @xmath255 under @xmath342 it is not changed .
@xmath341 can be viewed as a quadrivalent graph in @xmath25 with one directed edge . consider the banded graph @xmath344 in @xmath36 defined by definition [ graph height ] . for @xmath340 , we denote by @xmath345 respectively the number of preimages of @xmath346 . applying the rule to all the crossings of @xmath106 gives @xmath347 ) = \sum_{s : v \to \{-1,0,1\ } } \ , a^{4(a_s - b_s)}[2]^{n_s } \eta[\hat{\gamma}_s,3]\ ] ] consider one term of the sum associated to a map @xmath342 . to avoid heavy notations
, we remove the subscript @xmath342 to @xmath348 and @xmath88 .
we have the decomposition @xmath349 and @xmath88 is the number of vertices of @xmath341 .
the original graph @xmath334 was a quadrivalent graph with @xmath147 vertices hence satisfying @xmath350 .
the graph @xmath341 being obtained by smoothing @xmath351 vertices satisfies @xmath352 .
computing @xmath353 $ ] involves inserting idempotents in all edges of @xmath341 .
let @xmath354 be the set of edges of @xmath341 . for @xmath355
, we define @xmath356 to be the graph obtained by deleting all edges @xmath293 such that @xmath357 .
we define @xmath358 .
according to the rule of figure [ idempotent ] , we have @xmath359 = \sum_{\xi : e_s \to \{0,1 \ } } [ 2]^{-s_\xi}\eta(\partial \hat{\gamma}_{s,\xi})\ ] ] now fixing @xmath355 we want to bound @xmath360^{n - s}\eta ( \partial \hat{\gamma}_{s,\xi})$ ] ( here @xmath361 stands for @xmath362 ) .
note that @xmath363 and recall that as @xmath334 is euler - incompressible , this property still holds for @xmath356 .
hence lemma [ bound ] applies for @xmath356 and we have @xmath364^{n - s}\eta ( \partial \hat{\gamma}_{s,\xi } ) \le 4|a - b|+2(n - s+c)\ ] ] where @xmath365 is the number of components of @xmath366 which bound a disc in @xmath25 .
let us write @xmath367 where @xmath6 is the number of simply connected components of @xmath356 .
we denote by @xmath209 the left - hand side of equation . supposing that @xmath368 , we can apply lemma [ inequalities ] to @xmath356 .
hence ( [ eq1 ] ) gives @xmath369 ( we recall that @xmath370 and @xmath371 ) .
this is strictly less that @xmath372 unless @xmath373 and @xmath374 ( by lemma [ inequalities ] ) , that is unless we smoothed all the vertices of @xmath334 in the positive way .
doing so , @xmath341 is a collection of non - trivial curves ( since @xmath334 is euler - incompressible ) colored by @xmath10 , one of them being oriented .
the computation can be done locally in an annulus times an interval .
expanding the idempotent @xmath328 , an unoriented curve colored by @xmath10 is @xmath375=(1,0)_t^2-[\emptyset]=(2,0)_t+[\emptyset]$ ] and similarly , an unoriented curve colored by @xmath10 is equal to @xmath376=(2,2)_t+[\emptyset]$ ] . hence expanding @xmath377^n[(1,1),3][(1,0),3]^m)$ ] with the help of lemma [ area ] gives a positive result . repeating the argument for negative smoothing
, we finally proved theorem [ incompressible ] .
let @xmath378 be the basis of @xmath379 obtained by filling @xmath209 with @xmath380 and coloring the core curve by @xmath381 .
notice that we use the convention of @xcite , that is @xmath382 corresponds to @xmath383 where @xmath384 is the basis used in @xcite .
the algebra @xmath224)$ ] acts naturally on @xmath379 by stacking .
let us cover this action by an action of the quantum torus .
define @xmath385 as a @xmath14-dimensional hermitian vector space formally generated by @xmath386 where @xmath387 . by convention
, we set @xmath388 if @xmath389 and @xmath255 otherwise .
the quantum torus @xmath212 acts on @xmath390 by the formulas @xmath391 and @xmath392 .
let @xmath75 be the involution of @xmath390 defined by @xmath393 .
then the map @xmath394 defined by @xmath395 is an isometry onto the @xmath75-antisymmetric part which commutes with the action of @xmath396 .
this is well - known , see for instance ( * ? ? ?
* section 2 ) .
indeed , we get the expected formulas @xmath397 @xmath398 [ calcul]we have the formula @xmath399 this is a direct computation , replacing @xmath382 with @xmath400 .
we remark that because proposition [ multi_span ] and proposition [ skein_cyclo ] , it is enough to prove that any weighted multicurve has cyclic expansion .
let @xmath401 be a @xmath69-multicurve .
we denote by @xmath402 the surface obtained by surgery on every component of @xmath106 . for any @xmath403 , we add two marked points @xmath404 from each side of the handle used in the surgery at @xmath231 as in figure [ fig_surgery ] .
> 0 18inkscape -z -d file
= surgery.svg export - pdf = surgery.pdf export - latex finally , given a vector @xmath405 thought as a coloring of the points @xmath404 , we set @xmath406 because of let @xmath407 be a multicurve with weight @xmath239 .
let @xmath408 and @xmath409 be two parallel copies of @xmath235 such that @xmath410 .
consider the basis @xmath411 of @xmath412 obtained by filling @xmath413 with @xmath380 ( the same with @xmath409 ) .
the tqft axioms imply the following equation where @xmath88 and @xmath414 are elements of @xmath415 .
@xmath416 = \sum_{n , m } \left ( \prod_{j=0}^k \langle\langle a_j , b_j\rangle_t \ , e_{n_j},e_{m_j}\rangle\right ) \ , \frac{{c}_{\gamma , p}(n , m ) } { z_p(\sigma \times s^1)}\ ] ] using lemma [ calcul ] , we have @xmath417 = \sum_{n , m}\prod_{j=1}^k\big ( a^{2n_ja_j}(\delta^{p}_{n_j+b_j - m_j}-\delta^{p}_{n_j+b_j+m_j})+\\ a^{-2n_ja_j}(\delta^{p}_{n_j - b_j - m_j}-\delta^{p}_{n_j - b_j+m_j } ) \big)\frac{{c}_{\gamma , p}(n , m ) } { z_p(\sigma \times s^1)}\end{gathered}\ ] ] introducing signs @xmath418 and @xmath419 in @xmath420 , we can rewrite the sum in the following way @xmath416 = \sum_{n , m,{\varepsilon},\zeta } \prod_j a^{2{\varepsilon}_jn_ja_j}\zeta_j\delta^{p}_{n_j+{\varepsilon}_jb_j-\zeta_jm_j } \frac{{c}_{\gamma , p}(n , m ) } { z_p(\sigma \times s^1)}\ ] ] we conclude using proposition [ comptage ] where @xmath288 stands for the @xmath421-tuple @xmath422 , @xmath423 , @xmath424 for @xmath425 , @xmath426 and @xmath427 . [ comptage ]
let @xmath181 be an integer , @xmath428 be a surjective linear map @xmath429 , and @xmath430 some integers . then setting @xmath431 the quantity @xmath432 has a cyclic expansion .
let @xmath433 be a disjoint union of handlebodies bounding @xmath402 and @xmath334 be a unitrivalent banded graph embedded in @xmath433 such that @xmath434 and such that @xmath433 retracts on @xmath334 .
we will denote by @xmath435 the set of edges of @xmath334 and identify an element of @xmath333 with the edge incident to it .
then it follows from @xcite that @xmath436 is the cardinality of the set @xmath437 where we have set @xmath438 , \tau_i\le
\tau_j+\tau_k,\tau_i+\tau_j+\tau_k\le 2\}\ ] ] @xmath439 in these two expressions @xmath440 runs over all triples of edges in @xmath435 incident to a same trivalent vertex of @xmath334 .
this comes from the fact that @xmath441 has a basis obtained by coloring the edges of @xmath334 with @xmath365 for @xmath442 satisfying @xmath443 for all @xmath444 , see theorem 4.11 in @xcite .
observe that for all @xmath445 , @xmath446 where @xmath447 if @xmath448 and @xmath255 otherwise . plugging this into formula gives @xmath449 where @xmath450 has the same definition as @xmath451 with @xmath452 replaced with @xmath453 .
we will see that this sum is actually finite so that we are reduced to study the cyclic expansion of @xmath450 . as an example of what will follow , taking @xmath454
, we have the equality @xmath455 .
comparing the number of integral points with the volume gives the estimate @xmath456 here @xmath457 if @xmath458 and @xmath274 if @xmath459 , @xmath460 is computed from the lebesgue measure on @xmath461 and @xmath462 is the the volume of @xmath463 where @xmath464 is the vectorial part of the affine lattice @xmath465 .
one can compute @xmath466 if @xmath458 and @xmath274 if @xmath459 .
notice that up to a normalization factor , the leading order is the volume of the moduli space @xmath467 with the volume form associated to the atiyah - bott - goldman symplectic structure .
the proof of the general case will rely on the same kind of estimations .
more precisely , let @xmath468 be the affine subspace of @xmath469 given by the equations @xmath470 for @xmath425 and @xmath471 .
then we have @xmath472 where @xmath473 we would like to apply lemma [ integral ] to this situation in order to estimate @xmath474 .
the lemma applies to the polytope @xmath475 included in the euclidean space @xmath468 provided that @xmath475 has non - empty interior and @xmath476 is an affine lattice in @xmath468 .
if the first assumption does not hold , it implies that @xmath477 and hence @xmath478 .
next , the linear part @xmath479 of @xmath468 is by hypothesis the kernel of the linear forms @xmath480 which are integral and linearly independent .
it follows that @xmath481 is a free abelian group of rank @xmath482 where we have set @xmath483 .
hence , @xmath476 is a lattice in @xmath468 if and only if it is non - empty .
we observe that this last condition only depends on @xmath88 modulo 2 . indeed
, if @xmath484 satisfies the equation defining @xmath468 and @xmath465 modulo @xmath84 , we can add to @xmath365 a function @xmath485 which satisfies @xmath486 and @xmath487 .
such a function exists by the assumption that @xmath0 is surjective .
it follows that we need to specify the parity of @xmath88 . in the sequel
, we will denote by @xmath488 a solution of the preceding system and restrict to those @xmath88 s which are congruent to @xmath489 . for @xmath490 mod @xmath84
we have @xmath491 we have @xmath492 , showing that this quantity does depend on @xmath88 . on the other hand ,
the function @xmath493 is a continuous and piecewise polynomial function with compact support .
it follows that @xmath494 which can be written @xmath495 for a continuous and piecewise polynomial function @xmath55 with compact support .
this finally proves the proposition .
[ integral ] let @xmath496 be a polytope with non - empty interior in an euclidean space @xmath435 of dimension @xmath147 and @xmath465 be an affine lattice in @xmath435 .
we define the radius of @xmath465 as the constant ( independent of @xmath497 ) @xmath498 .
@xmath499 in this formula @xmath500 runs over the faces of @xmath496 of positive codimension @xmath501 and @xmath502 is the volume of the unit ball in @xmath503 .
for any @xmath504 define its voronoi cell @xmath505 then we have @xmath506 , @xmath507 and @xmath508 for @xmath509 .
consider @xmath510 such that @xmath511 .
any point @xmath288 in the symmetric difference @xmath512 is at distance at most @xmath513 from the boundary of @xmath514 as we have @xmath515 .
hence @xmath516 where we have set @xmath517 .
any point in @xmath518 has its closest point in @xmath514 belonging to some face @xmath500 of @xmath496 of positive codimension .
this gives the following estimation , proving the lemma .
@xmath519 consider a surface @xmath25 and a weighted multicurve @xmath232 .
the aim of this subsection is to provide a geometric interpretation for the evaluations @xmath520 $ ] .
recall that we have set @xmath521 , @xmath489 the volume form and @xmath165 .
every component @xmath231 of @xmath106 defines a map @xmath522 $ ] by the formula @xmath523).\ ] ] suppose that @xmath106 is a pants decomposition of @xmath25 and denote by @xmath334 the corresponding graph . then , the functions are well - known to be action variables on @xmath161 , see @xcite .
precisely , these maps poisson commute and the joint map @xmath524 has image the polytope @xmath496 defined in the proof of proposition [ comptage ] .
the hamiltonian flows of the maps @xmath525 induce an action of @xmath469 which act transitively on the fibers of @xmath526 . moreover , the kernel of this action is precisely @xmath527 .
we prove here the following theorem which implies theorem [ witten - sigma ] and extends the main result of @xcite .
[ sigma ] let @xmath75 be an odd integer and set @xmath528 .
suppose that @xmath14 goes to infinity such that @xmath75 and @xmath1 are coprime .
then @xmath529=\frac{1}{\nu_g}\int_{x(\sigma)}\prod_j \operatorname{tr}\rho(\gamma_j)^{a_j\sigma } d\nu(\rho)\ ] ] if @xmath530 is even and @xmath255 otherwise . before entering into the proof ,
let us see how this theorem implies theorem [ witten - sigma ] .
for any banded link @xmath131 , we set @xmath531 .
the compatibly with kauffman relations implies that @xmath532 is a linear form on @xmath533 . on the other hand ,
let @xmath534 be a pants decomposition of @xmath25 and @xmath535 be the sub - module generated by banded links living in @xmath147 , the product of a neighborhood of @xmath106 with the circle .
we denote by @xmath536 the function satisfying @xmath537id .
any element of @xmath535 may be viewed as a function on the character variety @xmath155 which depends only on @xmath538 and @xmath539 .
that is , we can associate to any banded link @xmath540 a function @xmath541 satisfying @xmath542 in this setting , we define a linear form @xmath543 by the formula @xmath544 where @xmath545 is the image of @xmath546 by @xmath547 . hence , by construction , we have @xmath548 .
theorem [ witten - sigma ] consists in proving the formula @xmath549 for a banded link @xmath0 projecting without crossing . on the other hand , weighted multicurves
@xmath238 $ ] with fixed @xmath106 are generators for @xmath550 , hence it is sufficient to prove the equality @xmath551=\lambda_\sigma ' [ \gamma , w]$ ] for all weights @xmath233 .
fix a weight @xmath239 . in the torus
@xmath552 $ ] , the element @xmath553 corresponds to the function mapping the representation @xmath554 $ ] to @xmath555 .
hence @xmath556 and @xmath557=\frac{1}{2\nu_g}\int_{x(\sigma\times s^1)}\prod_j \operatorname{tr}\rho({\gamma_j}^{\sigma a_j})\tau^{\sigma b_j } d\nu(\rho)$ ] . as @xmath155 is a disjoint union of two copies of @xmath161 defined by the equations @xmath558 and @xmath559 , the integral vanishes if @xmath530 is odd , and otherwise reduces to the integral in the right hand side of the equation in theorem [ sigma ] .
the conclusion follows . with the notation of the proof of proposition [ existence ] , we have the formula @xmath560=\sum_{m , n,{\varepsilon},\zeta}e^{\frac{2i\pi\sigma}{p}\sum_j { \varepsilon}_ja_j n_j}\prod_j \delta^p_{n_j+{\varepsilon}_jb_j-\zeta_jm_j}\frac{c_{\gamma , p}(m , n)}{z_p(\sigma\times s^1)}.\ ] ] for @xmath561 to be non - zero we need to have @xmath562 . at the same time there exists @xmath563 such that @xmath564 .
hence if @xmath565 , we have necessarily @xmath566 .
if @xmath567 , we have either @xmath566 and @xmath568 or @xmath569 and @xmath570 .
the number of solutions is bounded with respect to @xmath571 , and such terms can be neglected in the sum .
hence , we can suppose that @xmath565 and replace @xmath452 with @xmath453 .
this gives @xmath560=\sum_{n,{\varepsilon}}e^{\frac{2i\pi\sigma}{p}\sum_j { \varepsilon}_ja_j n_j}\prod_j \frac{c_{\gamma , p}(n+{\varepsilon}b , n)}{z_p(\sigma\times s^1)}+o\big(\frac{1}{p}\big)\ ] ] where @xmath572 is a shorthand for the tuple @xmath573 .
let @xmath334 be the graph adapted to @xmath106 introduced in the proof of proposition [ comptage ] .
we also introduce the polytope @xmath496 and the lattice @xmath465 .
applying lemma [ integral ] , we get that @xmath561 is either zero or equivalent to @xmath574 where @xmath575 is the affine space given by the equations @xmath576 and @xmath577 for @xmath425 .
supposing non - triviality , we get further @xmath578 denote by @xmath579 the data corresponding to gluing the vertices @xmath580 and @xmath581 of @xmath334 for @xmath425
. then , the same reasoning gives @xmath582 .
as @xmath583 and @xmath584 are equal up to @xmath585 , the volume @xmath586 is equal to @xmath587 up to @xmath585 where @xmath588 is the subspace of @xmath589 defined by the equation @xmath590 .
let us deal with the covolumes by considering first the case of @xmath591 .
let @xmath592 be the cellular complex associated to @xmath591 .
a coloring @xmath593 reduces modulo 2 to an element of @xmath594 . in this setting ,
the condition defining the vectorial part of @xmath595 is simply to be a cycle .
it follows that the index of @xmath595 in @xmath596 is the cardinality of @xmath597 which is @xmath598 .
the same reasoning applies to @xmath334 , supposing that @xmath599 is non - empty .
the vectorial part of @xmath600 is a map @xmath601 which vanishes at boundary points and such that @xmath602 is even for all triples @xmath440 incident to a same vertex .
its reduction modulo 2 is a cycle in @xmath603 , hence , the index of @xmath600 is @xmath604 . supposing it is non - zero , we get the following estimate where @xmath605 .
@xmath606 let @xmath381 be the number of connected components of @xmath334 .
the exact sequence of the pair @xmath607 gives the formula @xmath608 which replaces the power of @xmath84 in equation with @xmath609 .
consider now the problem of the non - vanishing of @xmath561 .
an element @xmath610 satisfies @xmath611 , @xmath612 and @xmath613 for all triple of edges @xmath614 incident to a same vertex .
summing @xmath615 twice over each edge of any component @xmath616 of @xmath334 gives modulo 2 the equality @xmath617 where @xmath618 .
summing over the connected components and observing that @xmath619 modulo 2 , this gives the identity @xmath620 .
conversely , a simple argument involving the homology of @xmath621 modulo 2 implies that @xmath561 is non - zero if and only if @xmath622 for all @xmath403 .
hence , assuming that @xmath530 is even , we get @xmath560=\frac{1}{r^k}\sum_{\substack{n,{\varepsilon}\\ l_i(n+{\varepsilon}b , n)=0}}e^{\frac{2i\pi\sigma}{p}\sum_j { \varepsilon}_ja_j n_j } 2^{l-1}\frac{\operatorname{vol}p\cap \hat{v}_{\frac n r}}{\operatorname{vol}\hat p}+o(\frac{1}{p})\ ] ] the parity conditions @xmath623 should divide the sum by @xmath624 but the sum @xmath625 vanishes modulo @xmath84 . hence they divide the sum by @xmath609 and this factor cancels . we recognize a riemann sum : setting @xmath626 we get @xmath627&=&\int_{\mathbb{r}}\sum_{{\varepsilon}}e^{2i\pi\sigma\sum_j { \varepsilon}_ja_j x_j}f(x)dx\\ & = & \int_{\mathbb{r}}\prod_j 2\cos(2\pi\sigma a_jx_j)f(x)dx.\end{aligned}\ ] ] this proves the proposition as @xmath628 is equal to the push - forward @xmath629 and by construction @xmath630 is conjugate to @xmath631 .
we would like to thank laurent charles , gregor masbaum and maxime wolff for valuable discussions and encouragement . | for a banded link @xmath0 in a surface times a circle , the witten - reshetikhin - turaev invariants are topological invariants depending on a sequence of complex @xmath1-th roots of unity @xmath2 .
we show that there exists a polynomial @xmath3 such that these normalized invariants converge to @xmath4 when @xmath5 converges to @xmath6 , for all but a finite number of @xmath6 s in @xmath7 .
this is related to the amu conjecture which predicts that non - simple curves have infinite order under quantum representations ( for big enough levels ) . estimating the degree of @xmath3
, we exhibit particular types of curves which satisfy this conjecture . along the way we prove the witten asymptotic conjecture for links in a surface times a circle . |
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.